6th International Conference on the Design of Asphalt Pavements - Volume contents and preliminary pages

n/a

Structural Design of Asphalt Pavements for Heavy Loads
J.F. Shook, J.A. Burton

A thickness design method for Full-Depth asphalt pavements, based on the use of multi- layered elastic theory and subgrade vertical strain criteria, is described. Design charts and tables are provided in the manual for two climatic conditions, a cold climate with frost conditions, and a warm climate with relatively constant subgrade conditions. Guidelines are given for selecting design wheel loads for a variety of vehicles, such as fork-lift loaders, off-road haulers, log-handlers and straddle carriers. The procedure is applicable to single load, dual load and multiple wheel load problems. Ten years experience with an earlier version guided the authors in selection criteria and typical material properties used to produce the design procedure.

ILLI-PAVE Based Full-Depth Asphalt Concrete Pavement Design Procedure
Marshall R. Thompson

The basic concepts and the development of a FULL-DEPTH ASPHALT CONCRETE THICKNESS DESIGN PROCEDURE are presented. The proposed procedure is based on resilient soil and material testing procedures, the ILLI-PAVE structural model, and design algorithms developed from an extensive ILLI-PAVE data base. Traffic (18k equivalent single axle loads), subgrade modulus, location (pavement temperature effects), asphalt cement grade (AC-10, AC-20) and design reliability factors (AVERAGE-INTERMEDIATE-HIGH) are considered.

Comparisons of ILLI-PAVE, SHELL, and The Asphalt Institute thickness requirements indicate ILLI-PAVE thickness requirements are quite reasonable. In general, ILLI-PAVE thicknesses are 'intermediate'. The ILLI-PAVE based procedure is 'modular', easy to use, and can be easily modified to accommodate a range of AC mixtures (gradation, asphalt cements, fatigue characteristics, etc.), subgrade conditions, and local climatic effects

An Integrated Approach for Determining Additive Requirements in Hot Mix Recycling
V.P. Sepvas, A.C. Edler, M.A. Ferreira, E.J. van Assen

Additives are needed in hot mix recycling in order to rejuvenate aged binders when this cannot be accomplished by using only new penetration grade bitumens. However, the use of additives can have significant adverse effects on the behaviour of asphalt mixes. Despite much progress in this area, questions relating to aspects such as time-dependency and chemical compatibility remain largely unanswered. The paper examines some of the earlier work undertaken in the development of additives and evolution of specifications, and the subsequent research aimed at developing a comprehensive method for the effective determination of the need, correct type and amount of additive for particular recycling situations.

Development of a Structural Design Procedure for Asphalt Pavements With Crushed Rubble Base Courses
G.T.H. Sweere, A. Penning, E. Vos

This paper describes the first phases of a major research project carried out in the Netherlands into the possibility of recycling demolition waste as a base course material for asphalt pavements. Since no long term experience regarding the structural contribution of base courses built with recycling materials was available, a fundamental research into the behaviour of base courses in general was initiated. The following phases in the research project can be distinguished:
- Material characterization, with emphasis on cyclic loading triaxial testing of unbound base course materials.
- Development of a finite element computer program for calculation of stresses and strains in pavements.
- Construction of a series of full scale test pavements, equipped with transducers for measurement of stresses and strains.
- Verification of the finite element model, using the data from the test pavements.

The first two phases of the project are described in detail, followed by a description of the first test pavement.

In the material characterization phase of the project, a number of recycling base coarse materials such as crushed concrete, crushed bricks and mixtures of these materials were tested, together with Eifellith lava as a reference material. Cyclic loading triaxial tests were carried out to determine the resilient modulus M sub r as a function of the stress level. Also, static loading triaxial tests were carried out in a search for a simpler test for determination of M sub r. A Large number of standard tests was carried out on all the materials investigated in order to check the possibility of a quick assessment of material quality.

The general purpose finite element program DIANA was modified in such a way that it can take into account the stress dependent resilient behaviour of granular materials. A secant approach to the contour model developed by the University of Nottingham is used to model the data from the cyclic Loading triaxial tests for input to the DIANA program. Details of the calculation procedure are given, together with the results of a first series of calculations.

Finally, the layout of the first of a series of test pavements is described. The test pavements are equipped with transducers for measurement of stresses and strains to provide the data needed for validation of the design procedure being developed in the final phase of the project.

Rational Model for the Flexible Pavements Deformations
P. Jove, J. Martinez, J.L. Paute, E. Ragneau

The improvement of flexible pavements design methods required for the "Laboratoires des Ponts et Chausses" in France the development of researches and experimentations in three ranges:

1. The unbound granular materials and soils behaviour study.

2. The creation of a numerical analysis model.

3. The pavements behaviour observation.

The first part of the paper presents the experimental results on the materials. The studies we achieved with the repetitive loadings triaxial on the unbound granular materials allowed the modelization of the nonlinear elastic behaviour of these materials and the drawing of relationships between a "permanent deformations modulus", the loadings number and the stresses which are applied. We demonstrate also that, for a family of UGM completely crushed, the elastic performances are above all dependent on the chippings shape and that the permanent deformations are very related to the moisture content. For the subgrades we show that the elastic deformations are related to the effective pressure due to the weight of the soil and the pavement structure. The relationships, determining the permanent deformations modulus, take also into account this parameter, the value of which is dependent on the hydrologial environment of the pavement. The second part of this paper is composed of the presentation of a design method allowing the determination of the internal forces and the rutting about a pavement. In a first time, we determine the stresses obtained in reversible deformations. For this, we use a nonlinear finite elements method, which allows to take into account the behaviour laws obtained experimentally. The advantage of this method is at first explained on an example, in comparing the results we obtained with those corresponding to a linear elastic behaviour of the materials. Then, we demonstrate how the systematic utilization of a such method allows the determination of charts for the pavements design. We show after how, from the stresses obtained in reversible behaviour, we can use the permanent deformations laws to determine the rutting of a pavement. An example illustrates this design method and we give the evolution of rutting depth in function of the cycles number.

The reader can refer to the paper by Messrs. AUTRET, de BOISSOUDY and GRAMSAMER (ref. 91, which shows the means used to align the theoretical models from experimentations in true scale.

Influence of Bitumen Hardness on the Fatigue Behaviour of Asphalt Pavements of Different Thickness Due to Bearing Capacity of Subbase, Traffic Loading and Temperature
W. Arand

The most well known and worldwide used fatigue laws are based on the results of bending tests on specimens of bituminous materials. Using one of these fatigue laws it can be demonstrated that the number of applicable load alternations increases with decreasing temperatures. Asphalt pavements however show increasing tensile stresses with drop of temperature because of the restrained thermal contraction within a range of adequate low temperatures. Superposition of these thermal induced tensile stresses to the bending stresses will result in a smaller number of applicable load alternations until a first cumulative damage will occur, if the pavement is kept below a certain temperature limit. The use of a harder bitumen type - par example a bitumen pen 20 instead of a bitumen pen 80 - entails a shifting of the critical temperature of about 15C in the direction of higher temperature. Therefore we can draw the conclusion that the use of harder bitumen types at lower temperatures involves a significant smaller number of applicable load alternations until fatigue occurs.

Because of the higher life expectancy of asphalt pavements with harder bitumen types within the range of higher temperatures, however, it will be difficult to take a decision on the application of harder or softer bitumen types with regard to the fatigue behaviour, if we don't have sufficient information about structural, traffic and especially climatic requirements.

In order to solve the above mentioned problem the following topics should be investigated:

- determination of hourly variation in temperature spread as well as in traffic with regard to traffic volume and traffic structure spread for one representative year and definition of classification schedule for both;

- determination of bending stresses and thermal induced tensile stresses in asphalt pavements of different thickness due to traffic loading and temperature as well as to hardness and sensitiveness of bitumen to temperature and superposition of both kinds of stresses:

- evaluation of the number of applicable load alternations until the occurrence of cumulative fatigue damage on the base of known fatigue laws and interpretation of the results with the aid of MINER's hypothesis;

- development of selection principals for the application of proper bitumen types with regard to structural, traffic and climatic requirements.

The Influence of Stiffness-Progress of the Different Pavement Layers on the Size and Shape of Rut Depth in the Pavement Surface
A. Gerlach, A. Loizos, H. Lucke

Various institutions in the F.R. of Germany and abroad are presently considering the problem of ruts as a criterium of damage. Observations in practice have shown that rut formation is one of the main reasons for repair work on roads and is an important characteristic relating to traffic safety.

In the most widely developed road dimensioning concept up to date, the VESYS-system, only the permanent deformations formed at the centre of the applied load are considered in the theoretical determination of rut depth. Furthermore, only central traffic loading along a line is taken into account. In addition to the rut depth considered here, the shape of the rut normal to the direction of traffic movement is also a governing factor in relation to the present serviceability. This follows from the assessment of the rut depth in terms of the present serviceability index PSI, which refers to a base width of 1,22 m. The importance of the rut shape in relation to traffic and travelling comfort is also evident by further literature research.

With the aid of the VESYS rut model and based upon the multi-layer theory, a method has been developed for theoretical determination of the size and shape of ruts. By this method, the contributions of individual layers to rut formation in the pavement surface may also be determined.

Effects of Load Distributions and Axle and Tire Configurations on Pavement Fatigue
Herbert F. Southgate, Robert C. Deen

Damage factor relationships for axle and tire configurations are presented. Adjustment factors are provided to account for variations in load distributions within axle groups, distances between axles of a tandem, and variations in tire pressure for both dual and flotation tires.

Properly accounting for accumulated fatigue of a pavement requires a reasonable measure of traffic volume, proportions of vehicle styles (classifications) within the traffic stream, dates of service, estimate of the average damage factor for each classification, and estimates of tire contact pressures.

