4th Conference

code ISAP 4th Conference — Titles & Abtracts
04000 4th International Conference on Design of Asphalt Pavements – Volume I – Preliminary pages and Table of Conents
04001 The Use of Distress Prediction Subsystems for the Design of Pavement Structures
F. Finn, C. Saraf, R. Kulkarni, K. Nair, W. Smith, A. Abdullah

This report presents the results of an investigation designed to develop procedures for predicting distress in asphalt pavements using mechanistic parameters of stress, strain, and deformation. The specific forms of distress considered were fatigue cracking, permanent deformation, and low-temperature cracking. Two computer programs were developed, one for fatigue cracking and permanent deformation (PDMAP) and one for low-temperature cracking (COLD).

The programs should provide capability in simulating the occurrence of pavement distress and can be used in pavement management systems, diagnostic investigations, formulation of design criteria, and preparation of material and construction specifications.

Field verification of such programs is limited, since most of the available field data were used to develop the damage prediction models.

The programs are considered implementable and have been successfully executed on state highway department computer equipment.

One of the conclusions of the investigation is that field verification is necessary in order to calibrate the basic predictions to local conditions.
04002 Asphalt Pavement Design – The Shell Method
A. J. M. Claessen, J. M. Edwards, P. Sommer, P. Uge

The original Shell method for the structural design of asphalt pavements has been up-dated and extended to incorporate all relevant design parameters. The method is based on a model in which the pavement structure is regarded as a linear elastic multi-layered system in which the materials are characterised by their modulus of elasticity and Poisson’s ratio. The computer program BISAR is used to compute all stresses, strains and displacements at any point in the system under any number of vertical and/or horizontal surface loads. In this way, the primary design criteria have been established, i.e. the compressive strain at the top of the subgrade and the horizontal tensile strain in the asphalt. Secondary criteria such as permissible stresses in cementitious base layers, permanent deformation of the asphalt, etc., are also included.

The permissible value for compressive subgrade strain has been derived from analysis of AASHO Road Test sections and structures conforming to CBR design. The permissible asphalt strain was determined from extensive laboratory measurements for various mix types at different stiffness moduli of the asphalt. In the application of the asphalt fatigue criterion allowance is made for the influence of the transverse distribution of wheel loads and for effects of healing and intermittent loading.

The traffic data are converted into an equivalent number of standard design axle load applications. To introduce the influence of the ambient temperature a procedure has been developed to relate the mean annual or monthly air temperature to an effective asphalt temperature, depending on the thickness of the asphalt. The moduli of subgrade and unbound base layers should be determined at appropriate stress levels whereby the latter modulus is a function of the subgrade modulus. The asphalt moduli of a large number of typical asphalt mixes have been determined from extensive laboratory measurements. It is demonstrated that the modulus of a given mix, relevant for the structural design, can also be derived with sufficient accuracy using a nomograph. For practical design purposes it is appropriate to take a loading time of 0.02 seconds and a Poisson’s ratio of 0.35 for asphalt and unbound materials and 0.25 for cementitious base layers.

To provide a practical system for road engineers, sets of design charts have been prepared from which combinations of thicknesses of the asphalt and unbound base layers can be derived for various mean annual temperatures, for a number of typical mixes and for various subgrade moduli. Examples of these curves are given. Special attention is paid to a method of predicting the permanent deformation (rut depth) of the asphalt layers during the expected service life of the pavement. In this case the traffic data are converted into an equivalent number of standard single wheels.

The application of the design method to the design of pavements for aircraft with multiple wheels is also dealt with. Examples of design curves for some types of aircraft are given.

The various laboratory tests and full scale road trials carried out to investigate the validity of the design method have been summarised. The practical use of the design method is illustrated by means of some examples.
04003 Fully Monitored Motorway Trials in the Netherlands Corroborate Linear Elastic Design Theory
W. G. Bleyenberg, A. J. M. Claessen, F. Van Gorkum, W. Heukelom, A.C. Pronk

In the Netherlands, the Studie Centrum Wegenbouw (Study Centre for Road Construction) set up a Working Party to investigate the applicability of theoretical design models by means of comprehensive experiments on fully monitored motorway trial sections. The road trials were designed to cover a wide variety of conditions. In two types of construction – one with full depth asphalt and one with a sand-cement subbase – strains at different levels in the structures, surface deflections and soil pressures were measured under a variety of conditions of temperature and of load, speed and lateral position of the loading vehicle. Both test sections were constructed in stages. The paper gives representative results of measurements for two stages of construction of the pavements.

Material properties of subgrade, sand cement and asphalt and the thicknesses of the layers were measured both in situ and in the laboratory. These data were then introduced into the BISAR computer program, based on linear elastic multi-layer theory, to calculate values of strains, deflections and soil pressures corresponding to the measured quantities and the results have been compared with one another. In addition, BISAR computations have been made to investigate the influence of a variation in test parameters (moduli, Poisson’s ratios, shape and stress distribution of wheel contact area, etc. The widths (durations) of recorded regular strain signals were used to characterize the loading time.

Measured and computed quantities are mostly presented as a function of the stiffness modulus of the asphalt (incorporating the effect of temperature and loading time). All data could be converted into values per unit load, demonstrating that the behaviour of both structures was linear. Generally, the results showed increasing scatter at lower asphalt moduli. Computations showed that variations in test conditions also have their strongest influence at low asphalt stiffness. In some cases a stronger difference between longitudinal and transverse strain was found as compared with the computed values.

On the whole, measured values of strains, deflections and soil pressures showed good agreement with the corresponding computed values, particularly with higher asphalt moduli.
04004 Predictive Design Procedures – A Design Method for Flexible Pavements Using the VESYS Structural Subsystem
William J. Kenis

This paper describes a procedure for the analysis and design of flexible pavements. The analytic techniques upon which this procedure is based have been assembled from the application of sound fundamental principles and their evaluations from results of laboratory and field studies. The concepts used have been formulated in State Highway Planning and Research, National Cooperative Highway Research Program, Administrative Contract, and FHWA staff studies, Work on the procedure was initiated by the Pavement Systems Group of the Offices of Research and Development in order to develop a reference document to assist the pavement designer in analyzing the structural integrity of flexible pavements. Work was accomplished under Federally Coordinated Program Project 5C "New Methodology for Flexible Pavement Design."

The design procedure described includes use of the most recent developments for the analysis and design of flexible pavements. Although it may be used to evaluate the distress and performance of in-service pavements, the procedure is written specifically for application to the structural analysis and design of new pavements. In either case a pavement section of known geometry is chosen and its behavior over time is predicted by sets of mechanistic models which have been computer programmed. The computer program package, known as the VESYS Computer System, computes distress and performance in terms of rutting, roughness, and crack damage. These damage indicators are then used in a distress performance relationship to predict the serviceability history of the pavement. Results of the analysis must be compared with sets of design criteria imposed by the user agency. The analysis process is repeated until the computed levels of damage and serviceability meet the acceptable levels established by the design criteria.
04005 Comparison of VESYS IIM Predictions to Brampton/AASHO Performance Measurements
J. Brent Rauhut, R. C. G. Haas, Thomas W. Kennedy

VESYS IIM is one of the most advanced models to date for simulation of a flexible pavement structure. However, this does not within itself imply its adequacy for practical use. Thus its ability to predict the performance of inservice pavements is of primary importance to its eventual utilization for analysis and design. The paper provides comparisons of VESYS IIM predictions and measured performance of four sections of the Brampton Test Road and four sections of the AASHO Road Test.

VESYS IIM as currently configured normally requires 47 input variables, twelve of which include multiple values in arrays. For simplicity, these input variables may be classed into four categories having rational engineering significance when considered separately. These classifications include:
1) pavement thickness dimensions,
2) traffic and wheel load data,
3) materials characterizations, and
4) control variables.

The most important departure of VESYS IIM from more conventional analytical methods is in material characterizations. These material characterizations include creep compliance curves for viscoelastic responses, permanent deformation coefficients to describe permanent deformation potential for each layer, and the fatigue coefficient and exponent commonly used for the asphalt concrete characterized with respect to temperature. A very detailed contemporary test program has been conducted for the Brampton Test Road so much useful data was available. For the AASHO Road Test, however, it was necessary to obtain material samples and to conduct specialized testing to obtain the necessary material input.

Comparisons of predicted and measured values show that performance predictions may not as yet be made at a sufficiently high confidence level. This is not surprising, however, considering the state-of-the-art for characterizing materials and inability of theoretical structural models in current use to deal with material changes with time. A significant finding of this study is that the calculated rut depth values approximated the measured rut depth for three of the eight sections when reasonably accurate input values were used. This success, even though limited, suggested that VESYS IIM holds considerable promise and may be expected to offer better predictive capabilities as improvements are made to both the program and to the material characterization procedures.

While no definite conclusions can be made concerning predictions for cracking, present serviceability, and service life, these models do appear to function rationally and should also provide improved predictions as the state-of-the-art for materials characterizations and for the models themselves progress.
04006 Sensitivity Analysis of FHWA Structural Model VESYS
J. Brent Rauhut, John C. O’Quin, W. Ronald Hudson

VESYS IIM is one of the most advanced models developed to date for simulation of a flexible pavement structure. As currently configured, it normally requires 47 input variables, 12 of which involve arrays of multiple values. As the development of the input values can be time consuming and some require data not normally available, it is important to understand the relative effects of the individual variables upon practical solutions. For instance, there is no reason to implement a detailed material testing program or an analytical study to obtain accurate values for an input variable that has little or no effect on the calculated responses. On the other hand, a variable to which the solution is very sensitive may warrant an extra amount of effort for accurate characterization.

This paper describes and reports results of an extensive sensitivity analysis for VESYS IIM conducted by Austin Research Engineers Inc under contract to the FHWA. Preliminary screening reduced the number of independent variables to be considered and allowed division of the analysis into two independent studies, one for the VESYS IIM roughness model and one for the VESYS IIM cracking model. These considerations permitted a reduction of the problem to two factorial experiments, one of 11 variables for the cracking model, and one of 15 variables for the roughness model. Fractional factorial techniques were then employed to effect an enormous reduction in the number of solutions required for the sensitivity analysis. The result was a regression equation for each calculated response representing the VESYS IIM model for that response and leaving little information "unexplained." The independent variables in the equation were coded in an orthogonal scheme, which allowed relative comparisons of the variations for each factor in terms of expected range of values in the field.

The results of this sensitivity analysis are reported and it is clear that a number of the variables have so little effect that little effort is warranted to obtain relative high accuracy for their values. This fortunately includes a number of variables that are difficult to evaluate such as the time-temperature shift function for the asphalt concrete, load duration and distribution, permanent deformation characterization for the base material; the correlation between the fatigue coefficient K sub 1 and the exponent K sub 2, creep compliance for the various layers, and stochastic distributions for several of the variables. Estimates will suffice for a number of these variables and can also be added to the computer program as constants for use when values are not available and are not furnished as input.

In general, the importance of the independent variables to the computed responses is consistent with known physical realities. As might be expected, primary emphasis for the rutting predictions should be placed on reliable values for the permanent deformation characterization for the asphalt concrete surface material. If a prediction of cracking damage is of special interest, a reliable relation for fatigue life potential must also be obtained. The relative importance of all significant variables is described.
04007 Field Verification of the VESYS IIM Structural Subsystem in Utah
D.I. Anderson, J.C. McBride, D.E. Peterson

The need to encorporate elastic and viscoelastic models into a practical flexible pavement design procedure is well founded. A limited verification of the VESYS IIM design procedure, and an investigation of the possibilities of implementing the system in Utah were carried out.

The program is based on a 3-layer model corresponding to the bituminous, base gravel and soil materials. The test samples were formed using the identical asphalts and gravel materials placed in the selected pavements. Air-tight containers were used to store the asphalts from the time of construction, and were found to be relatively unaged. The indirect tensile test was employed to obtain the creep compliance, permanent deformation, and fatigue data needed to generate the projected performance data. The base gravel and soil materials were tested using a cylindrical test specimen under confining pressure to obtain the creep compliance and permanent deformation data for the second and third layers. Actual traffic data measured over the pavement life was used as input.

Estimates of the performance of some pavements were produced, and compared with the measured performance data. These predictions were found to be adequate for the magnitudes and times measured. Future work in the study will evaluate the procedure in a wider range of performance data, pavement materials, traffic loadings, and environmental conditions.

Implementation of the procedure in its present form would require significant changes in the structure of Utah’s present overall design scheme. The obligations of various state organizations would require alteration. Also, some form of prediction of transverse, temperature- associated cracking would be necessary in many states to obtain an optimum design based on performance.

