code	ISAP 3rd Conference -- Titles & Abstracts
03000	3rd International Conference on the Design of Asphalt Pavements - Volume 1 - Preliminary pages and Table of Contents n/a
03001	The Bearing Capacity of Pavements with Frost Retarding Layers. A Test Road Study Olle Andersson, Bjorn Orbom, Goran Ringstrom For an extensive study of frost penetration and bearing capacity of roads with frost retarding layers of different composition and different thickness and surrounded by different types of bases and sub-bases a test road was built in western Sweden in 1966-67. During the construction period and the following year the various layers were followed up by measurement of dynamic elastic modulus by the wave propagation method, and several sections were subjected to repeated plate loading applying up to 50000 loads at each point in order to study the fatigue properties. For investigation of the development of bearing capacity the test road under traffic is followed up by measurements each year implying plate bearing tests, levelling, cross profile and longitudinal profile by the CHLOE profilometer. Most frost penetration retarding materials caused an increase in rebound deflection and did in some cases cause a disastrous permanent deflection, which necessitated rebuilding. Some materials had no detectable influence upon the permanent deflection. The serviceability index as measured by the CHLOE meter decreased faster on some sections having insulating layers but showed in some sections a slower decrease than on the reference sections. The dynamic elastic modulus of the base course increased during the first year by more than 30% also in those having insulating layers beneath. The test road will be followed up for several years to come but has already given positive results for the frost protection aspect of road design.
03002	The Measurement of Traffic Axle Load Distributions for Pavement Design Purposes J. E. B. Basson. G L. Dehlen, R. G. Phillips, P. J. Wyatt One of the required inputs for the design of new pavements, or of overlays to existing pavements, is accurate data on the axle loads of the traffic they will carry. Estimation procedures based on visual traffic counts together with the results of earlier axle load measurements made with loadometers or static weighbridges are sometimes employed to obtain such data for the present traffic. As the variables associated with the conversion of traffic counts to axle loads at a particular location include the commodities carried, the type of vehicle, the season, and police control of illegal axles, such an estimation procedure cannot be expected to render acceptably accurate results at all locations. By means of traffic surveys the effects of some of these variables have been quantified. These surveys have revealed that there is no estimation procedure which does not result in significant inaccuracies in the design thickness of a pavement. By contrast, the direct measurement of axle load distributions, using one of several dynamic weighing techniques, reduces the errors in the evaluation of present traffic to negligible proportions, and is considered to be the approach most suitable for pavement design purposes.
03003	Ste. Anne Test Road - Flexible Pavement Design to Resist Low Temperature Cracking R. A. Burgess, O. Kopvillem, F. D. Young Low temperature transverse cracking of bituminous pavements is a form of pavement distress prevalent in Canada and the northern United States. This distress induces a deterioration in pavement performance through spalling, heaving or settling at the cracks and reduces pavement service life. In 1967, a test road, designed and instrumented for the investigation of this problem, was constructed at Ste. Anne, Manitoba, This road incorporated twenty-nine test sections involving a number of different bituminous mix materials, pavement structures and subgrade types, believed to be potentially important in the study of transverse cracking. After five years of service, the following conclusions can be drawn with respect to the relative performance of the test sections: (1) Pavements incorporating high viscosity type asphalts and softer grade asphalts exhibit a greater resistance to transverse cracking. (2) Pavements with thick asphalt bound layers exhibit a lower frequency of transverse cracking. (3) The type of subgrade influences the frequency of transverse cracking. (4) The asphalt content of the mix, in the range of one percent below Marshall optimum to one-half percent above Marshall optimum does not appear to be significant in affecting pavement transverse cracking frequency. A laboratory study was conducted in conjunction with the field programme to investigate the possibility of correlating laboratory predicted fracture temperatures with the actual field performance of the St. Anne test sections. Predicted fracture temperatures were determined by calculating that temperature at which accumulated thermal stresses exceed the tensile strength of the compacted mix. Additionally, since the tensile properties of an asphalt binder are proportional to the tensile properties of a mix made with that binder, the possibility of predicting pavement field performance by a knowledge of binder properties alone was investigated. Accumulated thermal stresses of the binders and their mixes were calculated from their respective stiffness moduli at one-half hour loading time (cooling rate} over the appropriate temperature range. Accumulated thermal stress and breaking stress (tensile strength) were plotted as a function of temperature. The predicted fracture temperature is the intercept of these curves. Comparison of the laboratory and field results revealed that there is an excellent correlation between the laboratory predicted fracture temperatures of the binder and mix and the temperature of initial cracking of the asphaltic pavement in the field. For practical purposes, therefore, the tendency of an asphaltic pavement to crack can be predicted by a knowledge of the binder stiffness modulus at low temperatures and long loading time. Conversely, the binder, or mix, stiffness parameter may be used as a pavement design criterion to alleviate the transverse cracking problem. These research findings would also imply that it is necessary to be more selective in the use of materials and pavement designs with due consideration being given to their performance under the prevailing climatic conditions.
03004	The Response of Asphalt Pavements to Low Temperature Climatic Environments J. T. Christison, K. O. Anderson The response of asphalt pavements to environmental conditions is well recognized, but not easily quantified. Variations in temperature , particularly freezing conditions, may result in non-traffic load associated distress. Analyses of several pavement systems in Western Canada where continuous temperature records were collected, has yielded specific information on temperature gradients and time variations at various depths throughout the pavement structures. To extend this information to other climatic areas, a computer program has been developed that utilizes a finite difference method to predict the thermal regime within pavement systems. The necessary input parameters for the program are the thermal properties of the component layers and readily available meteorological data. Comparisons of predicted and recorded temperatures have shown excellent agreement. Calculated temperatures throughout the asphalt concrete surface layer, together with laboratory evaluation of fracture strength have been used to predict the pavement susceptibility to thermally induced fracture. Results of field studies support the analysis technique described.