All adjustment factors presented are based on analyses of a limited number of structures and should be used with caution. The accuracy of these analyses is not in question, but the range of structures investigated was limited. They are intended to indicate the trend, shape, and sensitivity of various inter-relationships and their relative magnitudes. Modifications may have to be made upon the analyses of additional pavement structures. Kentucky traffic may differ from that in other areas, both in types of vehicles in the traffic stream and the type and direction that cargo is being transported.

Seven Years' Experience With the Structural Aspects of the Shell Pavement Design Manual
A.H. Gerritsen, R.C. Koole

The experience gained with the Shell Pavement Design Manual (SPDM) since its publication in 1978 has been evaluated in 1984/85, which resulted in two essential conclusions. First, its use in practice has not revealed any systematic deviation in the structural design procedure and second, the users of the manual generally express a desire for more explicit information with respect to various aspects of the design procedure, especially with respect to the incorporation of safety margins.

This paper discusses some of the aspects of the structural design procedure in more detail and gives practical guidelines to raise the level of confidence of the design result from 50% to 85% or even 95%. With proper safety margins incorporated, the SPDM design results are found to be very similar to those of other design procedures. Unlike many other design procedures, however, the SPDM makes it possible to select the level of confidence suited for a specific design situation.

Two case studies serve to confirm that (i) the actual service life of a pavement is close to the designed value provided the actual material properties are used for the design, and (ii) the estimation of the material properties generally includes some uncertainties, which require the use of proper safety margins.

Influence of Mix Design on Reflection Cracking Growth Rates Through Asphalt Surfacing
T. Brooker, M.D. Foulkes, C.K. Kennedy

The mix design of asphalt surfacings rarely recognises the different functions these layers perform when laid over granular, bituminous and cement bound roadbases and the consequent need for different stiffness and fracture toughness properties. This paper defines the conditions under which reflection cracking due to thermal stresses will occur and enables estimates of the combined influence of thermal and traffic stresses to be made. The influence of mix variables and fabric interlayers are also discussed in relation to test results.

A Temperature model has been developed to determine roadbase daily temperature range and mean surfacing temperature for each month of the year, within a composite pavement; these determine the magnitude of crack opening movements and the brittleness of the surfacing. Thermal reflection cracking is considered to result from a daily cyclic fatigue mechanism rather than an extreme low temperature mechanism.

Test rigs have been developed to simulate both thermal and traffic movements but only a limited number of tests have been performed with traffic simulation; results are still being evaluated. Finite element analyses of the pavement structure indicate that stress intensity factors decrease as the crack develops under thermal stresses but increase under traffic stresses. Thus initial crack development is controlled by thermal stresses and final cracking to the surface is assisted by traffic stresses. The finite element model shows that thermal stress intensity factors are related to cyclic crack opening, crack length, mix stiffness and surfacing thickness. It also enables crack growth rates during testing to be determined from cyclic displacements at the surface of the samples that were monitored by specially developed portal frame gauges.

These finite element results enable a fracture mechanics interpretation of the test results to be made that serves as the basis of a predictive model for thermal reflection cracking of surfacing thicknesses differing from the 100 mm used in tests. The predictive model is partially validated by limited full-scale observations of cracking on a 100 m section of untrafficked road after 7 years.

Finally, the effect of accelerating simulative tests from a 24 hours cycle to 0.1 Hz has been considered in this study where the crack growth rates and thus fatigue lives have been shown to be related to bitumen stiffness as defined by the SHELL Nomograph. Bitumen stiffness is a useful parameter for 'low stiffness' fatigue testing because it accounts for the influence of test temperature, test frequency and bitumen grade on fatigue life.

Latest Developments in the Analytical Methods for the Design of New Pavements and Strengthening Overlays in Belgium
L. Heleven, J. Verstraeten, V. Veverka

The practical methods presented are primarily intended for project designers and those responsible for road networks (State, Provincial or Municipal authorities). The object of the proposed method for the structural design of new bituminous pavements is to avoid fatigue cracking of bituminous layers and excessive permanent deformation of road structures.

The practical solutions recommended offer a choice between two variants:
- flexible structures (bituminous layers, crushed stone base, granular sub-base),
- semi-rigid structures (bituminous layers, lean concrete base, granular sub-base).

The object of the proposed method for the design of strengthening overlays of old flexible pavements is the determination of the thicknesses of the strengthening layers and also the thicknesses of the layers to be eventually rebuilt. Several variants of the solution are proposed with a view to a technico-economic comparison.

The paper is a synthesis grouping the latest developments of analytical methods based on results of fundamental researches presented to the former Ann Arbor Conferences (1967, 1972, 1977 and 1982).

Characterization and Structural Assessment of Bound Materials for Flexible Road Structures
L. Francken, C. Clauwaert

In order to generalize pavement design and management systems, prediction methods have been developed at the Belgian Road Research Centre.

These methods allow the assessment of the dynamic modulus, the fatigue law of bituminous mixes on the basis of a limited number of input data describing the mix volumic composition and the binder characteristics. The general formulas proposed were obtained after a statistical analysis of a large variety of compositions.

The properties measured on mixes containing modified bitumens have been used to illustrate a procedure for the determination of equivalent thicknesses. A structural factor allowing the generalization of design methods can be derived in this way.

A stress analysis procedure based on the method of finite differences was used for the study of the stress distribution in cracked road structures or in the vicinity of buried joints.

A Design Procedure Based on Experimental Results
H. Buseck, H. Hurtgen

Structural design of pavements essentially means that layer thickness and material properties be so selected as to ensure that the end of service life is not reached within a given time. End of service life in this context is defined by a rut depth limit. In the Federal Republic of Germany, the long-term behavior of most pavements , studies on test roads, those in a full scale testing laboratory, and laboratory tests, all point to rut depth as failure criterion. Fatigue of bituminous road materials has been a very rarely observed phenomenon.

Experimental design methods require that an adequate number of variables of the pavement in question be subjected to real traffic till the end of service life. Full scale pavement tests can shorten this time. However, transfer functions then have to be set up to translate the results of the experiment into real life conditions. In theoretical pavement design methods, the material properties determined in the laboratory are used in a model and the long- term behavior of the pavement is forecast based on this model. It is assumed that the model is able to describe reality. Verification by means of the real behavior of an experimentally designed pavement will not be possible before the end of design life is reached. The pavement design strategy proposed here, being largely based on experimental results, also makes use of a calculation model. However. the results of the model are calibrated on the behavior of a full scale pavement subjected to fatigue loading.

The model calculation can thus be verified within a relatively short time, although applying to a special case only. The authors believe that the probability of describing real-life conditions by a model are much higher if it also applies to the full scale pavement test. In addition, the use of permanent deformation, in terms of rut depth, as the only failure criterion is considered as sufficient and is also justified. The calculation model is thus greatly simplified, becoming more realistic at the same time.

Although full application of this method has yet to be made, the results obtained so far are considered worth presenting here. In this connection, the relationship between the permanent deformation of bituminous road materials and above all the number of load applications is of importance. It exists in the same form in laboratory tests, in the full scale.pavement test, and in the real life conditions of a test road. With this and its temperature dependency it is possible to forecast the long-term behavior of a pavement, provided normal pavement behavior can be assumed, i.e., without excessive stresses and strains on the unbound subgrade. If that is not the case. it should be detected by the calculation model beforehand to be prevented from occurring.

Application of Reliability Concepts to Pavement Design
Paul Irick, W.R. Hudson, B.F. McCullough

This paper is an overview of the major aspects of pavement design reliability that were developed by the authors for inclusion in the 1986 revision, of the AASHTO Pavement Design Guide. Reliability (R) is generally defined to be the probability that a designed pavement will perform satisfactorily over a specified design period.

The measure of 'performance is taken to be the actual number of equivalent axle loads, N sub t, that the pavement carries during its performance period, i.e., the time during which a particular distress indicator, d, does not exceed a specified terminal value, d sub t. The indicator may be for singular distress such a cracking or rutting, or for composite distress such as roughness or serviceability loss. If N sub T is the actual number of equivalent axle loads tnat occur during the design period of T years, then reliability is the probability that N sub t will be at least as great as N sub T.

It is assumed that the design is based on two specific prediction equations or algorithms. The first is a function of design period traffic factors that gives predictions, w sub T, for actual design period applications, N sub T. The second is a function of load applications, other loading factors, environmental factors, and structural factors that gives predictions, ^d for the distress indicator, d.

If the second equation is solved for load applications with ^d = d sub t, the result is a design equation that gives predictions, W sub t, for actual performance, N sub t. In design practice, W sub t represents design applications, and is the product of a designer-selected reliability factor, F sub R > 1, and the design period traffic prediction, w sub T. The probability basis for F sub R is the set of all quadruples (N sub T, w sub T, W sub t, N sub t) that would result from (say) hundreds of independent pavement projects for which the same design procedure and reliability level has been used.

Probability distributions associated with these quadruples are assumed to be log-normal. Deviations between log W sub T and log N sub T are traffic prediction errors, deviations between log N sub t and log W sub t are performance prediction errors, and, by definition, the deviation between log W sub t and log w sub T is log F sub R. The algebraic sum of these three deviations is the overall process deviation, delta sub 0.

It is shown that R = Prob (delta sub 0 > 0). and that delta sub 0 is normally distributed with mean value log F sub R and variance S^2 sub O, where S sub 0 is the standard deviation of the overall design-performance process. If delta sub 0 is transformed to a standard normal variate, z, and if z sub R corresponds to delta sub 0 = 0, then R = Prob (z > z sub R), and for a given value of R, the reliability factor is given by log F sub R = -z sub R S sub 0.

The process variance, S^2 sub O, is the sum of the two prediction error variances, and each of the latter can be decomposed into higher level variance components that represent replication variance and lack- of-fit of the prediction equation. For performance predictions, replication variance components are unexplained variance and variance attributable to differences between design levels and as-constructed levels of the prediction factors.