Preliminary results show that the VESYS IIM program has great potential, and every effort should be made to fine tune the models and encorporate the procedure into practical application.
04008 Evaluation of Flexible Pavement Design Methodology Based Upon Field Observations at PSU Test Track
M.G. Sharma, W.J. Kenis, T.D. Larson, W.L. Gramling

Field observations as related to distress and performance of test flexible pavements under accelerated traffic loading at the Pennsylvania Test Track Facility have been presented and correlated with the corresponding value as predicted by the FHWA computer program (VESYS IIM). Eight flexible pavement test sections carrying about 1.5 million 18 kip equivalent axle loadings were evaluated. On each section, rut depth, cracking and present serviceability index (PSI) were measured at periodic intervals. An account was made of the environmental and traffic conditions throughout the test period and samples of materials used in the facility were tested in the laboratory to determine the properties necessary for input to the VESYS Computer Code. Using the VESYS IIM program estimates of the extent of fatigue cracking damage, rut depth, and PSI were computed and compared with measured performance data. Results indicate that it is possible to predict the relative magnitudes of these distress and performance indicators. In this paper an attempt is made to point out reasons for deviations between measured and predicted values. Suggestions are made concerning the eventual use of mechanistic structural subsystem for the analysis and design of flexible pavement systems.
04009 Implementation and Verification of Flexible Pavement Design Methodology
Jatinder Sharma, L.L. Smith, B.E. Ruth

Presented in this paper is the evaluation of two computer models(VESYS and PDMAP) for the design of flexible pavements. The Chiefland, Florida test section was chosen to verify the above mentioned models. This section is the first project of a series of ten projects to be studied over a three year period.

Laboratory tests on the Asphaltic Concrete Surface Course consisted of an Incremental Creep Test, a Dynamic Compression Test and a Flexural Fatigue Test; on untreated materials, the tests consisted of Incremental Creep and Dynamic Compression at ranges of stresses expected at in situ. The Kneading compactor was used to make samples and a MTS electrohydraulic closed loop system was used to run the tests.

Fatigue tests on the fabricated beam specimens showed a higher potential for fatigue life than beam specimens from the undamaged portion of the roadway which is about 10 years old. Aging is a significant factor in the prediction of pavement performance.

It is possible to predict cracking and rut depth with reasonable degree of accuracy, provided that the Modulus of Elasticity for different materials is based on the measurements over the entire length of the sample.
04010 Selection of Pavement Models in a Rational Pavement System to be Used by the Dutch State Road Laboratory
R. Frank Carmichael, Jan Eikelboom, Peter M.W. Elsenaar, W.R. Hudson

Simulations of the observed pavement performance parameters (i.e., skid resistance, rutting, structural cracking, and evenness) have been prepared using behavioral models, and data from in-service Dutch pavements. The study was performed to evaluate and compare various available mathematical models for predicting pavement behavior, for possible use as elements in the development of a Rational Pavement Management System (RVBS), by Rijkswegenbouwlaboratorium (RWL) for use in Holland. The predictions of the behavioral models were compared with available field measurements from different test sections in an attempt to verify each model’s capabilities. Rutting and structural cracking models were of primary interest because of their effects on structural design, evenness and skid resistance models because of their effects on safety and comfort.

The study included the following work items:

1. The Selection of Pavement Structures – Existing pavement structures in Holland were selected for the simulations and data concerning pavement construction, traffic, and pavement condition survey data were used as input for and verification of the simulations.

2. Model Selection – After review of 15 mathematical prediction models of pavement behavior, a total of eight behavioral models, two in each category of interest, skid, evenness, rutting, and cracking, were selected for structural simulation. Simulations of performance of selected pavement structures were prepared using these eight models, and the results were reviewed to ascertain the validity of the models.

3. Simulation Results – The simulation results from the eight behavioral models, recommendations as to future developments of RVBS, and improvements which are necessary for the use of any of these models in RVBS are presented.

This is the first phase of a three stage development of a Rational Pavement Management System, RVBS, for Holland. The study will help RWL to select behavioral models for use in the first working RVBS, which is to be developed in the second phase of work.
04011 A General System for the Structural Design of Flexible Pavements
W.R. Barker, W.N. Brabston, Y.T. Chou

Presented herein is a procedure for design of three types of flexible pavement: conventional, bituminous concrete, and chemically stabilized. These represent nearly all flexible pavements being constructed at this time. Designs are based on analytically determined strain values as computed by a layered elastic model of a pavement structure. The subgrade strain is related to performance through empirically determined criteria whereas the strains in structural layers are related to performance through laboratory determined material fatigue strengths. With the exception of granular materials and cracked stabilized materials, stiffness properties of the pavement material are determined through laboratory testing. For the granular materials and the cracked stabilized materials, empirically based charts are provided for determination of the stiffness. An adaptation of the cumulative damage permits the consideration of cyclic variation in bituminous materials due to variations in temperature, the variations in subgrade strength resulting from freeze-thaw cycles, the mixture of different traffic vehicles, and the distribution of traffic across the width of the pavement. An example problem is provided which illustrates the use of the procedure.
04012 Esso Road Design Technology
B. Celard

The pavement design method developed at ESSO embodies a rutting and a fatigue subsystem. The input data encompasses the climate, traffic and subgrade variables on the one hand, and the asphalt mixes dynamic properties on the other. These properties are measured through three tests: dynamic modulus, dynamic creep and uniaxial push-pull fatigue tests. They are performed on realistic specimens prepared with the Esso laboratory compactor. The test results are processed statistically and the damage laws are obtained: these connect the elemental damage undergone by the material to each stress and temperature conditions. The stresses inside the pavement are computed assuming the pavement can be modelized as a three-layer elastic system. The total damage caused to the pavement is obtained by integrating the related elemental damage over the climate and traffic variables:
– the rutting subsystem attempts to provide an actual prediction of the rut depth development for surface courses, asphalt overlays and asphalt layers laid on top of stiff foundations,
– the fatigue subsystem only provides a comparative index which allows arraying the different structures, and can be used as an optimization criterion.

Besides the method itself, a simplified testing equipment for asphalt mixes was developed. Its simplicity and low cost allow performing the dynamic tests on a routine basis and therefore contribute to make the design method available to practising engineers.
04013 Kentucky Research: A Flexible Pavement Design and Management System
Herbert F. Southgate, Robert C. Deen, James H. Havens, William B. Drake jr.

Various strategies for designing pavement structures are discussed. Initial full-life design, stage designs and planned extensions of service life, final design, surface renewals for deslicking, no-defect designs for high-type high-volume facilities, and allowable-defect designs are considered. Economics enter in terms of salvage value of existing pavements and alternate designs using different proportions of materials within the structure.

The elastic model represented in Chevron’s n-layered computer program is the basis for theoretical relationships. Ranges of values are given for input variables such as Young’s moduli, Poisson’s ratio, thicknesses for layers, tire pressure, and load. The Kentucky CBR is related to modulus by E = 1500 x CBR and is correlated with the AASHO Soil Support value and other strength relationships. The modulus of crushed stone base is shown to be a function of the moduli of the asphaltic concrete and subgrade. Appropriate relationships are given.

Graphs show interrelationships between asphaltic concrete thickness and asphaltic tensile strains, subgrade strains, and surface deflections. Structures are interpolated as a function of asphaltic concrete thickness and percent of asphaltic concrete in the total thickness. These graphs are interpolated to illustrate the relationship of Kentucky CBR versus total pavement thickness for structures in which the asphaltic concrete has a specified modulus and accounts for a specified proportion of the total thickness. The CBR-total thickness graphs are combined to make the Kentucky design nomographs. All of the above are theoretical and empirical relationships which are not dependent upon fatigue criteria.

A tensile strain-fatigue criterion is derived and correlated with Kentucky experience. The analysis is based on limiting elastic energies and may be treated thereafter in terms of limiting strains or limiting stresses.

The subgrade strain criterion is also associated with Kentucky field experience and is extended as a function of axleload; load and repetitions are interrelated to provide equivalent axleloads in kips for single axleloads, and EAL’s for single axleloads. These values are illustrated as log strain – log repetitions of 18 kip (8OkN) equivalents. A method for determining a design CBR value is discussed.

Methods of determining design periods, or design life, are discussed. Ways of determining appropriate traffic volumes are discussed and include available AADT data, use of classification and weight data from W-4 Tables, and modifications to account for known and unusual traffic variations such as for recreational areas, rural routes, and heavy hauling operations.

The methodology for computing load-damage factors is given, discussed, and compared to damage factors developed by the AASHO Design Committee from Test Road data. The value of subgrade strain correlated to Kentucky experience is that strain developed for an 18 kip (80 kN) axleload applied to typical Kentucky designs. Other axleloads producing other strains in the same structures are expressed as ratios to the 18 kip (80 kN) axleload. Thus, the coupling of repetitions, axleloads, and subgrade strain is accomplished in the load – repetitions equations. Graphs relating repetitions of 18 kip (80 kN) axleloads to subgrade and asphalt strains allow determination of the respective strain values for input to the CBR – thickness graphs or the Kentucky design nomographs to obtain design thicknesses with respect to the asphaltic concrete moduli.

The Kentucky criterion for rutting allows maximum rutting for the least traffic but minimum rutting for the highest traffic. A procedure is developed for adjusting thicknesses to a fmal design thickness for each of the four moduli given in the design nomograph. The appropriate modulus for a given locale is determined from pavement temperature histories. A method is given in this paper. For Kentucky conditions, the recommended modulus is a function of the percentage of asphaltic concrete thickness in terms of the total design thickness.

The choice of percentage of asphaltic concrete of the total design thickness and their respective moduli permit development of design charts showing the thickness versus log repetitions of 18 kip (80 kN) axleloads. Each line on the chart is for a given CBR. All the designer needs to use these charts is the design CBR and number of repetitions. Since the design charts are mathematical solutions coupled with fatigue criteria, changes in the fatigue criteria can be properly made by simply relabeling the CBR lines to appropriate equivalent values. The design charts can be used to design overlays and to determine when overlay construction should be scheduled in terms of accumulated EAL’s. Again, AADT charts and yearly W-4 table classification and weight data can be readily used to estimate accumulated EAL’s and to modify the design period, or design year, to account for increased weight limits, changes in style of cargo hauling equipment, routing, and distribution of vehicle types in the traffic stream.

Verification of this design system involve several sources, including some 30 to 40 years of design and behavioral experience within Kentucky and including a full-depth asphaltic concrete research pavement. Recent investigations involve re-analysis of AASHO Test Road data — both single and tandem axleloads. Results indicate the AASHO Test Road data can be analyzed and explained by elastic theory. The design level of terminal serviceability is shown to be a function of design repetitions. This relationship is also expressed in terms of axleloads. Coupling the variable terminal serviceabilities with axleloads and, in turn, with Kentucky equivalent damage factors petit superpositioning of the AASHO Test Road data in terms of 18 kip (80 kN) EAL’s for single axleloads. An additional relationship equating equal damage for single and tandem axles using AASHO Road Test data permits superpositioning of the tandem-axleload data over the single-axleload data with virtually no increase in data spread. Superimposed on the AASHO Road Test data are Kentucky moduli solutions showing excellent agreement of field data and theory.

Airport pavement thickness designs obtained from The Asphalt Institute’s Full-Depth Asphalt Pavements for Air Carrier Airports is closely duplicated by converting aircraft wheel loads to equivalent 18 kip (80 kN) axleloads. A method is given equating moduli with frequency of loading cycle. It is shown that the Kentucky asphalt tensile strain criteria can be expressed by an equation (given in the paper).

The verification of this design system involves the incorporation of test data from Kentucky, Canada, the AASHO Test Road at Ottawa, Illinois, and test roads in Colorado, San Diego, and Kentucky. In addition, The Asphalt Institute’s airport pavement design method uses aircraft strut and wheel loads which can be converted to an equivalent highway 18 kip (80 kN) EAL loading for pavement design by the Kentucky method. In one example, The Asphalt Institute’s airport pavement thickness was 17.0 inches (332 mm) full depth asphaltic concrete. The same aircraft loadings resulted in a Kentucky design thickness of 17.2 inches (337 mm).

Pavement management concepts are discussed and a method is presented illustrating the required data and its use to accumulate EAL’s annually for comparison with the design EAL. This method can be used to determine overlay priorities, overlay design thicknesses, and scheduled financing. Condition survey analyses coupled with economic analyses may indicate a pavement structure should be reconstructed, or even relieved by constructing a new corridor. A discussion of automatic feedback of field data is presented. Pavement condition reports may be fed into the data bank. However, such data should be analyzed separately to prevent improper adjustments to the design system due to causes of failure other than pavement fatigue. Construction of overlays for any purpose should be logged into the data bank. The overlay, whether for extending service life or improving skid resistance, provides an additional structural thickness and will modify the design life.
04014 The Belgian Road Research Center’s Overall Approach to Asphalt Pavement Structural Design
J. Verstraeten, J.E. Romain, V. Veverka

A structural design system for asphalt pavement, based upon the utilization of measured fundamental properties of materials, is presented, While the experimental methods for the determination of the fundamental properties of the materials are now well defined, they involve sophisticated tests which are not well adapted to everyday practical purposes (long time, high cost). For this reason predictive methods, based on numerous laboratory results obtained for the different kinds of materials, are made use of, mainly for bituminous mixes and granular materials.

The different elements of design, namely, criteria for riding quality, loading conditions, stresses and strains in layered systems, fundamental properties of road materials and climatic factors are reviewed. Models for cracking and for rut depth are supplied. A field verification is reported,

For the pavement design method itself, two approaches are presented. The first is a complete rational one, while the second is more approximative, but simpler than the first and, consequently, more practical.
04015 The Application of Simplified, Fundamental Design Procedures for Flexible Pavements
S.F. Brown, P.S. Pell, A.F. Stock

Prior to acceptance in practice of complete analytically based pavement design procedures, engineers have to be introduced to the basic concepts involved and this can best be achieved by use of a simplified approach. The Simplified Design Method has been successfully used for such educational purposes and is described in detail. It makes use of charts and equations but has also been programmed onto a computer of modest size. The Analytical Design Method is a more versatile, but still somewhat approximate, design tool in the form of a computer program. The immediate use which can be made of a design package of this kind is demonstrated by investigating the effects of mix variables on layer thickness.
04016 Method for the Structural Design of Asphalt Pavements
J. Eisenmann, U. Lempe, C. Leykauf

In the outlined design procedure the flexible pavement structure is considered as a linearly elastic multi layer system in which the road building materials are characterized by Young’s modulus and Poisson’s ratio. By aid of the computer programs BISTRO and BISAR, developed by Shell, the maximum stresses at the interfaces of the layers were determined. The primary criteria for the behaviour of the asphalt layer has been found to be the main shear stress (tau)max at the bottom of this layer. The stressing of the unbound base courses and of the subgrade can be evaluated by the compressive stress (delta)(sub)z.

As to the dynamic properties of bituminous bound materials extensive laboratory tests have been carried out. The Young’s modulus has been determined in shear tests in subjection to temperature and loading time, whereat the parameters — grain shape, binder hardness, binder content and void content — have been varied. With the same test equipment also the fatigue under shear stressing has been investigated ; the paper presents the influence of the different parameters.

Using the hypothesis of Miner the consumption of life time of a flexible pavement can be determined by accumulation of the ratios occurring load cycles n(sub)i to permissible load cycles N(sub)i for all stressing conditions. For this the occurring traffic can be substituted by an equivalent number of load cycles with a standard axle load using the equivalence-factors known from the AASHO road test. Due to the strong influence of temperature to the asphalt properties different seasons with ambient temperatures in the pavement have to be taken into consideration. The results of continuous temperature measurements are presented.

By Miner a cumulative failure can be expected if the total of Sigma n(sub)i/N(sub)i is greater than 1. But this fictive life time is not identical with the end of the service life of the pavement. The results of a theoretical investigation are presented, in which the computed fictive life time of two test sections of the AASHO road test is compared with the measured Present Serviceability Index p ; it is evident that the theoretically worked out permissible number of load repetitions gives the time when cracking of CLASS 1 – as defined in the AASHO road test – occurs, that is at an index of p approx = 4,0.

This shows, that the theoretical worked out service life will differ more or less from the reality. Therefore, when designing a new flexible pavement construction first the consumption of service life for a standardized road construction approved in praxis has to be determined in the mentioned manner and then the thickness of the new construction has to be varied so that this critical value is not overpassed. In this paper an example of designing a new pavement construction with an insulation layer out of hard foam is given. It is expected that this design method can be improved by the results of the test track "Hilpoltstein" ; in a theoretical study statements have been given on the service life of this temporary by-pass which shall be compared with real behaviour.
04017 Procedure for the Structural Design of Pavement Used on Italian Motorways
Franco Giannini, Gabriele Camomilla

A description of new methods for the structural design of flexible pavement should dwell mainly on new procedures for determining fatigue resistance and resistance against the accumulation of permanent deformations. It may be of equal importance, however, to set out a complete calculation procedure, based on already known theoretical formulations, but integrated with a series of practical suggestions indispensable for the functional transformation of this "rational" method, into an effective design.

The practical assumptions and specific evaluations set out herein, have been proposed, though in hypothetical form, in a number of sources in existing literature, but are essentially the fruit of repeated application and test results. The latter data have been taken from tests conducted on a modern, highly controlled motorway network operated with particular concern for travel efficiency and supported by continual and costly maintenance. This report is intended to contribute to the effectiveness of methods long proposed, by suggesting the practical systems for their step-by-step application so as to obtain reliable results. These suggestions are presented in a general fashion so that prescinding from the numerical values offered as examples in the text, they can be applied to situations other than those encountered by the authors.

Another important factor in the application of the method is the confirmation of the validity of Miner’s Rule which permits us to isolate the influence of individual design parameters, control their variability so as to determine their specific contribution to fatigue damage, and finally, and above all, to summarize these contributions, which act under widely varying conditions, in a coherent fashion.

The verification of the validity of this rule enables us to transform the current calculation procedure, i.e. the fatigue structural design scheme, which up to now has only had conceptual validity.

No explanation is given in the text of the evaluation of damage due to the accumulation of permanent deformations, but a complete description can be found in the works listed in the bibliography, which provides detailed evaluation of the initial verifications made in Italy. Obviously, given the technical acceptability of the proposed procedure, it provides a valuable tool for optimum economy in both construction and maintenance of highway pavements.
04018 An Optimal Design Procedure for Multilayer Pavements
T.E. Glynn, J.C. Byrne, R.W. Kirwan, M.S. Snaith

This paper describes a procedure for finding the optimum layer thicknesses of conventional flexible pavements. The pavement is modelled as a multilayer system subjected to the pulsed load equivalent of wheel loading on a circular contact area. The optimization scheme is based on a direct search method that compares predicted responses with specified design criteria in a multistage analysis. The direct search produces a finite set of discrete alternatives which takes into account structural stiffness and flexural fatigue. Those trial sections that satisfy the primary design constraints, and are at the same time economically attractive, are reduced to the optimum layer configuration by referring to an objective function. The objective function includes weighted terms for permanent deformation, fatigue life, and cost of construction materials in place.

The set of feasible solutions is determined on the basis of linear elastic theory while the reduced set is evaluated by non-linear analyses. The objective function is not a unique expression as its aim is to find the optimum solution relevant to local or regional requirements. The procedure is illustrated by a design for a four-layer pavement. The design avails of an elaborate set of experimental data for soils and bituminous materials.
04019 Asphalt Pavement Design for Arizona
R.A. Jimenez

There is a general consensus that a procedure for the structural design of asphalt pavements must reflect considerations of the loads imposed, physical characteristics of the pavement layers, effects of the environment on those physical characteristics, definition of a layer failure, and a means for following changes in the load response of the pavement with time in service. The report presented is concerned with some of the details in developing and describing the proposed pavement design method.

The pavement system is to be composed of three linear elastic layers — an asphaltic concrete surface, a granular base, and a subgrade. The modulus of elasticity and Poisson’s ratio are of fixed values for the surface and base courses. These reference values are partially justified on the basis that both layers meet certain minimum requirements specified for their manufacture. The value for Poisson’s ratio is also fixed for the subgrade ; however, the modulus of elasticity is presently determined from correlation with test values such as CBR , R-value, and others.

The procedure’s objective is to determine the thicknesses of the surface (H1) and the base (H2) to satisfy two design criteria. These are based on limiting radial tensile stresses at the bottom of the surface course to preclude cracking of the surface and on limiting vertical compressive strains on the top of the subgrade to minimize settlement or rutting of the surface. These criteria for failure were selected since the primary state highways do not show distress originating with shear failure in either the surface or the base. Values fo r the critical stresses and strains are obtained by calculation (using the Chevron program) at three specified points in the system. The maximum radial tensile stress occurs at a point directly under a wheel and at the first interface; the maximum vertical strain on top of the subgrade may be located either directly under a wheel or midway between two dual wheels for single and tandem axled trucks.

Satisfactory values for H1 and H2 are those that yield "fatigue life" greater than the design life of the pavement. As a consequence, the limiting stresses and strains are related to the number of applications of these expected over the design period. The general "fatigue " equations for asphaltic concrete and the subgrade are assumed to be in the form (sigma)(sub)T or epsilon(sub)C = aN(super)- b where the a’s and b’s are material constants. A procedure has been developed to evaluate the constants a and b for asphaltic concrete. However, to date we have not satisfactorily established the values for these constants in the strain equation and so are accepting the Shell characterization for strain fatigue as epsilon(sub)C = 0.0105 N(super)-0.200.

The calculations for stress and strain in the system require that the loading system be identified with the magnitude of the wheel load and the tire inflation pressure. Through the examination of loadometer studies and questionnaires the traffic loads on the highways can be characterized for the design purpose. Traffic loads are grouped into five categories which are passenger, pick-up (2P), front axle (FA), single axle (SA), and tandem axle (TA). The design values of wheel load, tire inflation pressure, and percentage of each axle category are obtained for each type of highway from loadometer data collected in previous years. Environmental effects on the physical characteristics of the layers are considered to affect separately each of the two criterion. The effect of lower temperature is accounted by increasing K1 (E1/E2) and thus the tensile radial stress. The effect of higher moisture is achieved by decreasing E3. Temperature and rainfall variations have a linear relationship with elevation in Arizona and these have been assumed to also have a linear variation with a regional factor which is comparable to AASHTO’s regional factor. It is assumed that the effect of AASHTO’s regional factor on its Structural Number is the same as our correction factor (C.F.) is to our K1.

At the present, no definite procedure has been proposed to follow or detect structural changes in a pavement ; however, the Dynaflect is being used to obtain deflection data. Recent analyses of Dynaflect deflections show lack of time-of-year effects on K1 values and suggest a need for modifying the measuring technique. A complete example of a pavement design is presented to illustrate the procedure.
04020 New Method for Asphalt Pavement Design Adopted in the USSR
M.B. Korsunsky, P.I. Telyaev

The report deals with the method for flexible pavement design recently accepted in the USSR.

This method allows to design a structure resistant, to both multiple applications of loads and of natural factors which is at the same time the most effective economically. The ability of asphalt pavement structure to resist well the load action is estimated by means of three criteria considering both the work of each layer and of the whole structure. The design is made for each layer and the criterion for evaluation of a layer is chosen depending on a degree of layer discreteness. The design of asphalt concrete pavements and other monolithic layers is performed on the basis of the criterion of "bending tensile strength" in order to keep them from fatigue crack formation. The design of subgrade soil and discrete layers is made according to the criterion of "shear resistance" to prevent them from permanent deformations. The whole structure is designed on the basis of elastic deflection in order to avoid its degradation under multiple moving loads. A method based on the above mentioned criteria is developed for designing the road pavements, which should have practically reversible (elastic) deformations under service loads. To determine stresses and strains in the road pavement layers and subgrade, the use is made of the solutions of the elasticity theory problem for layered semispace.

The ability of road pavement structure to withstand the frost and water deleterious effects is evaluated by using the criteria of "frost-resistance" and "drainage capability" of a structure. This value is of special importance for a subgrade of silty and clayey soils in the regions of excessive moistening and seasonal frost penetration. The road pavement structure is considered to be frost-resistant if actual expected heave of a subgrade soil is less or equal to the allowable value of pavement frost heaving.

An assessment of the expected heave is based on the regularities of the migration of water from water-table to the frost boundary of a soil. The main regularities are those characterizing the rate of frost penetration through pavement structure and the intensity of water inflow to the frost boundary. The drainage properties of a structure are considered to be satisfactory if the thickness of filtration layers is sufficient to provide:
– temporary absorption of the arriving water without decrease in strength of subgrade soil before the drainage facilities begin to work during the first stage of thawing when the pavement structure beneath the middle of a carriageway has thawed and the edges of a drainage course are yet frozen and the drainage facilities have not operated;
– appropriate drainage of the water from the base courses during the second stage.

Relationships for designing such drainage are based on the regularities of a steady regime of movement of free and capillary water flows in a draining layer of a pavement. To make the results of calculations closer to the data obtained the theoretical relationships are corrected for experiments.

The new method permits the following main factors determining the behavior of pavement structure in service to be considered: parameters of the equivalent vehicle, traffic density, soil types, climatic and hydrological conditions, subgrade construction, properties of the materials used, and the requirements for pavement performance. This method takes into account both deformative and strength characteristics of soils and materials. Design parameters are specified. But if necessary they may be obtained by direct sample testing under design moisture and temperature.

The report contains formulae for determination of stresses and strains arising in different layers as well as for establishing their admissible values. The experimental data which show the acceptability level of the formulae are given. The relationships for determination of actual and required frost-resistance as well as those for determination of drainage capability are presented, along with description of the basic principles of design automation and search for an optimum decision. Reliability of structures designed by the new method is demonstrated. An example of flexible pavement design under adverse soil and hydrological conditions is given.
04021 Development and Field Verification of a Mechanistic Structural Design System in Ohio
Kamran Majidzadeh, Leon 0. Talbert, Moses Karakouzian

This paper describes the development and field verification of a generalized mechanistic design system presently under implementation in Ohio. This Ohio pavement structural design system has been developed through the merger of various mechanistic subsystems developed in recent years.