Rationale for reliability level selection is discussed, and an example is discussed for the application of reliability to flexible pavement design.

Research needs include evaluation of variance components for all pavement design equations in current use, and development of objective criteria for selection of reliability levels.

The Effect of Truck Tire Contact Pressure Distribution on the Design of Flexible Pavements
C. Saraf, R. Marshek, H. Chen, R. Connell, W.R. Hudson

This paper presents the results of a study to determine the effect of truck tire contact pressure distribution on the response of a flexible pavement. An experimental procedure was developed to measure the tire contact pressure distribution under a treaded truck tire, This contact pressure distribution was used to determine the effect of increased tire Inflation pressure and wheel load on the response of flexible pavements. A 3D finite element program called TEXGAP-3D was used to analyze the pavement sections selected for this study.

The conventional analysis of flexible pavements is generally performed by layer analysis programs. Therefore, a layer analysis program called ELSYM5 was also used to analyze all the pavement sections used in this study. A uniform and circular tire contact pressure was assumed for this analysis.

The results of the TEXGAP-3D analysis were compared with the results of the ELSYM5 analysis to determine the effect of truck tire contact pressure distribution on the design of flexible pavements. Limited data used in this study indicated that for a given pavement section the conventional analysis of flexible pavements overestimated the tensile strain at the bottom of the surface layer. The percent increase in tensile strain due to increase in tire inflation pressure is also overestimated by the conventional analysis. However, the conventional analysis underestimated the percent increase in the surface tensile strain due to an increase in wheel load.

The vertical compressive strains at the top of the subgrade are underestimated by the conventional analysis. The increasing inflation pressure had negligible effect on the vertical compressive strain as Indicated by both the conventional (ELSYM5) and the TEXGAP-3D analysis. However, both analyses indicated that a 20 percent increase in wheel load resulted in about 19 percent increase in vertical compressive strain at the top of the subgrade.

Calculation of the Rutting of Structures - Castor Program Method for Prediction of Permanent Deformations in Asphaltic Structures
Honore Goacolou

Asphaltic pavement structural design is generally based on criteria of two types. One concerns the fatigue behavior of the materials (failure), the other permanent deformations of the pavement profile (subsidence or rutting). A method of predicting the permanent deformations of asphaltic structures is proposed for use in connection with the latter problem.

The calculation of the permanent deformations is based on a finite-element numerical method (two-dimensional model). The behavior laws of the materials are of the linear elastic type for reversible deformations, while permanent deformations are described by laws of the viscoplastic type, which may or may not include a threshold of plasticity. The formulation of the model leads to the non-linearities resulting from the viscoplasticity being taken into account by forces equivalent to the antecedent viscoplastic deformations. In this formulation, the rigidity matrix is constant and only the "forces" change over time. The system of equations is solved at constant rigidity by an iterative method.

The calculation of the rut, over a reference interval, involves working out the diagram of the evolution of the input parameters and dividing it into homogeneous elementary intervals. A finite-element calculation is carried out for each interval; the resulting permanent deformations are multiplied by the corresponding number of passages (the permanent deformations are additive), then incorporated in the geometry of the structure by modifying the coordinates of the nodes of the mesh. After chronological processing of all of the intervals, the final mesh represents the rutted structure.

The input parameters of the model are:
- the master curves of the asphaltic materials;
- the laws of flow of the asphaltic materials;
- the vertical temperature distribution;
- the speed of the vehicles;
- the axle weight;
- the transverse position of passage of the axle;
- the elastic moduli and Poisson's ratios of the other materials.

The method is illustrated by a simulation of the LPC rutting tester, testing the susceptibility to rutting of asphaltic materials. The viscoplastic behavior of the test specimen under the passage of a rolling load is simulated using a law of flow in pure shear. The main results given by the CASTOR calculation program are presented:
- the initial, then deformed, meshes of the structure;
- the lines of creep of the materials;
- the level curves of a principal tensor characteristic of the permanent deformations.

Finally, two deformed meshes are compared, one obtained experimentally and the other produced by the calculation model.

Evaluation of Heavy-Vehicle Traffic and Its Application to Pavement Structural Design
M. Siffert, B. Lescure

In-depth knowledge of traffic make-up and loading conditions is of primary importance for the proper design and management of highway facilities. Current methods of evaluating loads on roads and bridges and their effects are overly complex and inadequate. The piezoelectric ceramic cable has undergone extensive investigation and development in France for use as a traffic detector, or more precisely a load detector. More economical than traditional methods and involving easier maintenance, this detector can find broad applications in many fields. A traffic evaluation method based upon the real aggressiveness of loads and using of this type of detector will doubtless make it possible to improve the accuracy of pavement structural design methods and the definition of maintenance priorities.

Rut Depth Prediction: A Practical Verification
B. Eckmann

This paper presents the contribution of the ESSO Research Centre at Mont-Saint-Aignan (France) to an overall research program monitored by the Road Engineering Division of the Rijkswaterstaat. Two test overlays of asphaltic concrete were therefore built in 1978 on National Highway No. 28 in The Netherlands. The main objective of the program was to check the ability of various design methods to predict the behavior of the test sections.

Within the framework of ESSO Road Design Technology, the rutting behavior of bituminous mixes is assessed by means of a specifically designed dynamic creep test. When recorded against the number of load cycles, the axial permanent deformation of the specimen shows a rapid initial growth followed by a steady creep rate regime. Rut depth prediction is derived from computed stresses and temperatures and with reference to the creep curves obtained at laboratory under various state of stress and temperature conditions.

In the usual ERDT rutting subsystem, initial creep is not taken into account. All the bituminous layers are supposed to be in a "linear phase" of creep. Calculations are thus simplified but provide only the rut depth evolution after the first period of rapid deformation. In this particular case, this was not sufficient as the short time elapsed since the overlay construction conferred a key role to the initial creep in the road test analysis.

The contribution of this early rutting stage was estimated while describing the first part of the experimental creep curves by analytical laws of the type epsilon = alpha.t(super) Beta ( epsilon = permanent deformation, t = loading time). Like the creep rates, coefficients alpha and Beta are related to state of stress and temperature by means of a regression analysis.

By contrast with steady stage rut depth computations, the rutting development calculated at any time then depends on the deformation already undergone by the pavement. The computations are therefore made according to a stepwise procedure.

When compared to the rut depths actually measured, the calculated values proved to be excellent in one case and too low in the other case. These results are discussed through a sensitivity analysis applied to temperature and traffic data.

It could thus be demonstrated that the contribution of high temperatures, even when limited in time, is of paramount importance. This entails that a temperature record as detailed as possible is required for reconciling calculated rut depth values with actual field measurements.

Sensitivity to traffic was studied as a function of axle load and tyre pressure. It was namely found that, due to the effect of higher lateral stresses developed within the upper layers, a lighter axle may be more damaging than a heavier one (at constant tyre pressure). A load equivalency factor such as those derived from the fourth power law thus does not apply to rut depth prediction. Moreover, the process for backing up such an equivalent factor in the case of rutting remains to be explored and is likely to be more complex than for the fatigue case.

Due to the large number of input data which are to be known with high accuracy, a precise rut depth prediction appears to be difficult. But this does not lower the interest of rational methods such as the one described here.

We see it as being twofold :
- understand and emulate the relative importance of each environment and structural parameter. - provide a comparative basis for assessing the rutting performance of a given pavement versus a reference one, under given conditions for temperature and traffic. Along these lines, the ERDT approach for permanent deformation is confirmed by this conspicuous field test as an efficient aid to the pavement designer.

Design Practice for Bituminous Pavements in the United Kingdom
N.W. Lister, W.D. Powell

A new method for the structural design of bituminous road pavements has been developed by the Transport and Road Research Laboratory. It is based on the systematic analysis of the performance of a large number of experimental pavements interpreted in the light of structural theory. The sub-base is designed primarily to carry construction traffic. Standard design curves are given for the thickness of roadbase required to carry the traffic expected to use the road during its design life, life being defined in terms of the timing of pre-emptive pavement strengthening.

The designs take into account the effect of variability in road performance and there is independent evidence of their validity. Probabilistic criteria for design against pavement deformation and fatigue cracking are established and a new procedure for predicting the internal deformability of bituminous pavement layers is introduced. The method offers the means of adapting the standard designs to take advantage of new materials and design configurations: this is particularly relevant for designs for very heavy traffic and for pavement reconstruction. Examples of developments that are leading towards changes in design practice are given.

Mechanisms of Surface Cracking in Wearing Courses
M. Dauzats, A. Rampal

Because of the extent of the damage, the appearance of surface cracking in asphaltic concrete wearing courses in Europe and across the Atlantic has led to the development of models for the prediction of thermal cracking and to investigation of the parameters that influence its spread.

The method presented in this article, in the context of this approach, is suited to the needs of highway managers. It uses a combination of two models. One is an adaptation to French pavements of the thermal cracking prediction model developed by Y. Shahin. The discontinuities constituted by these surface cracks are entered into the second model, which uses fracture mechanics to investigate the propagation of the cracks towards the roadbase under the influence of rolling loads.

The extent of cracking and the rate of crack propagation, as a result of thermal effects only, are estimated using the first model, which gives, from the climatic conditions of the site and the characteristics of the mix and asphalt used, or to be used, the degree of cracking per year in service.

This two-dimensional mechanistic model uses the finite-element method, which provides a good approximation of a cracking model and satisfactory stress intensity factor values. The rate of propagation is calculated by integrating Paris' law and determining the number of cycles required for the crack to reach the wearing course/roadbase interface. The investigation is carried out for three loading cases, and the damage to the wearing course is determined as a function of the transverse distribution of the traffic.