Since the basis of any rational design approach is the stress distribution theory, an elastic stress distribution theory has been utilized in this system. The stress distribution theory, however, is presented in a generalized form incorporating geometrical and material non-linearities. The geometrical non-linearities include the effects of cracks, joints and discontinuities in the pavement system. As a result, this design system is capable of solving various boundary value problems associated with overlay design, joint stress analysis and fatigue crack propagation. The stress distribution in the vicinity of joints, cracks and other discontinuities can easily be calculated and their effects on other functional relationships of other subsystems can be determined.

The Ohio pavement structural design system is composed of two basic, independent subsystems, permanent deformation (rutting model) and fatigue models, as well as a subsystem describing the effect of environment (such as thermal fracture and thermally-induced failure) on the pavement structure. The environment subsystem considers the effects of environment on pavement material input variables and its expected performance due to thermally-induced stresses.

Due to the mechanistic nature of the Ohio pavement structural design system, it is applicable to design and analysis of new pavements, as well as determination of asphalt overlay thickness requirements for rigid and CRC pavements.
04022 A Thickness Design Procedure for Pavements with Cement Stabilized Bases and Thin Asphalt Surfacings
J.K. Mitchell, C.L. Monismith

A thickness design procedure applicable to a pavement system consisting of a cement-stabilized layer resting on a subgrade and protected by a thin asphalt surface layer is described. Stresses and deformations are estimated using layered elastic theory. The first step is to select a thickness adequate to prevent fatigue in the cement-stabilized layer. The second step is to insure that the combination of load and thermal stresses will not crack the stabilized layer.

The procedure accounts for the fact that cement-stabilized bases will crack shortly after construction due to shrinkage stresses. Base and subgrade stiffnesses can be determined either by laboratory tests or estimated by approximate procedures. Comparisons of thicknesses obtained by this procedure for highway type loading conditions with those by existing procedures show that the new method gives comparable values. Generally, the thicknesses are great enough to minimize initial pavement cracking, and may, therefore, be conservative for low numbers of repetitions of heavy loads.
04023 Trends in the Development of Flexible Pavement Design in Hungary
Ervin Nemesdy, Imre Keleti, Tibor Boromisza, Laszlo Gaspar

The authors describe the Hungarian practice of flexible pavement structure design and pavement strengthening. The catalogue of typical pavement structures is currently being compiled. Two main deficiencies are attributed to the semi-empirical design method used:
– The cheracterisation of the soil load bearing capacity by the CBR value is uncertain, and
– the equivalent number characterising the quality of asphalt materials is independent of temperature.

Hungarian research is in progress to simplify the computer calculation of these multi-layer systems. Work is also being done on a computer interpolation of the Jones-tables of a three-layer system. A simplified program has been written in order to reduce computer time and capacity, which calculates four pavement structure layers by the finite elements method. The soil layer is assumed to be en elastic semi-infinite continuum characterised by the Young’s modulus and the Poisson value. This computer program may also be run on smaller computers. The authors investigated the sensitivity of the design method to changes in the soil load bearing capacity in the cases of pavement structures designed for low, medium end heavy traffic. The elastic moduli of the asphalt courses have been determined at the Central European annual average air temperature (+15C) and the typical temperature (+5C) of the thaw period. On the grounds of the results obtained it can be stated that the role of the soil load bearing capacity in the design of pavement structures is smaller than hitherto assumed.

The splitting strength of asphalt materials is tested and the elastic modulus is determined at the same time. The magnitude of the splitting strength is considerably influenced by the loading speed and the temperature, so these are specified at 50 mm/min and +5C, respectively. The calculated modulus values may be used for approximate calculations of the pavement structure, they may furthermore be correlated with the characteristic values of the complex deformation moduli obtained by the fatigue tests.
04024 Thickness Design Procedure for Asphalt and Emulsified Asphalt Mixes
L.E. Santucci

A simplified thickness design procedure for pavement structures constructed with asphalt mixes, dense-graded emulsified asphalt mixes, or cement- modified emulsified asphalt mixes is described. Elastic layer principles are used in the development of the procedure. Two critical strains–the horizontal tensile strain at the bottom of the treated layer and the vertical compressive strain at the surface of the subgrade– are examined in determining the proper pavement thickness. Allowable values for tensile strain are based on laboratory fatigue tests, with consideration given to the slower crack propagation time encountered in the field. Vertical subgrade strain criteria developed from field observations of pavement behavior have been selected to minimize surface rutting caused by overstressing the subgrade.

Environmental considerations, in particular, the effect of temperature, the curing effect of emulsified asphalt mixes, and the effect of frost on subgrade strength, are considered in the design. The diametral resilient modulus device is used to define the stiffness characteristics of the treated mixes over the expected range of in-service temperature and curing conditions.

A relationship is proposed showing the important effect of the volume of air voids and asphalt in a treated mix on its fatigue behavior and, hence, the thickness requirements for the pavement structure. In the case of emulsified asphalt mixes, this effect can have a much greater influence on the design thickness than the lower strength condition of the mix during its early cure.

The paper contains a step-by-step outline of the design procedure for the practicing engineer to follow in determining pavement thickness requirements. A design example is included with the necessary work sheets to show the proper use of design charts, graphs, and equations. The procedure does not require the use of computers. However, simple computer programs written in BASIC have been developed to eliminate the tedious hand calculations described in the paper. These programs cover the design of pavements made with asphalt, emulsified asphalt, or cement-modified emulsified asphalt mixes.

The tensile strain or fatigue criteria used in this procedure predicted well the 10-20% cracking observed on the San Diego County Experimental Base Project after seven years’ traffic. Support for the vertical subgrade strain or rutting criteria is shown by an analysis of performance of mixes examined in a circular test track at our laboratory.
04025 Moving Loads on a Viscoelestic Double Layer: Prediction of Recoverable and Permanent Deformations
G. Battiato, G. Ronca, C. Verga

In this paper a two layered viscoelastic incompressible system subjected to a sequence of moving loads is considered. Materials characteristics are described by a simple analytical expression for the Creep Compliance function, as verified by laboratory creep experiments carried out on a large class of asphalt concretes. The calculation of the viscoelastic deformations (vertical, transverse and longitudinal strains) and vertical surface displacements is performed with reference to three orders of effects:
1) the passage of a single load
2) the accumulation of the recoverable viscoelastic deformation due to a sequence of moving loads; this effect arises from the viscoelastic delay in deformation recovery
3) the calculation of the permanent deformations due to each single load passage when the existence of a Maxwell model in series is assumed in the rheological model defining the characteristics of the first layer.

Interface conditions can be varied choosing rigidly bounded layers or frictionless slip.

With reference to point 1) our viscoelastic calculations show that the prediction of the deformations can be obtained with sufficient accuracy by means of elastic methods in general for the first layer only. If the response of the second (subgrade) layer is not completely elastic, the elastic methods underestimate the deformability of the lower layer by a factor that increases with the depth and is more relevant for soft subgrade materials.

At a distance from the surface of the road, that corresponds to 4-5 times the radius amplitude, this factor may be of the order of 2. If one is interested in calculating the vertical displacements under the load, the predictions of the elastic methods are inaccurate even for the points belonging to the surface of the road.

The lack of symmetry of the deformations with respect to time reversal is observed. The slow viscoelastic recovery of the vertical and transverse deformation for times which follow the load passage makes it possible the accumulation of the deformations with repeated moving loads.

With reference to point 2) we show that the contribution of the residual viscoelastic deformations can be of the same order of magnitude as the contribution of the single load passage in some critical but realistic traffic conditions. Furthermore, when the subgrade behavior is nearly elastic, the residual viscoelastic deformation accumulated in the first layer is independent of the subgrade stiffness.

A considerable accumulation of the deformation is observed for the vertical and transverse deformation, not for the longitudinal deformation. With reference to point 3) our results show that if the subgrade material has an ultimate elastic behavior, the passage of each single lead produces a small permanent deformation which concerns the asphalt concrete layer only and is independent of the subgrade stiffness.
04026 The Validity of Design Procedures for the Permanent Deformation of Asphalt Pavements
S.F. Brown, C.A. Bell

The approach to predicting permanent deformation in asphalt pavements suggested by Romain and Barksdale is examined in detail with particular reference to its validity under controlled laboratory conditions as a preamble to applying the ideas in practice. Attention is given to the correct representation of stress by use of invariants and the careful planning of laboratory materials tests following a review of the relevant literature. Non-linear elastic analysis is required to estimate stresses under the high temperature conditions relevant to the permanent deformation problem. The results from pavement experiments including suitable instrumentation are compared with the predictions of permanent deformation carried out by various authors. A firm conclusion about the validity of the Romain/Barksdale approach is not considered possible on the evidence presented but the need for further work is indicated.
04027 Permanent Deformation Law of Bituminous Road Mixes in Repeated Triaxial Compression
L. Francken

The experimental work carried out at the Belgian Road Research Center in the field of permanent deformations of bituminous road mixes has for primary goal the determination of a phenomenological deformation law which is to be fed into a structural design method directed towards the limitation of rutting.

The experiments were carried out by means of an electro-hydraulic testing machine. Cylindrical samples of 5 bituminous mixes were submitted to dynamic triaxial tests at different temperatures, frequencies and stress conditions ; in these tests the vertical stress is a sinusoidal function of time and the lateral stress is a static one. The results obtained have been interpreted by considering two important mechanical characteristics : the dynamic stiffness modulus |E*| and the creep curve.

1) The stiffness modulus |E*| is a decreasing function of time (or number of cycles); its value measured after 10^3 seconds obeys the frequency-temperature interchangeability principle. The master curves obtained are similar to those determined in bending, and they may be predicted in many cases from the mix composition. The dynamic stiffness modulus is strongly dependent on the amplitude (sigma sub 1) of the dynamic component of the vertical stress; this nonlinear behavior may not be overlooked for vertical stresses exceeding .3 MN/m2.

2) The creep curves observed may be described by means of a time dependent function of the form

epsilon sub P (t) = At^B + C (exp Dt-1)

in which A, B, C, D are parameters to be fitted (by computer) to the experimental points.

In order to establish a general law of permanent deformation the influence of the physical conditions characterizing the test on the value of these 4 parameters has been investigated.

This leads so far to the following conclusions:

– A purely parabolic behavior (C = 0) may be observed if the highest value of the shear stress corresponds to an experimental point which is below a typical intrinsic curve. This condition may be used as a design criterion against plastic failure.

– The parameter B seems to be less dependent on the physical conditions and ranges from .1 to .3. A mean value close to .25 is a realistic value. A method allowing the evaluation of the creep curve of bituminous mixes is presented for two cases :

1/ When the vertical stress oscillates between zero and a maximum value, and provided that C = 0 (plastic failure criterion) the creep curve takes the form

epsilon sub P (t) = K [(sigma sub 1 – sigma sub 3) / |E*| ] t^B

where sigma sub 3 is the confining pressure, K an empirical constant and t the time (corresponding to the quotient of the number of cycles by the frequency of the vertical stress).

2/ An expression based on the energy dissipation concept is proposed in order to predict the permanent deformation under more general stress conditions, This more general formulation reduces to the former one when taking into account the dependence of the dynamic stiffness modulus on the amplitude sigma sub 1.
04028 Evaluation of Rutting Due to Viscous Flow in Asphalt Pavements
S. Huschek

This paper is to be considered as an attempt to develop a generally applicable method for calculating the permanent deformation of bituminous layers due to viscous flow. Using a computer program based on linear, elastic multi-layer theory, the stress/strain distribution in a road under a wheel load can be calculated. The static stress/strain distribution is transformed into a "moving distribution" with respect to speed v = x/t. So as a result, stress and strain at every given point becomes a function of time.

This permanent deformation in one layer caused by the passage of one wheel is calculated assuming viscoelastic properties characterized by a Maxwell element. The total rut depth is obtained by summing up the irreversible deformations of all viscoelastic layers. It is important to consider the viscosity eta of the dashpot as a time dependent value increasing with the number of loadings. Thus, the phenomenon of "consolidation" taking place in asphaltic layers under traffic can be described. The material characteristics are derived from an unconfined compression creep-test which is described in this paper.

The following conditions which influence rutting are taken into account: temperature distribution, number and amount of wheel loadings, transverse distribution of wheel passages, and speed of vehicles. The proposed method can be used to evaluate the mechanical properties of bituminous mixtures required to resist traffic forces under different climatic conditions. Though the presented method is to be considered as a practical engineering approach, it is useful to evaluate the relative performance of different bituminous mixtures and different structural configurations. The validity of this method is shown by comparing the results with measurements made on a Swiss test-road since 1972. The effect of consolidation is shown by comparing the mechanical properties of field core specimens from bituminous layers from the normal lane and from the passing lane.
04029 A Computer Based Subsystem for the Prediction of Pavement Deformation
R.W. Kirwan, M.N. Snaith, T.E. Glynn

This paper describes a subsystem capable of predicting the lateral profile of pavements after any number of axle loadings. The method employs a sophisticated computer program, together with a high level of materials characterisation. The paper is divided into three parts.

(a) A description of the research program DEFPAV is given. This program is capable of computing the load induced stresses and elastic strains beneath a moving wheel load. Furthermore it produces at the same time an estimate of the surface profile due to permanent deformation in each of the constituent layers of the pavement structure. A modification of the program which is suitable for use by road engineers is discussed.

(b) A description of the methods used to obtain both the elastic and creep properties of pavement materials is given. The methods discussed are designed to be usable by the average well equipped materials laboratory rather than a research laboratory.

(c) A description is given of the verification procedures undertaken by this research group to demonstrate that DEFPAV may be used with some confidence in multilayer pavement problems. Furthermore a computation is presented which demonstrates that a program such as DEFPAV may be used to estimate load equivalence factors for different environments and loading conditions.
04030 A Working Design Subsystem for Permanent Deformation in Asphalt Pavements
Frank R.P. Meyer, Ralph C.G. Haas

Rutting represents one of the most common forms of distress in flexible pavements throughout the world. In order to rationally design pavements, in terms of economics, safety, serviceability to the user, and structural damage, it is necessary to be able to predict the actual magnitude of rutting under any given set of conditions.

This paper describes a method of predicting rut depths in flexible pavements. It is based on the evaluation of permanent deformation characteristics of asphalt concrete and unbound base course materials. Statistically designed laboratory experiments provided information about the behaviour of these materials under simulated field conditions. From these laboratory results, mathematical models were developed to predict rutting. Verification was conducted through correlations with actual field test measurements of rut depths.

Asphalt concrete was initially investigated and predictive models were developed in terms of load applications, stress state, temperature, air voids and asphalt penetration as variables. These models were verified using actual measured rut depths from full depth sections at The Brampton Test Road in Ontario. A high degree of correlation was achieved.

The next phase of the study examined unbound base materials. Similar predictive models were developed, in terms of density, stress state and load application as variables. These laboratory based models were applied to the conventional flexible pavement sections at Brampton, again with a good degree of reliability.

Finally, after a sensitivity analysis of variables expected to be significant in affecting rutting, a simplified regression equation incorporating only the important variables was developed. Granular and asphalt equivalencies were used and good agreement between predicted and measured values were found for sections at The Brampton Test Road, Ste. Anne Test Road and San Diego Test Road.
04031 A Subsystem to Predict Rutting in Asphalt Concrete Pavement Structures
C.L. Monismith, K. Inkabi, C.R. Freeme, D.B. McLean

A design subsystem is presented to estimate the amount of permanent deformation (rutting) resulting from repeated traffic loading. Relationships between applied stress and permanent strain defined by repeated load triaxial compression tests are presented for fine-grained soils, granular materials, and asphalt concrete. Stresses resulting from wheel loads are estimated assuming pavements to be represented as layered elastic structures. The stresses, in turn, permit estimation of permanent deformation in each layer of a specific pavement by

(1) Computing the permanent strain at a number of points within the layer, the number being sufficient to define the strain variations with depth.

(2) Estimating the deformation by summing the products of the average permanent strains and the corresponding differences in depths between the locations at which the strains were determined.

Total rut depth is estimated by summing the contributions from each layer.

To illustrate the potential applicability of the procedure, comparison of the amount of rutting with that observed in an in-service pavement is presented. In addition, to illustrate how the method can be used in practice, a number of examples are included, representative of both unsurfaced (low volume) and surfaced roads.
04032 Methods of Predicting Deformation in Road Pavements
E.N. Thrower

Rigorous computations of the permanent deformation in pavements by methods which satisfy all the necessary equations of equilibrium, compatibility and material behaviour are difficult and extremely time-consuming, so that-simpler approximate methods are normally used. The implications of one such approach, which is in common use, are discussed. The method is based on deriving the permanent deformation from experimentally-determined material deformation properties, in association with a stress distribution determined separately ignoring the permanent deformation.

The further simplification is often made that in the computations, a moving load can be replaced by a pulse load, but the results of computations made by this separative method for pulse and moving loads, using both linear and various non-linear deformation laws, lead to the conclusions that this approximation can result in serious distortions of the assessment of the relative contribution to the total surface displacement from the individual layers.

Because of several doubts and difficulties associated with the use of separative approximations, the permanent deformation in a linear visco-elastic model of a pavement is discussed; this is a rigorous analogue of the linear cases treated by the separative approximation. It is shown that the permanent deformation in such a structure depends only on the deformational properties of the layers, and not at all on their elastic properties; this is in total conflict with the separative approach. Further, the permanent deformation due to a pulse load depends on the deformational properties of the layers in precisely the same way as the deflection of an elastic structure depends on the elastic properties of the layers, and can be computed using the same programmes. Comparison of the results obtained for pulse loads by this method and by the separative approach shows considerable discrepancies, both in the total displacement and in the relative contributions from the respective layers. and it is concluded that some care is needed in the use of the separative approximation.

The behaviour under a moving load can be obtained from that under a pulse load by superposition, but it is shown that the integrals involved do not converge in the linear case. It is concluded that to obtain physically reasonable results by either the separative approximation or a visco-elastic approach requires a knowledge of the deformation properties at low as well as at high stresses.
04033 Design of Asphalt Overlays for Pavements
J. Bonnot, P. Autret, A. De Boissoudy

An important program of systematic overlay of the national primary road network has begun in France in 1969. A pavement overlay design method has been established, and is described in this paper. The most important – and most difficult – phase of overlay design is the characterization of the old pavement: the principles of the method used are summarized.

The paper first describes the method used to confirm the design of a pavement structure incorporating an old pavement and an overlay. A multilayer elastic model is used; an equivalent traffic and an equivalent temperature are used to take into account the variations of axle loads and pavement temperature. The method attempts to take into consideration the probabilistic nature of pavement distress, resulting from the scattering of pavement thickness, overlay material composition, fatigue life, and bearing capacity of old pavement. Distress risk has been chosen as a function of traffic. The method attempts also to take into account in someway the effect of future maintenance on the overlaid pavement.

The overlay design method has the form of a Catalogue of overlay structures in which overlays are calculated once for all. For the overlay of old flexible pavements, the various old pavements have been classified in 36 cases, according to deflection and thickness of old pavement. The overlay thickness has been calculated for each of these cases and for several traffic classes. It has been found that for overlays with cement or slag treated base, the overlay thickness depends only on traffic and deflection of the old pavement; and that for overlays with bituminous base, the overlay thickness depends on traffic, deflection on the old pavement, and on thickness of asphaltic concrete in the old pavement.

The design method has been confirmed using data on the in situ behavior of thirty actual pavement overlays, which has been observed for 4 to 9 years; these data allow a comparison between theoretical and observed relations between strain (or stresses) in overlays and overlays life.

A method is also given for the design of asphalt overlay of old pavement with cement treated (or slag treated) base. Two cases must be considered : the cement treated base shows fatigue cracks but has kept its cohesion, or has lost its cohesion. For old pavements of that type, overlay thickness are always high.
04034 Design of Asphalt Concrete Overlays Using Layer Theory
Harvey J. Treybig, B.F. McCullough, Fred N. Finn, Richard McComb, W. Ronald Hudson

This report is a user’s manual for thickness design of flexible overlays for flexible pavements. The design procedure is limited primarily to fatigue cracking and rutting criteria. Three cases of existing pavement condition are recognized by the procedure. These subsystems are:
1) existing pavement with remaining life,
2) existing pavement mildly cracked, and
3) existing pavement severely cracked.

Each requires input from the following areas:
1) deflection testing,
2) condition surveys,
3) traffic data, and
4) materials characterization.

The deflection testing serves as an aid in establishing "design sections" and in characterizing the subgrade. The condition surveys are used to select the proper design subsystem. Traffic data must be in the form of 18-kip equivalent axle loads. The materials characterization consists of laboratory testing to determine modulus values for each material.

The overlay thickness design involves the use of inputs from the above four areas along with an elastic layered theory computer program, a fatigue equation and a rutting equation to determine a thickness that satisfies both the fatigue and rutting criteria. A complete example problem solution for the remaining life subsystem is presented.
04035 A System for the Prediction of Pavement Life and Design of Pavement Strengthening
N.W. Lister, C.K. Kennedy

Consideration about twenty years ago of possible approaches to forecasting the life of existing pavements and to designing overlays using, as far as possible, simple in-situ measurements of pavement strength led to the development of a systematic programme of deflection-testing on the full scale road experiments already in service, or being built by the Road Research Laboratory. Lack of knowledge about the real stress-strain behaviour of roads dictated the investigation of possible direct correlation between deflection and the deterioration of roads, deterioration characterized primarily by the development of rutting in the wheel paths.

The experimental programme designed to relate deflection measured by the Benkelman Deflection Beam to pavement condition and traffic is briefly described. The preliminary information presented at the last Conference has been greatly extended and confirm strongly defined relationships between deflection measured early in the life of the road and the onset of critical conditions defining the need to strengthen. When combined with further similar relations shown to exist between critical deflections and the traffic carried, design charts for predicting unexpired pavement life can be drawn up: these are presented for the main types of pavement.

The reduction of deflection brought about by overlaying has been quantified and pavements already overlaid have been studied in order to establish deflection performance relations for strengthened pavements. From the information design charts have been developed defining overlay thicknesses required to extend the life of a road to carry any given traffic. Considerable experience in overlay design has led to great emphasis being placed on the need to define, by closely spaced measurement, the considerable variability of strength which normally exists on a road in need of overlaying. A system for predicting pavement life and designing overlays from this type of information obtained on the road by surveys using the Deflectograph and the Benkelman Deflection Beam is briefly described, with reference to other published work.

The latest development in the design procedure is a computer programme capable of treating the variability of deflection in a consistent manner in such a way as to eliminate the risks of localised early failure or over design in the strengthened pavement. Assessment of successive 100 m lengths of road on a moving average principle is used to define minimum cost solutions which are then used by the engineer as the basis for designs which take into account engineering constraints.

Deflection surveys carried out on a number of normal roads having a range of base types are compared with the predictions of the design charts presented. Examples are given of both new construction and strengthened pavements; one example where the design life of an overlay has now expired is considered in more detail.
04036 Pavement Evaluation and Overlay Design – The Shell Method
A.I.M. Claessen, R. Ditmarsch

A method is presented for evaluating pavements by deriving the structural properties from the shape of the deflection bowl under a test load. The structural properties of the existing pavement are expressed in terms of effective layer thicknesses and Young’s moduli of the materials. The shape of the deflection bowl is characterised by the ratio of the deflection at a given distance from the load to that measured under the centre of the test load. Charts prepared from computations with the BISAR computer program for multi-layer elastic systems are used for the interpretation of the deflection measurements.

The method can be used, in principle, in conjunction with any type of test load, provided that the appropriate loading conditions are taken into account. Preferably, the measurements are conducted with the Falling Weight Deflectometer (FWD) since its characteristics of force level and loading time are more representative of heavy traffic than most other systems, which are affected adversely by the configuration of the loading and recording systems and/or by indirect measurement of deflections.

Preliminary studies demonstrated that satisfactory results were obtained with a prototype FWD. The data in this paper were obtained with a new FWD modified to allow quick operation with remote control for routine pavement evaluation. Special attention is given to routine operation, including data processing, and to the interpretation of measurements on more complicated structures such as those with cementitious base layers. From the structural properties of the pavement, derived from the deflection bowl measurements and the number of axle loadings carried during service, the residual life of the pavement, expressed in terms of the number of standard axle load applications, can be determined. This information is used to estimate the overlay thickness required for future estimated traffic, using the newly developed design method, in which influences of climate and asphalt mix type can be taken into account. Practical examples of the evaluation and overlay design method are given.
04037 A Practical Approach to Flexible Pavement Evaluation and Rehabilitation
A.C. Bhajandas, G. Cumberledge, G.L. Hoffman, J.G. Hopkins III

A practical scheme for evaluating and rehabilitating flexible pavements is presented. It considers safety, riding quality, and strength as primary factors for evaluating the condition of a roadway. Criteria for safety are not discussed.

Riding quality standards are expressed through the Present Serviceability Index, PSI. Terminal Serviceability Indices, TSI’s are established for Interstate, Principal, Minor Arterial, Collector and Local Access Highways from PSI distributions obtained on 5,492 miles of flexible pavement. These TSI’s are shown to be flexible and dependent upon funding and construction capability. Thus flexible failure criteria for each type of highway are established.

Spring deflections measured through the dynamic deflection measuring device Road Rater are used to assess pavement strength. Correction curve for surface temperature and adjustment factors for season are established to facilitate comparison under standard conditions.

Permissible deflections on crushed stone bases for various failure criteria are drawn from the AASHO Road Test data. Permissible deflections on bituminous concrete bases are established from a study of ten sections of a test track facility located at State College, PA.

The deflection-overlay relationship drawn from data of ten sections in south central Pennsylvania serves as the basis for overlay design. These overlays show no significant distress upon five years of service. The simplicity of the proposed scheme with a small additional cost for testing, evaluation and design make it feasible for implementation by State Highway Agencies.