The method described is aimed at giving highway managers information that can help them to choose, according to the site and its climatic conditions, the class of asphalt that will give the least thermal surface cracking, or, in the case of existing pavements, to determine the depth of penetration of surface cracks into the pavement so that the appropriate maintenance work can be done.

Employing Paving Asphalt Temperature Susceptibility in the Structural Design of Asphalt Pavements
Norman W. McLeod

Paving asphalt temperature susceptibility is defined, and a simple method for its measurement is described.

The influence of paving asphalt temperature susceptibility on pavement design and performance in hot countries without frost, and in cold climates with frost is described.
Requirements for paving asphalt temperature susceptibility that can be added to or incorporated into a paving asphalt specification are proposed and discussed.

It is shown that by the addition of suitable polymers, the temperature susceptibilities of paving asphalts can be changed dramatically. This development implies that what has always been an asphalt supplier's market could be changed to an asphalt user's market. In an Appendix, the validity of pen-vis number (PVN) as a measure of paving asphalt temperature susceptibility is examined, together with some of its implications for pavement design and pavement performance.

Estimation of Fatigue Life of Asphalt Pavement
Kenji Himeno, Takashi Watanabe, Teruhiko Maruyama

In this study, a new fatigue failure criterion of asphalt mix is presented, based on the energy dissipation theory. When the mix stiffness modulus is low, fatigue test data cannot be obtained from laboratory cyclic bending test, thus the phase angle between the stress and the strain sinusoidal waves is measured, employing the cyclic wheel tracking test, and the new fatigue failure criterion is proposed which is valid for the wide range of mix stiffness modulus, estimating the energy dissipation in the mix.

Examining the new criterion, it is found that bending fatigue damage at bottom of the mix slab is comparatively large in spring and can be ignored in winter and summer. It was also found that fatigue failure can initiate at top of the mix slab when the mix stiffness modulus is low.

A new system to predict the load associated pavement fatigue life considering the external factors affecting the life, such as wheel load, transverse wheel position, pavement temperature, as random variables. In the system, pavement fatigue failure is assumed to initiate at both top and bottom of the slab, and the shorter one of the predicted lives is defined as its fatigue life. From the results of case studies applied for Japan national highways, the predicted lives were found to correspond well with field data obtained by government agency.

Catering For Long Term Changes in the Characteristics of Asphalt During the Design Life of a Pavement
F. Hugo

Asphalt is subject to ageing. This causes hardening and volume change of the binder during the life cycle of a pavement. Both phenomena have already been extensively studied. However, with respect to the influence of these phenomena, certain aspects required investigation.

A laboratory procedure for simulating the ageing process was used to evaluate the variation in the engineering parameters of the asphalt during the life cycle of a pavement. Typical South African continuously and gap-graded asphalt mixes were tested. The engineering characteristics were determined for aged and unaged asphaltic mixtures at low, medium and high temperatures that is at -5 C, 25 C and 40 C.

By using elastic layered theory in conjunction with the analysis of stresses due to rapid cooling down of the asphalt layers it was confirmed that thin asphalt layers as used in South Africa would be prone to cracking under particular environmental conditions. This was found to be both load and non- load associated as well as due to the interaction of the two. In an attempt to validate previous findings and hypotheses a full-scale pilot study on experimental pavements is being carried out using a Heavy Vehicle Simulator in conjunction with a temperature chamber capable of varying the pavement temperature between -10 C and 50 C. The tests are being conducted on pavements especially prepared for this purpose and both continuously graded and gap-graded asphalt with and without precoated chips are being tested. The program also includes the testing of pavements which have been artificially aged by using heat and UV-radiation similar to a process utilised in the laboratory.

At temperate and high temperatures the variation in binder viscosity is most important since it causes the pavement to have variations in plastic behaviour in relation to depth. This in turn has been shown to cause residual stresses to build up. These stresses can either prolong or shorten the fatigue life of the asphalt depending upon the structural system and the location within the asphalt layer.

Results obtained to date have shown that the design of asphalt pavements should take into account the fact that the engineering parameters which influence the structural behaviour of such pavements, are subject to environmental influences. The parameters used in mathematical analyses have to be selected or varied to cater for changes throughout a single day and also for changes which are due to environmental influences during the life cycle of the asphalt. It will be shown that the life cycle of a pavement needs to be analysed in separate phases covering inter alia the following periods: initial construction, early life, middle age, aged and rejuvenated or recycled. This should lead to a clearer understanding of distress mechanisms and phenomena such as so called "premature cracks". Interim guidelines are proposed for the design of asphalt mixes and pavements.

Evaluation of Fatigue Properties of Recycled Asphalt Concrete
Elton R. Brown

Recycling of aged asphalt concrete pavements has been demonstrated to be cost-effective and to reduce the demand for natural resources such as aggregate and asphalt. Because of the advantages derived when using recycled materials, the capability to predict long-term performance is needed so that optimum benefits can be obtained. This study was undertaken to evaluate the flexural fatigue performance of recycled asphalt concrete mixtures and to compare these results to those measured for conventional asphalt concrete mixtures.

To make these comparisons, samples of aged asphalt concrete were obtained from three locations where recycling was planned. These samples were blended with new aggregate and new asphalt materials to produce six different recycled mixtures.

Two aggregate types, a crushed gravel and a crushed limestone, were used to produce two conventional mixtures and to blend with the reclaimed asphalt pavement to produce the six recycled mixtures. Three asphalt materials which were obtained to produce the various mixtures being evaluated consisted of AC-20 for preparing the conventional mixtures and AC-5 and a recycling agent for preparing the recycled mixtures.

The flexural fatigue properties ware evaluated for all mixtures. Tests were conducted on the asphalt binder (combined binder for recycled mixtures) prior to mixing with aggregate. Tests on the binder included penetration at 40F and 77F; ductility at 40F and 77F; softening point; viscosity at 275F, 225F, and 140F; specific gravity; and rolling thin-film oven test. Tests conducted on the asphalt concrete included all tests conducted during mix design and flexural fatigue test at 40F and 77F.

Test results indicate that recycled mixtures can be designed to perform as well as conventional mixtures when tested in flexural fatigue. The properties of the blended asphalt binder in the recycled mixture should be similar to the properties of a new asphalt binder to provide satisfactory results.

Higher Accuracy in Flexible Pavement Construction Design
F. Lehovec, F. Luxemburk

This paper concentrates on stress in flexible pavement constructions derived from:

- horizontal forces acting on the wearing course,
- vertical forces, assuming an imperfect interaction between the top layer and road base.

To facilitate the calculation of stress and deformation in an n-layer system, TANLAY program has been written which operates on the following assumptions:
- The individual construction layers behave like perfectly elastic three-dimensional bodies characterized by modulus of elasticity E sub i, Poisson number mu sub i and layer thickness h sub i horizontal dimensions are unlimited,
- the sub-base behaves like an elastic half-space characterized by modulus of elasticity E sub n and Poisson number mu sub n,
- measure of interaction between the layers and sub-base can be prescribed in the range from perfect friction to perfect slip,
- the surface of the system is subject, on an area with radius a, to a uniform horizontal directional (shear) load.

However, under these assumptions the radial stress on the surface of the critical wearing course at the circumference of loading area is infinitely great.

In order to determine the magnitude of stress in these troublesome singular points the construction system was expanded by an additional layer situated in contact with the wearing course. Material characteristics of this layer (E, u ) and its thickness are identical with those of a tyre tread. Under this simplification the stress of a number of road constructions was computed and their operational capacity, expressed as the number of design axle passages, was evaluated.

For calculation of stress and deformation in an n-dimensional system by vertical forces the LAYMED program is used. Theoretical premises in the LAYMED program are analogous to those in TANLAY. Only the loading is different. The LAYMED program assumes a vertical load distributed uniformly over a circular loading area.

In this case the investigation focused on determining the measure of interaction at the contact of top and road base layers. A number of constructions was investigated, with the interaction at the interfaces ranging from perfect friction to perfect slip. The obtained results have borne out the fact that whenever construction work fails to secure a perfect bond between construction layers, operational capacity and service life of the designed pavement are severely reduced.

The Behaviour and Mechanistic Design of Asphalt Pavements
C.R. Freeme, M. De Beer, A.W. Viljoen

The mechanistic method of the design of asphalt pavements, as reported to the last Conference, has received a wide degree of acceptance by road authorities in South Africa. Additional work has enabled the application of the principles embodied in the method to lead to the development of more effective rehabilitation methods.

This paper describes the importance of understanding the behaviour of asphalt pavements, with the emphasis on its state of distress. The importance of the type of support under the asphalt layer is emphasized, whether this be granular or cemented layers. Typical moduli for these layers in the distressed state is given for use in mechanistic rehabilitation methods. A variety of distress mechanisms is also discussed.

The present state of the pavement can be assessed more accurately and rehabilitation strategies better selected by using the methods and information discussed.

A General Analytically Based Approach to the Design of Asphalt Concrete Pavements
C.L. Monismith, F.N. Finn, G. Ahlborn, N. Markevich

This paper describes a general approach to the structural design of asphalt concrete pavement systems, making use of research which has been developed in the past 25 years. The approach is based on the assumption that pavement systems respond to loads elastically; however, the actual nonlinear responses of the materials comprising the pavement section are incorporated in an ad hoc but representative manner.

Analyses of representative pavement systems are made using an updated version of the ELSYM5 program developed originally in 1972. This updated version, termed ELSA, permits consideration of up to 100 loads applied to the pavement including 10 different load types (defined by tire pressure, radius of loaded area, or magnitude of load). The pavement structure can be represented by up to 10 different layers.

Examples of the use of the approach are presented and include:
1. Design of a new airport pavement system in the Middle East.
2. Analysis of the rutting potential in an airport pavement in the Middle East and modifications in asphalt mix design.
3. Assessment of increased tire pressures on the rutting potential of pavements in the southwestern United States.
4. Assessment of the potential of mixtures containing a modified asphalt to solve specific distress problems.

Developments to the Nottingham Analytical Design Method for Asphalt Pavements
Janet M. Brunton, Stephen F. Brown, Peter S. Pell

The method of pavement design developed at the University of Nottingham uses theoretical analysis and mechanical properties of pavement materials in a procedure which is implemented by use of computers. The techniques have been described to previous conferences as they have evolved, so the present paper provides a summary, together with a description of developments which have taken place recently. Amongst the detailed matters considered is the improvement to the subgrade strain criterion, to take account of differing deformation resistance offered by various asphalt mixes and the elastic stiffness which should be assigned to granular sub-base. Comparisons are made between the results of the Nottingham design method and that recently proposed by the UK Transport and Road Research Laboratory.

New consideration has been given to the appropriate terminal pavement condition for design purposes and the concept of "design temperatures" has been evolved. While the Nottingham design method has, quantitatively, been developed for British conditions, it can be used for other environments and an example is given of how it has been adapted for North America.

Prediction and Prevention of Surface Cracking in Asphalt Pavements
A.H. Gerritsen, C.A.P.M. Van Gurp, J.P.J. Van Der Heide, A.A.A. Molenaar, A.C. Pronk

In several relatively new road pavements premature cracking in the top layer(s) has occurred. It has been found that this type of cracking could be surveyed both in and outside the wheelpaths. According to the currently used mechanistic design procedures with linear elastic multilayer programs the greatest tensile strains are predicted in the bottom of the asphalt layers and not at the surface. To search for causes of this type of distress and to provide recommendations for the prevention, theoretical analyses, laboratory experiments and field studies have been conducted. In the theoretical analyses the currently used traffic models and fatigue Parameters have been evaluated. The tentative results show that radial shear forces under rubber tyres can attribute to surface cracking. Secondly it is shown that energy parameters can be useful in the structural analysis of top layers. Experiments have covered aspects such as mix composition, mix properties, fatigue, strength, thermal stresses and dynamic load induced residual stresses. The fatigue characteristics of the mixes appeared to be not sufficient mainly due to excessive hardening of the bitumen. The mixes have been tested for their response on low temperatures. Tests on stress relaxation could be described very well by the Burgers' model. This enabled the use of the model for investigating the influence of loading time, temperature, and load repetitions on the asphalt mix. Field Studies containing visual condition surveys and falling weight deflection measurements have been conducted to check the results of the laboratory experiments. Tentative recommendations are presented to prevent premature cracking of asphalt wearing courses.

Influence of Wheel Load and Inflation Pressure on the Rutting Effect at Asphalt-Pavements - Experiments and Theoretical Investigations
J. Eisenmann, A. Hilmer

The presented paper includes a study of the influence of wheel loading and inflation pressure on the rutting effect at asphalt pavements. The statements are based on laboratory tests with single tire equipment as well as with twin tire equipment and on theoretical investigations. The tests were performed with a test facility which enables full-scale rutting tests with loading by real tires, test specimens of natural dimension, and practical temperatures. In the here described test series only the parameters wheel loading, inflation pressure, and tire arrangement (single tire, twin tires) were varied. The pavement system, the temperature gradient (according to the relevant temperatures of a hot summer day), and the speed in test were not changed.

The rutting development was measured by superimposed transverse surface profiles. The dependence of rutting on the influence parameters loading number, wheel load, and inflation pressure was determined by regression analysis.

The theoretical investigations were based on the elastic multi-layer theory using computer program BISAR of SHELL whereby deformation was separated from general strain. The calculated deformations were compared with the test results. With the results of theoretical calculation the test results could be completed or extended.

The tests and the theoretical investigations have shown that the wheel load, the inflation pressure, the average contact pressure between tire and pavement surface (depending on both factors), and the tire arrangement have an important influence on rutting effect. With regard to the design of heavy trucks these parameters need a particular consideration.

Asphalt Mix Design for Optimum Structural and Tyre Interaction Purposes
G. Lees

In recent years there has been a noticeable trend towards redefining the desirable qualities of bituminous pavement surfacing materials and towards the development of design methods aimed at achieving these qualities. Resistance to deformation, to cracking, to adhesion failure and to wear, and durability in the sense of resistance to adverse physical and chemical changes as a consequence of weathering effects upon either aggregate or binder, are the structural properties required. In addition, for wearing course compositions, adequate skid resistance over the anticipated speed range of vehicles , plus low noise generation, low spray generation, reasonably low tyre wear, low vibration characteristics and low rolling resistance, are the main aims in design and construction. All in this second group are concerned with the interaction in one form or another between tyre and pavement material.

Notwithstanding the need for research studies involving more fundamental and complex test methods, the requirement of industry for a quick and reasonably reliable test method has led to the adoption on a fairly world-wide scale of the Marshall test both for design and control of bituminous mixes. This paper aims to provide a closer link between the Marshall method and some of the more important findings from fundamental research studies. Some long held misconceptions concerning the Marshall test are discussed and a new approach to Marshall asphalt mix design proposed involving:
a) a rationalisation of Marshall test parameters,
b) determination of a Design Binder Content by the method of 'ranges' rather by the traditional method of 'averaging',
c) the need to design asphalt mixes for thin carpets by testing specimens of corresponding thickness,
d) the greater engineering relevance of designing according to limits of permeability rather than of air voids,
e) via the latter the possibility of inclusion of design of deliberately pervious ('drainage') asphalts into the Marshall method,
f) recognition of the importance of the different stiffness requirements of asphalt mixes for fatigue life, as between structurally thick bituminous pavements and structurally thin bituminous pavements.

Probabilistic and Reliability Analysis of the California Bearing Ratio (CBR) Design Method for Flexible Airfield Pavements
Yu T. Chou

The California Bearing Ratio (CBR) design method for flexible airfield pavements was analyzed using a probabilistic approach. The design parameters considered were the load P (or the equivalent single-wheel load), the subgrade CBR, the tire contact area A, and the pavement total thickness t. The expected value and variance of the dependent variable performance factor alpha (which is logarithmically related to the number of passes to failure) were estimated by using the Taylor series expansion and the Rosenblueth method. Differences in computed results between the two methods were found to be small, although the derivation of the expressions for Taylor series expansion was very complicated. A procedure was developed to estimate the reliability of the designed pavement system based on known variabilities of design parameters. Results of the reliability analysis indicate that prediction of pavement performance is most influenced by variations of pavement thickness t and is least influenced by variations of tire contact area A. The effects of variations of wheel load P and subgrade CBR are identical. The weighting factors for parameters t, CBR, P, and A, in general, are approximately 1, 0.34, 0.34, and 0.01, respectively.

An Analysis of Load and Non-Load-Related Effects on Flexible Pavement Performance
Tien-Fang Fwa, Kumares C. Sinha

The determination and knowledge of the relative effects of traffic and environmental factors have significant implications in pavement design, pavement performance monitoring and evaluation technology. In addition, this information would also assist in pavement maintenance and rehabilitation planning and decision making. The paper presents an analysis of the effects of load and non-load factors on flexible and overlay pavement performance by means of:
(a) a performance-based procedure to estimate quantitatively the relative effects (or responsibilities) of load and non-load factors; and
(b) statistical correlation and regression analyses to investigate how individual load and non-load factors influence pavement performance.

The analysis presented is based upon the data of the state highway system in Indiana. The data used included the following main categories: pavement inventory data, traffic data, pavement performance data, pavement routine maintenance cost data, and subgrade soil data. The AASHTO serviceability concept was adopted as a measure of pavement performance. A parameter, PSI-ESAL loss, was introduced as a quantitative representation of the performance of a given pavement. This quantity is calculated by integrating the PSI loss of the pavement over imposed traffic loading expressed in ESAL. To investigate the influence of individual load and non-load factors on asphalt pavements in Indiana, statistical correlation and regression analyses were performed to study the relationships between the results of performance analysis and individual factor effects. Based upon the results of the performance analysis, conclusions were drawn concerning the relative shares of load and non-load-related effects on the performance of asphalt pavements in Indiana. Statistical analyses provided further information regarding the relative importance of various climatic factors, traffic load and pavement characteristic variables as to how they affect the performance of asphalt pavements in Indiana.

Predicted and Field Performance of a Thin Full Depth Asphalt Pavement Placed Over a Weak Subgrade
R.B. Smith, W.O. Yandell

The mechano-lattice method of analysis method, which assumes elastoplastic material behaviour is described and details are provided of both the bound and unbound options. Prior to this paper the method has been used successfully with a variety of flexible pavements, but this was the first opportunity for the method to be used on an asphalt pavement for which testing was particularly adapted to the requirements of the analysis procedure.

The asphalt pavement was 100 mm thick and overlay a weak imported subgrade, having four day soaked CBR values between 2% and 4%. Both the pavement and pavement materials were extensively evaluated during construction. Following construction, the pavement has been monitored and the results of rut depth determinations, roughness, and periodic visual assessment are presented.

Using the load/deformation behaviour of the constituent materials as input the mechano-lattice predictions were compared with the field performance after the passage of the equivalent of 350 000 standard axles (esa). Two prediction runs were performed. One was based on an asphalt modulus determined from a one second loading time and a relatively dry imported subgrade and the other based on an asphalt modulus determined from a one-tenth second loading time and a relatively wet imported subgrade.

The permanent deformation was overpredicted by approximately 50% for the dry imported subgrade condition and low asphalt modulus, and was close to the field value for the wet imported subgrade condition and high asphalt modulus. In both cases the absolute rut depth was predicted as 18 mm. The field survey data did not allow for the absolute rut depth to be determined.