The proposed scheme ensures that all type’s of highway are given equal consideration and that the worst mileage of each highway type is resurfaced or rehabilitated.
04038 Evaluation and Overlay Design for Flexible Pavements on Low Volume Roads
Pieter De Kiewit, Peter C. Koning, R. Frank Carmichael, W.R. Hudson

The evaluation and overlay design of low volume pavements is a major expense for many governing bodies such as small cities and counties. All over the world such governmental agencies have many miles of flexible pavements to maintain; usually with only small staffs to handle the job. The problem is often complicated further because the agency does not have adequate funding. It is also becoming clear that in the future the maintenance of road pavements will claim a greater part of the funds that are available for highway and road development. Because of these and other factors a rational pavement evaluation and overlay methodology is needed to assist the engineer and preserve the existing investment.

To fulfill these needs, an evaluation and overlay design procedure has been developed based on dynamic deflection measurement, elastic layered theory, and behavioral models for fatigue and rutting (Ref 1). The most important aspects of this procedure are that it is 1) currently operational, 2) based on the best theory available, and 3) developed for easy use. This paper describes the procedure and illustrates it with an example problem applied to conditions in Holland. The criteria according to which maintenance operations are carried out are divided into three main groups:
1. Traffic criteria (geometry),
2. Structural characteristics, and
3. Safety and comfort criteria.

A total evaluation procedure must consider all of these elements. For practical application we have chosen a method which emphasizes these aspects, and recognizes the more relevant items for the type of road which is to be evaluated. In the case of the low volume roads, the structural element and the consequences of different maintenance strategies is important and this will be a main part of the paper presentation. Consultation with the road owner and his needs will establish the requirements for the consideration of the traffic, safety, and comfort aspects.
04039 Optimal Design of Asphalt Overlays
M.D. Rafroiu

It is often economical and convenient to improve the structural condition of a pavement by means of asphalt overlays. Many different methods can be used for the design of these overlays. Only part of them are fitted for an optimal design. New methods based on a structural design approach have been proposed in the last ten years. Generally, the present design methods are based upon deflection measurements and limiting deflection criteria. The difficulties of this approach are: the measurements of the equivalent modulus of the existing pavement, the evaluation of the elastic modulus of the overlay, the correlation between constructive requirements and theoretical accuracy (which may not agree perfectly), the maintenance and overlay policy and other factors as well.

This paper presents a new method for designing asphalt overlays when optimization is required at least from an economic standpoint. Two main situations are taken into account : the existing pavement needs only to be strengthened as no important distortions are reported, the existing pavement is heavily rutted, cracked or disintegrated and it is supposed to undergo particularly deep investigation and, if necessary, a new and complete structural design would be done. Only the first situation is studied in this paper.

The new method is based upon an energy approach which requires that the energy absorbed by the highway body, when deflection under exterior loading occurs, should not exceed a limited amount which is related to the composition and the value of the traffic the pavement is supposed to bear through the design period.

The design criteria are based upon limiting tensile stresses at the bottom of the layers, shear stresses in the center plane of the layers and, most of all, requiring a given bearing capacity of the "rejuvenated" pavement. Since only one layer (as overlay) may not fulfill all requirements derived from the design criteria, it has been assumed that in certain situations two layers would be necessary, such a solution being supposed to offer both economic and constructive advantages.

The method also permits an iterative procedure in order to get an economic optimum.

Since new developments have been made, both theoretical and experimental accuracy checks have been done.

The method has been accepted for current overlay design and is now in use in parallel with the existing standardized method.
04040 A Pavement Analysis and Structural Design Procedure Based on Deflection
P.G. Rufford

Recognition of flexure cracking as a primary mode of distress of asphalt pavements in New South Wales, Australia, highlighted the need for a structural design procedure based on deflection. As the Benkelman beam was considered the most suitable equipment for measuring deflection for general design purposes, it was felt necessary in the first instance, to develop a pavement analysis technique to interpret the Benkelman beam measurement of deflection. This pavement analysis technique involved analysing the Benkelman beam deflection bowl using elastic theory to assess the flexibility of the subgrade and the overall flexural strength of the pavement. The elastic theory analysis was tested both theoretically and empirically to define the limits of its application and to provide the confidence required to use it in practice. In-situ non-destructive testing has particular application to reconstruction and overlay design and therefore the analysis was developed into a design procedure for this purpose. As the overall flexural strength of the pavement was expressed in terms of an equivalent thickness of a standard base material, the design procedure used the basic AASHTO pavement model. It involved designing the pavement to a design Benkelman beam deflection which was determined from the performance of the existing pavement in relation to the TRRL design deflection levels. This paper presents the pavement analysis technique and the structural procedure for reconstruction and overlay design. It also includes an example of an overlay design to demonstrate this procedure.
04041 Overlay and Stage-by-Stage Design
P. Ullidtz

A structural design procedure based on elastic theory is presented. The main aspects of the procedure are discussed, i.e. failure conditions, loading and climatical conditions, determination of elastic parameters and of allowable stresses and strains and finally calculation of the critical stresses and strains.

The main emphasis of the paper is on the determination of the elastic moduli of the different materials in a pavement structure, because a correct determination of the moduli is a prerequisite for the application of the theory of elasticity. A non-destructive method of determining the moduli of two- and three-layer systems is presented. The method is based on the use of dynamic plate loading tests (e.g. falling weight deflectometer tests) and the calculations may be carried out with the use of some diagrams and a programmable pocket calculator. For more complex structures, i.e. structures containing four or more layers, use of at least a programmable desktop calculator is required.

The moduli are calculated from the deflection basin, and information on the subgrade modulus may be obtained to an equivalent depth of 10 -12 meters. The method is therefore especially well suited for determining the modulus of the subgrade before construction of the road. The moduli may be determined both for structures containing linear elastic materials only and for structures having a non-linear elastic subgrade, for which the stress dependent modulus may be approximated by the following simplified relationship:

E = C x (sigma sub 1 / sigma prime)^n

where E is the modulus, sigma sub 1 the major principal stress, sigma prime a reference stress and C and n are constants.

With this relationship a modified version of the method of equivalent thicknesses may be used and some simple equations for calculating the critical stresses and strains in the centerline of the load are presented.

The validity of this model for determining the stresses and strains under actual traffic loadings has been verified through full scale experiments and the determination of the allowable stresses and strains is based on laboratory tests and the results of the WASHO and AASHO Road Tests, modified in accordance with the experience gained from Danish road design during the last decade.
04042 Structural Design of Asphalt Pavements and Control of Environmental Influence on Performance
William S. Housel

This paper reviews the development of field loading tests to measure the bearing capacity of soil masses and pavements supported by soil masses. Time, load, settlement or deflection, and the size of the bearing area are the four basic quantities which must be measured to determine load bearing capacity. The relationship between these variables has been clearly and accurately established by more than forty years of research and practice by various agencies.

The time, expense, and inherent difficulty of running full scale field loading tests over a wide area of variable soils involved in the design and construction of highway pavements places them beyond the resources of the majority of practicing engineers. Consequently it has been a major objective of the International Conferences on the Structural Design of Asphalt Pavements to develop laboratory tests to measure the fundamental properties of the materials used in the pavement structure and formulate design procedures based on these tests.

This primary objective and the corollary requirement of field verification have been achieved by an exhaustive study of data from the Hybla Valley project where both field loading tests on the pavement structure and laboratory tests of the materials in each component of the pavement were available. In every case including subgrade, base course, and the bituminous surface, the stress reactions and soil resistance coefficients measured by the field loading tests could be computed from the laboratory tests within an acceptable range of accuracy for the materials involved and the normal experimental error which would be expected in comparing field and laboratory tests.

The results of this correlation using the design procedures proposed in this paper are summarized in a tabulation. These design procedures are within the capabilities of any well informed and competent practicing engineer in the field of highway engineering and do not require the background of a research specialist.

Aside from routine laboratory tests for classifying the materials, the controlling laboratory tests to measure the fundamental properties required in the design are the triaxial compression and direct shear tests. These were the tests used in the early investigation of granular materials by Berry which are summarized in Table 3. These data demonstrated that internal stability of dense granular soils and highly consolidated granular materials was sufficient to control the design of flexible asphalt pavements. Internal stability was defined as that mechanical property of granular masses which produces resistance to displacement by mutual support of adjacent particles too large to be measurably affected by molecular forces.

Computations based on these data showed that the mechanical advantage of aggregate interlock produces resistance to displacement which in Berry’s tests, ranges from 1 to 17 times that obtained from theoretical equations. The key to the surprisingly high internal stability is measured by the "Aggregate Interlock Factor K," defined as an exclusively peculiar structural property of granular material unaffected by the theoretical capacity computed from traditional concepts of elastic behavior.

The test results under discussion, available since 1936, provide a body of experimental evidence to be evaluated which cannot be passed over without careful study. For many years internal stability based on these data has been successfully applied in design of caissons and high capacity bearing piles supported on highly consolidated granular soils and granular cohesive mixtures. These practical applications have demonstrated that these materials will develop internal stability up to the crushing strength of the grains without lateral displacement. It should be recognized that the vital role of such granular materials and mixtures becomes a controlling factor in design of flexible asphalt pavements.
04043 Pavement Management Design Considerations for Canadian Airports
G.H. Argue, B.B. Denyes, G.Y. Sebastyan

A synopsis is given of Transport Canada structural design practices for the provision of airfield asphalt pavements at Canadian airports. Pavement management design considerations in the related areas of construction, evaluation, operations, maintenance and rehabilitation programming are also discussed in a general manner. A short summary of Transport Canada research and development objectives as related to asphalt pavement structural design are outlined. A detailed treatment of standards and practices may be found in the Transport Canada manuals listed in the references.
04044 Pavement Management Guide: A Summary
Pavement Management Committee, Roads & Transportation Association of Canada

The Pavement Management Guide summarized in this paper is concerned with the practices and procedures illustrated within.

The Guide suggests that the basic purpose of a pavement management system is to achieve the best value possible for available public funds and to provide safe, comfortable and economic transportation. This is accomplished by comparing investment alternatives at both the network and project levels, coordinating design, construction and maintenance activities, and making efficient use of existing practices and knowledge. This paper provides a very brief summary of the Guide. It discusses benefits and implementation guidelines associated with pavement management, network investment programming, optimization of project investment, pavement evaluation, structural design of flexible and rigid pavements, construction, maintenance, data banks, and guidelines for research management.
04045 The Consideration of Frost in the Design of Asphalt Pavements
L. Caniard, C. Peyronne

The purpose of the article is to describe a new method for the consideration of frost in the design of asphalt pavements. The originality of the method developed by the Administration Francaise des Ponts et Chaussees resides in the fact that the harmfullness of frost is to be characterized by a frost index transmitted to the soil, and not by a depth of frost. Finally, a numerical model which is more precise than other models currently in use is used, which makes it possible to treat the case of structure containing several courses. The initial parameters involved in the method are:
– the thermal characteristics of the winter, and the manner of synthesizing them through the choice of a standard frost index,
– the thermal characteristics of the asphaltic materials making up the pavement. This article indicates how they have been determined,
– the mechanical properties of the asphaltic materials, and the influence of temperature,
– the heaving of the soil by frost, and its bearing capacity at thaw.

A numerical model, whose equations are elucidated, allows the calculation of temperature changes in the pavement, and forecasts the final position of the frost front, as well as the frost index transmitted to the foundation soil. The article presents some computation diagrams.

According to the result of the temperature calculation, three cases are possible:

– The thermic protection ensured by the asphalt pavement is sufficient, so that there might be no fear of any loss of bearing capacity of the underlying soil; the structure is to be retained.

– The thermal protection is inadequate, and there is a loss of soil bearing capacity. The consequences to the useful life of the asphalt pavement are examined. If the mechanical resistance is suitable, the structure is retained.

– In the case to the contrary, the method indicates how it is possible to either increase the thickness of the pavement (which improves thermal protection and mechanical resistance), or to make the soil less susceptible to frost. The article gives some examples of the practical application of the method, and describes verifications carried out in an experimental test station.
04046 Determination of Pavement Layer Moduli from Surface Deflection Data for Pavement Performance Evaluation
Lynne H. Irwin

In order to utilize mechanistically based methods for pavement design and evaluation, there is a need to know the moduli of elasticity of each layer in the pavement system. For pavement evaluation and overlay design it is preferable to know the in situ moduli, rather than resorting to elaborate and costly laboratory approaches to materials characterization.

Techniques for measuring moduli in situ have primarily been based on two techniques: surface vibratory testing and surface deflection testing. This paper reports on a method whereby surface deflection data may be evaluated using layered elastic theory to estimate the elastic moduli of each pavement layer. Moduli determined by this method are known within an accuracy of ten percent or better. Accuracies of this degree compare favorably with the point to point variability of in situ moduli due to construction and materials variation, and thus may be considered satisfactory.