Output from mechano-lattice analysis and the asphalt moduli were used to predict the onset of cracking using the Shell method. In each case the onset of cracking was predicted to occur at about 100 000 esa. Initial cracking was noted in the outer wheelpath during a visual assessment after the equivalent of 105 000 esa.

Fundamental Properties of Recycled Asphalt Mixes
V.P. Servas, M.A. Ferreira, P.C. Curtayne

Experience has shown hot mix recycling to provide a cost-effective rehabilitation option. To resolve some residual doubts about quality, recycled mixes have been evaluated in South Africa through a laboratory investigation and Heavy Vehicle Simulator testing.

The laboratory study was carried out to determine the initial engineering properties of asphalt mixes composed of different proportions of reclaimed material, yet meeting the design criteria of conventional mixes. It was found that the proportion of reclaimed material had no effect on permanent deformation and fatigue resistance. This study took no account of either the durability characteristics of recycled mixes or the effect of recycling additives. These factors are currently being investigated.

The Heavy Vehicle Simulator has been used to test recycled asphalt base layers. The results of this accelerated testing suggest that the field behaviour of recycled base mixes is comparable to that of conventional asphalt.

Dynamic Anaiysis of Multilayered Pavement Structures - Theory, Significance and Verification
Michael S. Mamlouk

The loads applied by traffic and by most deflection measurement devices on pavements are dynamic in nature. Until recently, analysis of the data obtained from dynamic loadings have been based on either empirical approaches or static models. This paper discusses the state-of- the-art of the use of dynamic analysis (considering the inertial effect) and provides better understanding of the true response of pavements under dynamic loadings. The theoretical formulation and the significance of considering the dynamic response of pavement is presented. The dynamic analysis is significant when shallow bedrock or frozen subgrade is encountered. The dynamic analysis is also more important in case of harmonic loadings as compared to the case of impulsive loadings. The dynamic analysis technique is verified on in-service pavements using field measurements obtained by deflection devices under different conditions. Field deflections were predicted by the dynamic analysis closer than did the static analysis.

Structural Evaluation of Asphalt Pavements in the Eastern Province of Saudi Arabia
Madan G. Arora, Faisal Saleem

In this paper, a framework for estimating the structural adequacy of in-service asphalt pavements is developed based on the Benkelman beam deflection measurements on typically selected highway pavements in the Eastern Province of the Kingdom of Saudi Arabia. Pavement condition survey was also conducted, side-by-side with the deflection survey, following the U.S. Corps of Engineers' procedure based on pavement rating (PAVER) technique. Field cores were extracted from asphaltic layers and tested for dynamic modulus under repetitive loading at varying temperatures simulating the on-site pavement conditions. Moduli of the supporting layers were determined by back-calculation by matching the observed beam deflections with the theoretically determined values using the BISAR computer software. Applications of the prevalent fatigue damage and subgrade rutting models to predicting the pavement life and overlay thickness requirement have been demonstrated for the prevailing in-service conditions in the Kingdom. The developed overlay design chart will be further refined as more data on deflection and performance of overlays of varying thicknesses become available.

A Field Verification of VESYS IIIA Structural Subsystem
N. Paul Khosla

Presented in this paper are the results of the field verification of VESYS IIIA structural subsystem.

Five pavement sections in different geographical locations of North Carolina were selected for this study. The layer materials from these pavement sections were characterized in the laboratory by subjecting specimens of the given materials to a series of creep and dynamic load tests under environmental conditions representative of those experienced in the field. The characterization of asphaltic mixtures was done using the direct compression test and the diametral tension test.

Based on the input of the mechanical properties of the layer materials, the actual traffic volume, and the local environmental conditions, the performance of the pavement sections was predicted using the VESYS IIIA structural subsystem. The predicted performance parameters (rutting, cracking, and present serviceability index) were compared with the actual measured performance parameters.

The predicted performance using the mechanical properties, as determined by the direct compression test, matched quite well with the actually measured performance. On the other hand, the input of the mechanical properties, as determined by diametral tension test, almost always overestimated the performance of the pavements.

Diagnostic Evaluation of In-Service Pavements Performance Using Pavement Condition Data
Waheed Uddin, John F. Nixon. B. Frank McCullough, J. Kabir

A combination of various pavement condition monitoring methods and equipment was used for diagnostic evaluation of in-service highway pavements' performance in Oklahoma. Condition monitoring data, analyses, and interpretation pertaining to the selected pavement sections are discussed in this paper. Pavement Condition monitoring data for this study included:
1) estimates of Present Serviceability Rating (PSR) based on ride quality measurement,
2) visual distress survey, and
3) nondestructive testing.

Results of PSR and visual distress surveys on the test sites were used to identify "good" (performing as intended) and "bad" (poorly performing and severely distressed) pavements.

The extensive nondestructive test data were collected in Oklahoma by operating both the Dynaflect and falling weight deflectometer (FWD) at the same locations, with the FWD immediately following the Dynaflect, providing a unique opportunity for side-by-side comparison of the two devices, and in situ moduli evaluated from their respective deflection basins. The mechanistic interpretation of dynamic deflection basins to ascertain in situ material properties and load carrying capacity was performed using self-iterative computer programs, FPEDD1 and RPEDD1. The methodology used in these programs ensures unique results of in situ moduli and eliminates the user-dependency aspect of the deflection basin fitting approach. The results were comparable. Neither of the devices appears to be significantly underestimating or overestimating in situ moduli.

The mechanistic analyses of dynamic deflection basins of both the Dynaflect and FWD did not provide any evidence of structural inadequacy. Fatigue was not the primary distress mechanism according to the results of distress surveys. Rutting and shoving were prominent among the primary distress manifestations on most of these pavement sites. Laboratory tests (Lottman tests and Boiling tests) and visual observation of core samples provided ample evidence to conclude that stripping in bituminous concrete surface and/or base layers is the most likely cause of premature distress on most of these in-service pavements. This case study shows that although pavement design may be structurally adequate, material problems and their interaction with load may become significant factors associated with pavement deterioration.

Verification of the Analytical-Empirical Method of Pavement Evaluation Based on FWD Testing
P. Ullidtz, G. Battiato, B.K. Larsen, R.N. Stubstad

This paper is based on several years experience with practical use of analytical-empirical (or mechanistic-empirical) methods for pavement evaluation. The paper addresses problems related to the analytical part as well as to the empirical part of the method.

The analytical part consists of determination of layer moduli and calculation of critical stresses and strains. Many models exist but none of them deals with all the complexities of real pavements. The use of the very simple Odemark-Boussinesq method is described and examples of verification of this method are given.

The crucial part of the method is the empirical relations between pavement response and pavement performance. The paper demonstrates that deflection criteria for predicting performance are incompatible with strain or stress criteria. A tentative strain criteria for predicting decrease of asphalt moduli is then presented and compared to a few observed values. Finally, a stress criteria for predicting functional pavement performance (PSR) is verified using data from 157 pavement sections with a combined length of 253 km, and covering a wide variety of pavement structures, subgrade soils, climatic conditions and traffic loadings.

The paper concludes that the analytical-empirical method is a powerful tool for pavement analysis at the present time, but that further improvements and/or verifications are needed for the analytical and, particularly, for the empirical parts of the method.

Derivation of Pavement Material Variability From Nondestructive Testing
Jacob Uzan, David G. Zeitoun, Raphael Baker

In pavement evaluation using nondestructive testing (NDT), a large amount of deflection bowls are analyzed in terms of the elastic moduli of the layers. The results are used to evaluate the material variability, which could serve in an overlay design procedure based on the concept of reliability.

The model currently used for interpreting deflection bowls is based on the random variable theory which neglects the spatial distribution of the elastic modulus of the material. Since the subgrade and pavement materials have a spatial distribution, the analysis of NDT could lead to an underestimate of the material variability. The random field theory, which is more adequate than the random variable theory, is presented and used to correct the NDT analysis. The theoretical approach is limited to the small variation range (i.e. coefficients of variation up to 30 percent).

Two cases - the Boussinesq and the two-layer system - are analyzed with the random field theory, and the covariance matrix of the deflection bowl is obtained and used to generate deflection bowls corresponding to the properties of the random field. These bowls are then interpreted with the current procedure and elastic modulus variabilities are computed. It is found that the current procedure for interpreting deflection bowls underestimates the variability of the subgrade, by a factor of 0.4 to 1.0. It is interesting to note that the average moduli of the Boussinesq layer and of the two layers are not affected by the type of theory used. The variability of the upper layer in the two layers system is also unaffected (for the small variation range).

Figures for correcting the variability obtained from NDT results are presented. The methodology is illustrated and discussed.

Accelerated Full Scale Testing of Heavy Duty Pavements - Experience with the Australian Accelerated Loading Facility (ALF)
P. Kadar

The Accelerated Loading Facility has completed more than 2.5 x 10^6 loading cycles up to date, and provided valuable information on the pavement response to accelerated loading. Once the initial mechanical problems were overcome, the ALF applied an average of 50 000 load cycles per week. Three pavements were tested, each typical to the current design and construction practice of heavy duty pavements of the host State Road Authority.

The pavement performance and structural response was monitored by means of surface and subsurface deflections, permanent deformations, in-depth strains and visual observation of surface deterioration.

A method for assessing critical strains based on the measured surface deflection bowls and in-depth deflections was worked out, which makes it possible to continually monitor (and analyse) the structural performance of the tested pavements. Cross-reference between analytical approaches based on the equivalent layer thicknesses and on the surface curvature index yielded consistent results, increasing confidence in both approaches.

All three tests indicated long pavement life under the given load conditions, consistent with the design life of the tested pavements.