The method of obtaining the layer moduli reported in this paper relies on the use of layered elastic theory computer programs. It is necessary that the thickness and Poisson’s ratio be known for each layer, although the results are not highly sensitive to the value assumed for Poisson’s ratio.

In the method, points on a two-dimensional surface deflection basin are fitted to the field data. Iteration is required to align the measured and computed points by adjusting the assumed values for the layer moduli. Presently the method relies on trial-and-error iteration using the BISTRO computer program. It is suggested that a new program be developed which would enable a more direct determination of the moduli.

A case study is presented to demonstrate the application of the method to the evaluation of a test pavement for a heavy duty haul road to support 400,000-pound gross vehicle weight lignite coal trucks. Deflections of the test pavements were obtained using the Dynaflect apparatus and a fully loaded haul vehicle. The major difference in the magnitude of surface loading served to illustrate that the assumption of linear elasticity is not valid over such a range.

The evaluation of one of the eight test pavements is illustrated by example, and it is predicted that the pavement will not be satisfactory for the anticipated loading and traffic. The prediction was borne out after several years of use.
04047 OPAC Design System
Ramesh Kher, W.A. Phang

OPAC (Ontario Pavement Analysis of Costs) is a computerized system that compares the performance and cost of hundreds of design alternatives for flexible pavements within just a few hours. Using the system, pavement design engineers can select the most effective pavement design that has the least cost.

OPAC predicts the life of a pavement. The deterioration in Riding Comfort Index (RCI) has been made a function of repeated traffic loading and annual cyclic environmental changes. The deteriorations caused by the two factors have been respectively modelled using data from the AASHTO (Illinois) Road Test and Brampton (Ontario) Road Test. Subqrade surface deflection under a standard wheel load has been used as a predictor of future behavior of a pavement. Successful thickness designs in the province were analysed using Elastic Laver Theory and several iterations of such analyses resulted in a set of modulus values which were thereafter assigned to various paving materials and six different categories of subgrades in the province. Through AASHTO Road Test data, an excellent correlation between subgrade surface deflection, traffic, and pavement performance has been developed to allow calculation of that component of pavement deterioration which is caused by traffic. The other component, caused by environment, has been modelled from the Brampton Road Test data. It is based on the difference in deterioration predicted due to traffic, as previously indicated, and actual measured deterioration of Brampton Road Test sections over a period of eight years. The two submodels for pavement deterioration are combined to predict the performance of any pavement section,

OPAC predicts pavement costs throughout the life of a pavement. Various cost components are: initial capital expenditure, subsequent resurfacing and maintenance expenditures and salvage return. OPAC is so comprehensive that it also includes road user delay cost in its economic analysis of design alternatives. OPAC provides an evaluation of the various cost components of a pavement on the one hand and various possible consequent costs to the users on the other and makes it possible to examine design trade-offs, The final decision regarding the selection of a design remains with the pavement design engineer who must also consider such location information as construction problems, aggregate depletion and traffic safety.

OPAC is very accessible, simple to use. Communication between the pavement design engineer and OPAC is achieved through a terminal which is linked by telephone to an IBM 360 computer. A question and answer dialogue is established between the computer and the engineer using the keyboard of the terminal. In response to questions posed by the computer, the pavement design engineer enters basic design specifications such as subgrade condition traffic projections,
, performance requirements,
available material and their costs. Within seconds the computer returns its analysis to the terminal , printing out various design alternatives to meet the engineer’s specifications.

In operation since March 1974, OPAC’s ability to analyze and predict cost and performance provides the basis for effective pavement management, and for assisting planners in providing, within the available budget, an economically designed road system that considers the needs of the province’s travelling public.
04048 Prediction of Sulphur-Asphalt Pavement Performance with VESYS IIM
Robert L. Lytton, Donald Saylak, Daniel E. Pickett

The utilization of sulphur in sulphur-asphalt pavements is receiving considerable attention for a variety of reasons. One reason is that sulphur is one material which is expected to be in ample supply in the foreseeable future. Because of its ability to function both as the aggregate as well as an integral part of the binder, sulphur has been used to upgrade locally available, poorly graded sands in sulphur-asphalt sand mixes and as a partial replacement for asphalt cement in conventional bituminous concrete pavement mixtures. Because of limited field experience with these pavements, the viscoelastic layered pavement analysis program VESYS IIM was used to predict their performance in a variety of climates and to compare this performance with that of conventional asphaltic concrete pavements placed in identical conditions. Material properties of the mix designs used in the analyses were all measured in the laboratory. The resulting predictions showed that sulphur asphalt pavements may be expected to have less rutting and maintain a higher serviceability index but to have a comparatively greater susceptibility to fatigue cracking than conventional asphaltic concrete pavements.
04049 Evaluation of Existing Pavement Based on Deflection and Radius of Curvature and Overlay Design
Y. Miura, T. Tobe

To perform maintenance and repair of an asphalt pavement having reached its service limit, it is necessary to make a mechanical evaluation of the existing pavement and subgrade. On a road being in service, however, it is difficult to perform destructive examination. It is therefore needed to establish an evaluation procedure by non-destructive method. In view of this, the author et al. have dealt with a procedure for evaluating existing pavements by determining the stiffness moduli of the subgrade and pavement from data obtainable from the surface of pavement, i.e. deflection and radius of curvature induced in the neighborhood of the center of loading, and with the aid of the two-layer system. Further, concerning the method to determine the radius of curvature, an examination was made of the deflection curve taken as the basis for this.

There are two methods of dividing the existing pavement into two layers: one into the subgrade and pavement and the other into the asphalt treated layer and bearing layer. The former is suitable for evaluation by compressive strain of the subgrade and deflection and the latter, by tensile strain at the bottom of asphalt treated layer and radius of curvature,

Field verification was made on local roads being in service about 10 years and by comparing the stiffness moduli of the subgrade and pavement, which were estimated from material testing of the existing pavement with due consideration for environmental and traffic conditions, to those estimated by nondestructive method. For the subgrade, the resilient modulus from a repeated compression test performed with the field suction condition reproduced in laboratory was compared to that obtained by nondestructive method. As a result, it was confirmed that the nondestructive method permits to estimate the stiffness modulus with good accuracy.

As for the asphalt treated layer, the values determined by the Shell procedure were compared with those obtained by nondestructive method, and it was confirmed that there existed good correspondence. It should be noted that this comparison was done at a reference temperature (20C) and that temperature correction was therefore made for all of the field data.

Since it was ascertained that not only deflection but also the stiffness moduli of the pavement and subgrade follow a logarithmic normal distribution, field measurements of deflection and radius of curvature were statistically treated to present an adequate procedure for determining a typical structure of existing pavements.

As the existing pavement was evaluated by the two-layer system, overlay design was made by the three layer system. That is, examinations were made assuming the pavement structure as being divided into the subgrade, pavement and overlay when a limit is placed upon strain on the subgrade or deflection and as being divided into the bearing layer, asphalt treated layer and overlay when a limit is placed upon tensile strain at the bottom of asphalt treated layer or radius of curvature.

The values of deflection and radius of curvature estimated by computation for two types of overlay used on the site were compared to the values measured after the overlaying. As the result, the computed values were found close to the average of the measurements, proving adequacy of the design.
04050 A Stochastic Model for Pavement Performance and Management
F. Moavenzadeh, B. Brademeyer

A methodology for the analysis and selection of pavement systems, given a set of goals and constraints is presented. A set of models at three different levels of analysis is discussed.
1) analysis of the physical behavior of the pavement system,
2) analysis of the distress and performance of the pavement system, and
3) analysis of the selection and optimization of maintenance activity given the initial pavement design configuration.

The first involves a mechanical model of the pavement system, which is taken as a stochastic linearly elastic or viscoelastic layered halfspace subject to a static circular normal loading, which predicts the primary responses of the pavement system. These are the stress, strain, or deflection at any point in the halfspace.

The second model uses the Boltzmann Superposition Principle to convert the primary responses into limiting responses under a realistic operating environment, with mechanical and phenomonological models to convert these limiting responses into cracking and permanent deformation. These distress parameters are then introduced into a probabilistic version of the AASHO serviceability equation, and the serviceability and reliability of the system are predicted. Changes in the state of the serviceability are modeled as a homogenous Markov process.

The third level of analysis utilizes transfer functions to convert maintenance activities into improvement in system performance. Decision analysis, with traffic and maintenance levels being the independent variables, is then used to rank and optimize various maintenance alternatives with or without budget constraints. In all of the levels of analysis, the system responses and performance are random variables, as is traffic, with environment and maintenance level being deterministic variables. Sensitivity analyses performed at M.I.T. have shown that the trends predicted by the model are in reasonable agreement with the anticipated behavior of real world systems.
04051 The Analyses of Asphalt Pavement Research Results on the Tomei and Chuo Expressways
Nihon Doro Kodan

This progressive report summarizes two topics obtained from the pavement research program carried out on the heavily travelled expressways.

1. Relationship between Thickness Index, TI and 10 t EAL Applications. To examine the relationship between thickness index, TI and 10 t EAL applications, the following three basic variables were evaluated.
a. Evaluation of pavement performance by means of pair-wise comparison.
b. Evaluation of traffic condition in terms of 10 t EAL applications.
c. Evaluation of pavement structure in terms of the equivalent asphalt thickness, TI, as placed incorporating subgrade and underlying materials into the calculation.

Relationship between TI vs. 10 t EAL applications to the terminal level revealed that there were two types of pavement lives, the structural one represented by cracking and the functional one represented by rutting. The former showed the curve linear relationship between TI and 10 t EAL applications. However, the latter did not correlate with axle load applications but is limited to around 10 x 10^6 axles.

2. Development of a Mathematical Model for Rut Depth Prediction. In the foregoing analysis, rutting resistance of the bituminous mixture was recognized to be very important to extend the overall pavement life. Therefore, significant factors affecting the occurrence of rut depth were investigated by a statistical method and then mathematical models to predict rut depth were developed.

The most significant factor selected was 10 t EAL applications followed by PI of recovered asphalt, air void, type of aggregate, annual sunshine duration, TI and slope index. Rut depth prediction model could predict rut depth at the accuracy of +/- 3.2 mm with 95% confidence level.

Modification of the structural design curve is now in consideration and mixture design criteria were modified based on these test results.
04052 Failure Models and Pavement Design and Rehabilitation Systems Developed and Adapted for Conditions Prevailing in the Nordic Countries
The Nordic Cooperative Research Project for the Application of the AASHO Road Test Results (STINA)

Application of the road test results in other areas than the test site faces three main problems
– pavement failure mechanisms
– subgrade characteristics and their seasonal variation
– the relation between failure and axle load spectrum.

These problems have been dealt with in a joint research effort in the Nordic countries (Denmark, Finland, Iceland, Norway and Sweden), which has been in operation since the beginning of 1974.

Three failure models (AASHO modified, Texas and Ontario) have been fitted to serviceability data from test roads and found applicable when using the same parameter values as those giving the best fit in their original application.

In an analytical examination of pavements complying with Nordic standard specifications, assuming current asphaltic pavement fatigue criteria, a range of subgrade criteria was reached, which contained within its limits the Shell subgrade formula. Examination of the power law for load equivalence factors gave an exponent range around four.

From a previous quite extensive axle load study at some 200 test points in the Swedish road network vehicle equivalent factors have been calculated, using different exponent values in the "fourth power formula". It was found that variation in the exponent had much greater influence upon the vehicle equivalent factors than variation in the traffic load spectrum.

Application of the analytical design and rehabilitation models (SAMP) to Nordic conditions have been examined and an extension of such systems applicable to decision processes for rehabilitation is suggested.
04053 Design System for Minimizing Asphalt Concrete Thermal Cracking
Mohamed Y. Shahin

A design system for minimizing asphalt concrete thermal cracking has been developed. Asphalt concrete "thermal cracking" occurs in the form of transverse, longitudinal, and/or block cracking. Major factors causing these cracks include (1) very-low temperatures resulting in tensile stress that exceeds mixture tensile strength, and (2) daily temperature cycling resulting in repeated tensile stresses and strains that cause asphalt-concrete thermal fatigue. Therefore, thermal cracking occurs not only in cold climates, but also in climates where there is a large daily temperature range and high solar radiation. The author has found several pavements in Phoenix, Arizona having severe block cracking, even though the minimum air temperature is 30°F. However, the average daily temperature range is about 30°F and the solar radiation is quite high causing a large pavement daily temperature range. Similar cases were found in Texas, Alabama, Florida, and So. California. Although the initial occurrence of the cracks has little effect on pavement performance, routing and filling the cracks becomes a maintenance problem. If cracks are not maintained they may rapidly deteriorate. As water infiltrates the cracks, swelling or consolidation of pavement or subgrade layers may occur causing considerable roughness. Spalling of the cracks in airfields can cause foreign object damage (F.O.D.) to aircraft engines. As severe as the problem may seem, thermal cracking can be minimized or completely eliminated for new asphalt pavements through the use of a rational mixture design system. Such a system has been developed and verified and is presented in this paper.

The verification showed that the system is a reliable tool for selecting the most suitable asphalt grade, asphalt supplier, and mixture design for minimizing thermal cracking. The inputs to the system are the environmental conditions where the pavement is to be constructed, asphalt grade, asphalt concrete mixture characteristics, and variability associated with the inputs. The output of the system is the amount of thermal cracking as function of age, expressed in ft^2/100 ft2 or ft/100 ft ^2. A procedure is also described to assess the effect of different amounts of thermal cracking on pavement structural integrity and surface operational condition.
04054 Volume II of Proceedings – preliminary pages

Table of Contents
Sponsoring Organizations
Conference Committees
General Program
Opening Session Address by Egil Nakkel, General Director, Dept of Road Construction Techniques, Federal Road Research Institute, Cologne, Germany
04055 Session I — Design Methods (discussion)
Moderators: W. R. Hudson, R. C. G. Haas

Moderators’ Opening Remarks . . . W. R. Hudson
Presentation by Author . . . A. I. M. Claessen
Discussion of Paper by . . . A. I. M. Claessen et al
Discussion Form Questions/Comments
Presentation by Author . . . F. N. Finn
Discussion of Paper by . . . F. N. Finn et al
Discussion Form Questions/Comments
Summarization of Paper . . . R. C. G. Haas
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04056 Session II — Design Methods (discussion)
Moderators: J. F. Shook, L. E. Santucci

Moderators’ Opening Remarks . . . L. E. Santucci
Presentation by Author . . . W. J. Kenis
Presentation by Author . . . J. B. Rauhut
Summarization of Papers . . . J. E. Shook
Summarization of Papers . . . L. E. Santucci
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04057 Session III — Structural Design Systems (discussion)
Moderators: R. G. Aalvin, T. F. Mcmahon

Moderators’ Opening Remarks . . . R. G. Ahlvin
Presentation by Author . . . J. Verstraeten
Presentation by Author . . . W. R. Barker
Presentation by Author . . . B. Celard
Discussion Form/Questions Comments
Moderators’ Summary . . . T. F. McMahon
04058 Informal Session A — a New Pavement Design (discussion)
Chairman: G. Y. Sebastyan

Opening Remarks . . . G. Y. Sebastyan
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04059 Session IV — Design Methods (discussion)
Moderators: P. S. Pell, S. F. Brown

Moderators’ Opening Remarks . . . P. S. Pell
Summarization of Papers . . . S. F. Brown
Presentation by Author . . . W. D. 0. Paterson
Review of Design Papers . . . P. S. Pell
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04060 Session V — Permanent Deformation (discussion)
Moderators: R. D. Barksdale, R. G. Hicks

Summarization of Papers . . . R. G. Hicks
Summarization of Papers . . . R. D. Barksdale
Discussion by Authors
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04061 Conference Banquet

04062 Session VI — Overlay Design (discussion)
Moderators: B. F. McCullough, R. A. McComb

Moderators’ Opening Remarks . . . R. A. McComb
Presentation by Author . . . J. Bonnot
Presentation by Author . . . H. J. Treybig
Presentation by Author . . . N. W. Lister
Presentation by Author . . . C. K. Kennedy
Presentation by Author . . . A. I. M. Claessen
Moderators’ Summary . . . B. F. McCullough
Discussion Form Questions/Comments
04063 Session VII — Overlay Design Methods (discussion)
Moderators: K. Wester, W. Visser

Moderators’ Opening Remarks . . . K. Wester
Moderators’ Summary . . . N. Visser and K. Wester
Discussion of Summarized Papers
Discussion Form Questions/Comments
Miscellaneous Written Remarks
04064 Informal Session B — Overlay Design Papers (discussion)
Chairman: G. B. Sherman

Opening Remarks . . . G. B. Sherman
Discussion Form Questions/Comments
Miscellaneous Remarks
04065 Session VIII — Summary Session (discussion)
Moderator C. L. Monismith

Conference Summary . . . C. L. Monismith and F. N. Finn
Appendix A – Paper Categories
Appendix B – Pavement Maintenance and Rehabilitation Considerations
Remarks . . . J. F. Shook
Remarks . . . N. W. Lister
Remarks . . . C. F. Scheffey
Discussion Form Questions/Comments
Concluding Remarks . . . C. L. Monismith and F. N. Finn
Conference Closing Remarks . . . E. Tons
04066 Corrections to Papers Printed in Volume I

Corrections to Papers Printed in Volume I.

These corrections have been placed at the end of each relevant file.
04067 List of Registrants

List of Registrants
04068 Composite Bituminous Pavement Design to Prevent Transverse Cracking
I. Deme, F. D. Young, R. W. Culley

Previous research projects have demonstrated that asphalt type and grade are the major factors influencing transverse cracking of asphalt pavements at low temperatures. Paving mixes made with asphalts of low temperature susceptibility (high penetration index) and/or soft grades have been found to be resistant to transverse cracking.

Composite asphalt pavements, comprised of an upper non-cracking and a lower transverse crack susceptible layer of asphalt mix were constructed over sand and clay subgrades in Manitoba and Saskatchewan. The composite pavements did not crack providing evidence that the asphalt in the top layer was the prime factor responsible for the initiation of transverse cracking at low temperatures. The project also demonstrated that transverse cracking of mixes, normally susceptible to cracking, can be prevented or reduced when such mixes are restricted to use as binder courses or pavement bases in conjunction with a non-cracking asphalt surface course. Pavement temperature studies indicated that the insulation provided by the surface course(s) was sufficient to shield the temperature susceptible underlayers from their critical cracking temperatures.

A procedure has been developed for selecting composite asphalt pavement lift thickness in areas with low winter temperatures. It is based on preventing cooling of the layers of various asphalts to their critical cracking temperature. The design approach takes into consideration the critical cracking temperature of the asphalts used, the pavement temperature distribution with depth, and the intensity and recurrence intervals of low temperature extremes in the area. The use of composite asphalt pavements in accordance with these design principles optimizes both the low temperature performance of the pavements and the use of asphalt resources.
04069 Urban Pavement Management on a Coordinated Network-Project Basis
M. A. Karan, R. C. G. Haas

This paper describes the basic components of an urban area pavement management system, and the development of some of the key technology. The approach is primarily concerned with determining investment priorities on a network basis, rather than with sophisticated within- project design.

A user-related "Urban Serviceability Index", USI, measure has been developed. It is a combination of subjectively established Riding Index and Appearance Index, which can be correlated with objective, mechanical measurements of roughness and damaged area of the pavement, respectively.

A performance prediction model, incorporating the USI measure and various pavement classes, has been described. It can be applied to alternative improvement "strategies" and the priorities to be applied to these strategies can be determined through an optimization model. A key feature of the optimization model is the ability to evaluate trade-offs associated with project timing or scheduling.

Finally, the basic components of a pavement inventory system, based on periodic acquisition of performance and other data, have been described. This data is used as "feedback" and the inventory system becomes an inherent component of the overall pavement management system.
04070 Materials Characterization for Asphalt Pavement Structural Design Systems
T. W. Kennedy, A. S. Adedimila , R. C. G. Haas

Elastic and viscoelastic analyses of pavement structures must rely on proper characterization of the pavement layer materials. Variations in materials properties through testing error, lack of uniformity, or differences between tests can significantly affect the design thicknesses selected. This paper suggests that inadequate or improper materials characterization can invalidate the results of structural analyses.

The paper first relates the structural responses of major interest in design to the appropriate structural subsystems and the basic materials properties required. Criteria for selection of methods to determine these properties are then described and a classification of available tests, matched to properties, is presented. Both elastic and viscoelastic characterization is considered.

A simple example of the potentially significant effect of variations in materials properties on the number of traffic applications carried by a given pavement structure is presented.
04071 Evaluation of Subgrade and Its Overlying Flexible Pavement Layers
S. K. Khanna, D. Raja Rao

In a rational design approach for flexible pavements, stresses and strains at critical points in the pavement structure are to be limited to acceptable values. But, the difficulty arises in determining the strains at critical points and their acceptable values. Again, to evaluate the performance of flexible pavements, the maximum surface deflection alone under standard wheel load condition is not sufficient. Some investigators have shown that the product of deflection and the curvature of the deflected basin under the standard wheel load reasonably is constant for a given type of pavement. In terms of critical points of the pavement structure, the maximum deflection controls the limit vertical strain over the subgrade and the curvature of the deflected basin controls the limit horizontal tensile strains in the stiffened layer.

In this study theoretical design charts have been developed for evaluating the structural performance of flexible pavements under the standard wheel loads by deflection criteria which takes into account the maximum deflection and the curvature of the deflection basin in terms of spreadability. A method to determine the elastic moduli of the subgrade and its overlying flexible pavement layers using plate load tests under standard wheel load conditions has also been suggested.

To verify the above evaluation chart, an experimental investigation was conducted through full scale laboratory tests under static wheel loads using rigid plate as well as dual wheels. These tests were carried out over the subgrade, three layers of water bound macadam and bituminous carpet. During these load tests, the surface deflections as well as deflections over the subgrade and other layers were measured.

From this test data, the effect of deflection ratios, plastic deformations and the subgrade deflections with variation in plate types and pavement thicknesses are studied. The suggested method also includes the determination of elastic moduli and equivalent elastic moduli of pavements.

In India, the base course construction generally consists of water bound macadam layers. Such layers do reduce enormously the deflections under the standard wheel loads up to an optimum thickness, but due to lack of cohesion, it does not always improve spreadability with increase in its thickness. This drawback of water bound macadam as base course for heavy traffic conditions is clearly brought out in this study.
04072 South African Mechanistic Pavement Design Procedure
R. N. Walker, W. D. 0. Paterson, C. R. Freeme and C. P. Marais

The mechanistic procedure for pavement design described in this paper is total in its concept and largely complete in its content and verification. The design philosophy, which is based on mechanistic rather than empirical models distinguishes between maintenance which strengthens the pavement structure and that which merely improves the surfacing condition. By using for structural design purposes a life with a high confidence value, and average lives for the estimation of the nature and timing of maintenance, a realistic economic analysis is made using a discount rate that includes the effect of inflation. The uncertainties involved in relating models of material behaviour to models of pavement behaviour, and especially performance, are identified as areas for future work. The errors involved in traffic prediction are quantified and these show that strong preference should be given to dynamic weighting techniques when a mechanistic design is considered. Controlled trafficking tests are used to show that load equivalency factors vary enormously depending on pavement composition and the criterion of distress.

Bituminous materials are fully characterized with respect to fatigue, crack propagation and deformation, and some new original work is included. It is recommended that stiffness and repeated load tests be used in measuring these properties, and that it is essential to use a viscoelastic response in the case of deformation analysis.

Much new work on cementitious materials is presented. These materials are characterized by linear elastic behaviour in flexure, and by fatigue behaviour as a function of maximum tensile strain. The effects of shrinkage cracking and thermal stresses are quantified in the design. The characterization of granular and cohesive materials draws on published data together with initial results from recent work.

A highlight of the paper is the strong verification obtained for the traffic-associated behaviour of bituminous and cementitious materials in full-scale pavements under controlled accelerated trafficking tests using a Heavy Vehicle Simulator. Mechanistic evaluation of the test data yielded a fund of unique information, especially in permitting a precise identification of the timing and location of cracking in cementitious layers. A number of full-scale experimental pavements under normal traffic were also used to verify the behaviour of bituminous materials.

Finally, a two-level approach is proposed for implementing mechanistic design techniques. The comprehensive procedure which can handle any combination of loading, environment and material properties is normally only warranted in special cases, and en application of this to the design of container terminal pavements for wheel loads in the range 70-450 kN is used as an example. A second and much simpler procedure, which utilizes a catalogue of designs, is used as a preliminary analysis to the comprehensive one, and is also recommended for routine pavement design use for design traffic not exceeding 20.10^6 equivalent 80 kN axle loads.
04073 Simple Elastic Models for Pavement Evaluation Using Measured Surface Deflection Bowls
G. Wiseman, J. Uzan, M. S. Hoffman, I. Ishai and M. Livneh

A simple procedure for pavement evaluation is presented, and its applicability verified. This procedure is based on the interpretation of measured surface deflection bowls, using the Hogg and Burmister two-layer elastic models, with a rigid bottom at a finite depth below the subgrade. The real multi-layered pavement structure is replaced either by a plate or by a single elastic layer. Therefore, influence charts and evaluation nomograms can be easily prepared for any specific case of pavement thickness and loading configurations, and are presented in the paper.

Of the two models presented, the Hogg model is simpler to use at the present time, as separate evaluation charts are not required for each pavement thickness. Since both models result in approximately equal values of the subgrade modulus (E sub 2), there is little justification for using the more complicated Burmister model for this purpose. The Burmister model, however, shows more promise for quantitative evaluation of the pavement structure.

Examples of the interpretation of measured surface deflection bowls, using these simple models, are given in this paper. They have proven to be a valuable aid in flexible pavement evaluation. Their very simplicity should make them attractive to pavement engineers.
04074 Index of Contributers and Discussers

Index of Contributers and Discussers