The Circular Test Track of the Laboratoire Central Des Ponts et Chaussees (L.C.P.C.) Nantes - First Results
P. Autret, A. Baucheron de Boissoudy, J.C. Gramsammer

The circular test track of the Laboratiore Central des Ponts et Chausses (Public works Research Laboratory) has been in service for two years. The properties and performance of this sophisticated machine are remarkable, especially as regards the applied load. The first part of the article gives a brief description of the installation and of the measuring apparatus and how it is used. The second part of the article deals with the role of the circular test track in highway research and with the first results obtained on asphaltic pavements. These results concern the factors of equivalence among axles, which vary both with the type of pavement and with the type of damage considered, the representation of flexible pavements by mathematical models, the laws governing rutting and cracking, and the deadlines for repair work in vim of the speed of evolution of the damage once the process has begun.

New Methods Developed in France for Road Network Survey and Maintenance
M. Cuvillier, J.F. Godard, P. Retour

The present increasing use of pavement management systems and the high sensitivity of highways managers to maintenance policy are reinforced by setting up new devices to facilitate road data collection and compilation or to better know characteristics of existing pavements. This paper presents three new methods developed by the Laboratoire Central des Ponts et Chaussees (LCPC):
- the "PULSE RADAR"
- the "VIDEOROUTE"
- the "DECRIROUTE".

Strain Measurements in Bituminous Layers
I.F. Scazziga, A.G. Dumont, W. Knobel

An international expert Group on "Full scale pavement tests", created by OECD in 1983 with the task to improve the use and benefits from accelerated load tests through international co-operation, carried out a joint program of strain measurements on the Italian test facility of Nardo in April 1984. The comparison of the results from various types of instrumentation used to measure strain in asphalt layers was intended to give an answer to the question whether it is possible to consider in an equivalent way strain values obtained through different techniques and finally also the conclusions of tests based on such results of strain measurements. Nine groups of researchers and technicians from 8 different countries participated in the common field experiment, placing their instrumentation in a special test section and carrying out simultaneous readings under the same transient loads. The sum of the results obtained in the course of 3 days by all the participants have been analyzed together with all the boundary conditions for the test. It was found that a major part of the dispersion in the results was due to variations in the test conditions, i.e. homogeneity of the test pavement structure, temperature, speed and exact position of the load. The different types of instrumentation used also showed a certain amount of scatter. However, a detailed analysis indicates that certain solutions are less subject to additional influencing factors such as stiffness of the gauge, method of preparation and placement, use of asphalt materials prepared in the laboratory. Knowledge of all gauge characteristics is thus an important factor in the estimation of its reliability and of the value of the results obtained.

Estimation of In Situ Elastic Moduli of Pavement Structural Layer with Failing-Weight-Deflectometer Deflection Basion
A. Kasahara, H. Kubo, T. Sugawara

This paper describes the estimation method of the elastic moduli of a subgrade (E sub 3), a granular subbase (E sub 2) and an asphalt bound layer (E sub 1). In this study, SAS-BISAR system is fully utilized and the pavement is simplified into three-layer structure by using deflection data, i.e. D sub 0, D sub 300, D sub 750, (suffixes denote distances (mm) from the center of the loading plate), which are obtained by the testing device of FWD. SAS-BISAR system was accomplished by combining BISAR and Super Application System (SAS). BISAR is an elastic analysis method for multi-layer, and SAS is a statistical software, which was developed by SAS Institute in U.S.A. and has valuable functions for data management and drawing figures.

First, we attempt to find the most appropriate deflection variables for estimating the elastic coefficient of each layer of the pavement. The relation between D sub 750 and (D sub 300 - D sub 750) is useful to estimate the value of E sub 3. In a similar way, the relation between D sub 750 and (D sub 0 - D sub 300) and the deviation (D sub 0 - D sub 300) are required for the estimation of E sub 2. It becomes possible to estimate the elastic modulus of the whole asphalt bound layer (E sub 1) by using the relation between D sub 0 and E sub 1.

The elastic modulus of the whole asphalt bound layer, cited above, has a good agreement with the elastic modulus of the asphalt mixture which was obtained in the dynamic indirect tension test with a loading time of FWD. We propose the relatively simple estimating method of E sub 2 and E sub 3 with the known variable E sub 1. In this method, a diagram of the relation between delta sub 750 and (delta sub 300 - delta sub 750) is drawn by the use of SAS-BISAR system. Plotting D sub 750 and (D sub 300 - D sub 750), values which are actually measured, on this figure, E sub 2 and E sub 3 are estimated by interpolation. By the use of this method, E sub 2 and E sub 3 are estimated for various kinds of the asphalt pavement structure. The results are as follows:

1) Seasonal fluctuation in the bearing capacity of the asphalt pavement can be shown in terms of the change in the elastic moduli of the granular subbase and the subgrade.

2) It is possible to estimate the elastic moduli of the granular subbase and the subgrade even for some special types of pavement structure such as the pavement with insulation boards or with a thin asphalt layer.

The Verification of Design Methods - Test Sections
M. Siffert

Pavement structural design, reconstruction and maintenance methods involve the use of theoretical models and varying degrees of experimental verification which should make it possible to adjust the parameters defined in the models and check, on real cases, the assumptions upon which the methods were based.

With the appearance of the first rational pavement structural design and reconstruction methods in France in 1965, a program of "test sections" was set up to allow closer analysis of pavement performance over long periods in order to determine the effectiveness of the methods used and to gradually improve their precision.

The sections were defined and selected in order to match real and representative cases. An experimentation plan was developed using many pavement surveying methods applied at adjustable intervals. These sections, most of which are between 10 and 20 years old, allowed the experimental verification of pavement behavior. It was thus possible to analyse the effectiveness of predictions provided by the theoretical model. Among the numerous results already made possible by this experimentation, we can mention the development of a structural surveying method, medium-and long term behavior laws for the represented structures, an assessment of the results of measurements and observations, the effect of environmental conditions (climate, drainage, etc.) on each type of structure, making it possible to determine the important effect of surface waterproofing, the limit values or thresholds on different survey test parameters, the overall trends in pavement damage, the search for surveying test indicators most sensitive to the structural evolution of pavements and their meaning, the improvement of methods of using the theoretical model (correlation coefficients), and so on.

An Integrated System for the Evaluation of Road Pavements
P.G. Jordan, B.W. Ferne, D.R.C. Cooper

A high-speed road monitor (HRM) for surveying the surface condition and alignment of roads at traffic speed is described. This new computer-controlled equipment uses laser and inclinometer sensors to survey up to 200 lane kilometres per day of longitudinal profile, rutting, macrotexture, crossfall, gradient and horizontal curvature; usually it does this without interference to other traffic using the road. The development of a comprehensive computer-based system for processing the survey data to provide data summaries, exception reports and trend analysis is described. Levels of unevenness, and of change in unevenness are proposed for use in assessing aspects of road serviceability including structural adequacy.

The survey equipment provides a rapid and economic method of locating damaged sections of pavement so that more detailed and costly inspection and deflection surveys can be more efficiently deployed. Sometimes a regulating course may be all that is required to restore surface profile. A method of estimating the material required to restore a deformed surface is presented.

Measurement of deflection under a rolling wheel is firmly established in the United Kingdom as the basis for assessing structural condition and designing strengthening. An automated analysis system for deflection measurements is described that can assist in establishing priorities and designing cost-effective strengthening. The analysis system embodies empirical performance relationships derived from observations on pavements ranging from undesigned paved roads to pavements designed to carry traffic in excess of 100 million standard axles. An example is presented of a least-cost strengthening design provided by the system of analysis.

Prior to strengthening more detailed investigations of sections of pavement identified by rolling wheel deflection measurements are often carried out to identify the causes of deterioration and to establish where strengthening by overlay is appropriate and where reconstruction is needed. A falling weight deflectometer is being used to augment the information obtained from the excavation and materials testing.

The combination of these assessment techniques into an integrated system will allow the best use to be made of the limited resources available for the rehabilitation of road pavements.

Evaluation and Comparison of Various NDT Devices in Side-by-Side Testing on Indiana Highways
J.K. Lindly, N.D. Pumphrey jr., T.D. White, V.L. Anderson

As part of a research program designed to provide an overlay design procedure for pavements in Indiana, four Nondestructive Deflection Testing (NDT) devices were compared side-by-side on 73 pavement sections in Indiana. The devices were the Dynaflect, the Road Rater 400, the Road Rater 2000, and the Dynatest Falling Weight Deflectometer (FWD). The 73 sections represent overlaid flexible pavements and both overlaid and non-overlaid JRC pavements. Brief descriptions of activities performed prior to field testing - including data base development, statistical experiment design, and test section selection - are presented, as are preliminary findings from the first of two planned test sequences. Comparisons of the deflection basins and the deflection variations within pavement test sections are made. Equations are presented correlating deflections from the FWD with deflections from the other three devices. Finally, variables are calculated from the deflections and are used as indicators of relative pavement strength characteristics. These indicators are analyzed to predict which of the NDT devices offer the most usable deflection data for investigating pavement structural properties.

Evaluation of Flexible Pavements and Overlay Design Based on F.W.D. Tests
A. Marchionna, M.G. Fornaci, M. Malgarini

In this report a procedure is described that evaluates the remaining life of flexible pavements from tests carried out with the Falling Weight Deflectometer. A method was developed that allows an evaluation of pavement layers moduli using the deflections measured under the FWD load (PA.STR.EV. Program). A bituminous layers fatigue distress model was also developed that takes into account the effect of cracks propagation. Using this model it is possible to evaluate the allowable load repetitions related to different cracking stages of pavement surface. The input data required by the model are the tensile strains at the bottom of the bituminous layers. The stress/strain distribution induced by traffic loads was calculated using a finite element program (NOL.A.P.) that utilizes the elastic characteristics obtained from the PA.STR.EV. Program. The pavements are schematized as an elastic four layer body: a linear elastic law was adopted for the bituminous layers, a nonlinear elastic law was used for the unbound layers. Using these models it is possible to forecast the evolution of the pavement condition and to determine the pavement surface distress curve. Afterwards it is described a procedure that evaluates the percentage of the initial thickness of a bituminous layer affected by cracking. The fatigue distress model was used to calculate the residual life of the sound part of a partially cracked bituminous layer. The result was a diagram in which the percentage of residual life is linked to the percentage of cracked thickness. Finally this procedure allows the calculation of an overlay working life.

Performance of Flexible Airfield Pavements Subjected to High Traffic Volumes
Starr D. Kohn, Ross A. Bentsen

Presently, the Federal Aviation Administration's (FAA) pavement design and evaluation procedure is based on the Corps of Engineers' CBR procedure. The Corps' procedure is based on accelerated traffic testing of pavement test sections. The highest traffic levels in the testing program were approximately 5,000 coverages of a given gear configuration for flexible pavements. Thus, the design curves in the current advisory circular (AC 150/6320-6C) have been drawn to include pass levels of 25,000 annual departures. However, the traffic levels at many of the major hub airports are receiving traffic in excess of this level. In order to accommodate the thickness design for these pavements, the design curves have been extrapolated to include pass levels of 200,000 annual departures. Since adaptation of this extrapolation procedure, some concern has arisen concerning the adequacy of the design procedure. In response to this concern the FAA initiated a study to determine the adequacy of this design procedure. The study included evaluation of in-service pavements currently receiving more than 25,000 annual departures. Flexible pavement sections at two major hub airports were surveyed and included in this study. This paper will summarize the findings of the field survey of these pavements. Included in the field survey was a visual condition survey using the "pavement condition index" (PCI) procedure and a nondestructive evaluation Using the Waterways Experiment Station's (WES) 16-kip vibratory testing device. Data from the condition surveys and nondestructive testing (NDT) will be presented.

Application of the International Roughness Index to Response Type Measuring Systems
Michael W. Sayers, Thomas D. Gillespie, Cesar A.V. Queiroz

Characterizing the roughness of a road in a universal, consistent, and relevant manner has proven to be a persistent problem over the past 40 years. Two major problems are:
1) measuring methods have not been stable with time; and
2) measuring methods have not been transportable.

In response to these problems, The World Bank and The Brazilian Ministry of Transportation initiated the International Road Roughness Experiment (IRRE), held in Brasilia, Brazil, in 1982. The experiment and subsequent analyses included the participation of researchers and equipment from Brazil, England, France, Belgium, Australia, and the United States. Representative roughness measuring equipment was operated over test sites in the area around Brasilia that included asphalt pavements and unpaved roads. The data obtained were analyzed with the objectives of determining correlations between the various equipment and limitations on their use. The findings showed that it is possible for all of the equipment to measure roughness on the same scale, if the scale is carefully chosen. Thus, an International Roughness Index (IRI) was proposed for future work to eliminate much of the difficulty experienced in earlier studies.

The IRI is based on the concept of the ubiquitous response-type measuring system -- a vehicle equipped with a roadmeter. These systems are used throughout the world in both developing and developed countries. The IRI is defined, however, by a mathematical procedure that is applied to a measured profile, called a quarter-car simulation. The IRI can be measured directly by rod and level methods or by high-speed profilometers. In a validation study, all of the response-type systems that participated in the international experiment could be calibrated to the IRI without loss of accuracy.

Pavement and Material Evaluation by a Dynamic Cone Penetrometer
Moshe Livneh, Ilan Ishai

This paper presents a non-destructive pavement evaluation method based on continuous measurement with depth of pavement layers and subgrade parameters, using a Dynamic Cone Penetrometer (DCP). The DCP pavement evaluation method was developed based on extensive laboratory and field correlation testing. In the laboratory, conventional and DCP tests were made on a wide range of undisturbed and compacted fine grained soil samples, with and without saturation. Compacted granular soils were tested in flexible molds with variable controlled lateral pressures. Field tests were made on natural and compacted layers representing a wide range of potential pavement and subgrade materials. Pavement evaluation tests were also performed for pavement and material evaluation and for correlation with pavement condition.

The correlative laboratory and field testing program resulted in a quantitative relationship between the CBR of the material and its DCP value. This relationship was highly correlated for a wide range of pavement and subgrade materials.

Based on the above extensive laboratory and field testing and analysis, a model was developed for realistic flexible pavement evaluation procedure. This is based on the DCP test through the entire pavement structure and down to the subgrade layers. The model relates the measured strength of individual pavement and subgrade layers and their relative contribution to the entire bearing capacity of the pavement structure.

The model finally resulted in an equivalent thickness of a designed conventional pavement. This equivalent thickness can be interpreted either to the residual pavement life, i.e., the number of coverages of existing load down to a given terminal serviceability, or to the required overlay thickness (or other pavement remedy) for obtaining the designed pavement life.

The experiments on subgrades and pavements have proved that the DCP has a high degree of repeatability under controlled conditions, and is sufficiently sensitive for use in practical pavement and material evaluation. The test was found to be simple and fast, allowing performance of scores of tests in one day.

Despite the asphaltic core extrusion needed for the DCP testing, the evaluation method can still be considered non-destructive. Moreover, it has a major advantage of measuring realistic pavement and subgrade layer parameters with no need for any supplementary destructive testing.

The paper presents the basic DCP test, the results of the correlative laboratory and field testing program, the models for highways and airport pavement and subgrade evaluation and results of the current method of pavement and subgrade evaluation using the DCP and its related models. A numerical example illustrates the practical use of the method.

On Improvements of the Existing VESYS-Concepts
H. Beckedahl, A. Gerlach, H. Lucke, W. Schwaderer

The existing VESYS programs provide engineers not only with a new philosophy how to handle a management system by means of a performance predicting calculation program but also support design engineers with a suitable calculation method. The programs enable us to calculate the progress of PSI (Present Serviceability Index) by means of theoretical treatments. Otherwise PSI could only be obtained by measurements.

About 8 years ago we started to scrutinize those VESYS programs which were available for us. A lot of special measurements were carried out to support our theoretical considerations. Instead of improving parts of existing programs we started to develop a new version which is called VESYS 3H (H stands for Hannover). In this program particularly the calculations of ruts and cracks had been changed. Measurements have shown that longitudinal roughness of roads is correlated to non-uniformities of the road bodies, which are spatially distributed by chance. So far, it seems to be very hard to predict slope variance.

Structural Design and Prediction of Asphalt Pavement Life-Time in Hungary
Laszlo Gaspar jr., Erno Toth

After having mentioned the former main Hungarian research activities in the field of asphalt pavement structural design and life-time prediction, an evaluation method of highway network bearing capacity is presented. Its main advantages are as follows: it makes possible to perform a realistic evaluation, different permissible deflection values can be used, it utilizes the "traffic ratio", a correction because of actual surface damage mark is applied. It is expedient to evaluate every pavement condition parameter simultaneously. A 8-year long research work was being done connected with the deterioration process of asphalt pavements. Hereby the actual condition of some 30 selected highway sections of 500 m length were rated twice a year using several measurements and visual observation. The sections were divided into four classes according to their level of dimensioning (over-dimensioned, well-dimensioned, under-dimensioned, very under-dimensioned groups). Their still expectable life-times were determined by means of graphs where the subjective condition notes can be seen on the vertical axis and the time passed after construction or strengthening on the horizontal axis. The limits obtained are as follows: over-dimensioned sections 11-15 years, well-dimensioned sections 9-13 years, under-dimensioned sections 7-11.25 years, very under-dimensioned sections 6-9.5 years.

Heavy Vehicle Simulator Aided Evaluation of Overlays on Pavements With Active Cracks
A.W. Viljoen, C.R. Freeme, V.P. Servas, F.C. Rust

Conventional overlay life estimates are often invalidated by the reflection of existing cracks in a pavement through an overlay. This is attributable to a lack of understanding of the factors influencing relative crack movements and their mechanisms. Inadequate knowledge of the ability of overlay techniques to tolerate induced strains originating from the discontinuities in the existing overlay pavement also contribute to the problem.

This paper identifies the factors and mechanisms and illustrates their effects by means of field data from actual pavements. The South African Heavy Vehicle Simulator was also used to evaluate a variety of conventional and innovative asphaltic overlays on a severely cracked concrete pavement of which the mechanisms and extent of relative crack and joint movements were determined prior to overlay placement. The results of this testing programme are discussed with special emphasis on the ability of the overlays to inhibit reflection cracking.

It is believed that the improved knowledge of the parameters involved in reflection cracking will contribute to more realistic prediction models and result in more cost effective rehabilitation strategies.

Evaluation of Mechanical Parameters of Inservice Pavements from Field Data
M.G. Sharma, W.J. Kenis, M. Mirdamadi

In recent years, many analytical models have been developed to evaluate damage in flexible and rigid pavements. Most of these models determine each of the three primary modes of distress (rutting, cracking and roughness) separately and require as inputs, the material constants for pavement layers obtained from laboratory testing. It is very well known that laboratory-determined mechanical parameters do not reflect the in situ response of the material in the field. In addition, field condition surveys have indicated that accelerated damage to a pavement generally occurs when a primary distress leads to a secondary distress generated by increased traffic and/or severe environmental conditions.

The paper describes a material parameter identification model for the determination of mechanical properties of pavement layer materials from the field measured data such as the vertical deflections, rutting and cracking. The material identification model has been developed by considering the highway pavement as a layered elastic-plastic medium. The material parameters for the layers have been determined by solving a nonlinear optimization problem using the so-called Simplex method. The mechanical properties determined by the method are the elastic constants, permanent deformation and fatigue parameters for layer materials. The identification model has been used to determine in situ mechanical parameters for in-service highway pavements from New