03005	A Study of Subgrade Moisture Conditions in Connection with the Design of Flexible Pavement Structures T. Y. Chu, W. K. Humphries, S. N. Chen In the structural design of flexible pavements, the supporting power of subgrade soils should be evaluated on the basis of anticipated moisture conditions. It is, therefore, necessary to predict subgrade moisture conditions and to simulate these conditions in performing laboratory tests for subgrade evaluation. In this study, field investigations were conducted at selected sites in South Carolina for determining subgrade moisture variations under existing pavements. Findings from these investigations were utilized In the development of laboratory procedures for treating subgrade soil specimens in order to simulate anticipated moisture conditions. While laboratory investigations reported in this paper were carried out in connection with a particular type of subgrade evaluation test, the procedures developed for specimen treatment may be adapted to other types of test for similar purposes.
03006	The Influence of Climatic Factors on the Structural Design of Flexible Pavements D. Croney, J. N. Bulman Current design standards for flexible pavements in Britain are based on the results of full-scale road experiments extending over 20 years. The influence of seasonal and climatic factors is taken into account by such a protracted period of observation. In extending the use of the standards to other climatic environments or in applying methods of structural analysis to them, the influence of climate on the performance of the various elements of the pavement needs to be appreciated. The first part of the paper considers the influence of climate on the strength of soil foundations. Recent research in the field of soil mechanics is used to isolate three broad subgrade moisture conditions based on rainfall, evapo-transpiration and the position of the water-table. The role of vapour transfer in influencing subgrade strength is also discussed. High temperatures reduce the modulus of the bituminous elements of the pavement and again the effects of this on the stress regime are discussed. Augmented vertical stresses in the subgrade, as well as a loss of stability in the bituminous material may lead to increased deformation under traffic during hot weather. However, the ability of the materials to accept higher levels of strain reduces the risk of cracking under these conditions. Sub-zero temperatures normally lead to a much increased modulus both in bituminous materials and subgrades. The transfer of water into the subgrade during the downward progress of the zero isotherm will however reduce the elastic modulus of the subgrade drastically immediately ofter the thaw. The implications of this on pavement stresses must be considered. Published data have been used to estimate for four soil types and for two granular sub-base and base materials the probable effect on strength of a wide range of climatic conditions. Strengths are expressed in terms of CBR or elastic modulus. Used in conjunction with structural design procedures these values will give some indication of the influence of climate on pavement thickness requirements.
03007	The Classification of Traffic for Pavement Design Purposes E. W. H. Currer, P. D. Thompson 03007.pdf For the structural design of new pavements the information available about traffic is normally restricted to census data obtained from visual counts. Such counts are made in Britain at about 4-yearly intervals on a national basis, but for major projects supplementary studies will normally be required. It has been usual practice to differentiate between private cars, light commercial vehicles (delivery vans of estimated unladen weight not exceeding 1.5 tonnes), heavy commercial vehicles and public service vehicles. The total of heavy commercial and public service vehicles only has been used in the past for structural design purposes. More recently however it has been recognised that the axle-load spectrum of the commercial traffic is of major importance. An electronic weighbridge suitable for installation in the carriageway surface was developed many years ago and effort is now directed towards the production of an alternative capacitor-type pad which can be placed on the carriageway surface. Equipment of the former type has been used to obtain the axle-load spectrum of the traffic on existing roads, but the cost precludes its use for routine survey purposes. The objective of research in this field in Britain has been to determine by limited surveys mean spectra for commercial axles appropriate to different classes of road, e.g. industrial motorways and major trunk roads, rural major roads, secondary rural roads and roads in residential areas. A summary of results to date is given in the paper. Typical spectra obtained in this manner are converted to an equivalent number of standard 8200 kg (18,000 lb) axles per 100 commercial axles, using the A.A. S. H.O. equivalence factors. In using this approach, to evaluate from the commercial traffic flow the cumulative number of equivalent 8200 kg (18,000 lb} axles carried by a particular carriageway lane during the design life of a pavement, two other pieces of information are required:- (1) the average number of axles per commercial vehicle (2) the distribution of commercial traffic between the carriageway lanes. Studies of these two factors have recently been made in Britain and the results are summarised in the paper.
03008	A Contribution to the Establishment of Design Loads for the Thickness Design of Flexible Road Pavements H. Keller Theoretical and empirical research is constantly producing new ideas on flexible pavement design, but these are not being incorporated in the principles used in actual practice for assessing useful design loads. This report is designed to draw attention to research that has been done on the specification of service load data for roads in the Federal Republic of Germany. It also outlines the work that remains to be carried out in the future. It lists initially the requirements that are currently made for loading data for flexible pavement design and indicates possible ways of simplifying data presentation. The report deals with surveys already carried out on service loads: starting with the pattern of stress distribution between truck tires and carriageway surface, it discusses vertical axle loads assessed in permanent surveys as well as sample and indirect surveys. It also deals with the data required on loading periods and loading points. Explanatory measurement examples are cited for various characteristics of truck traffic, Finally, the report takes current loading data requirements and available data as a basis for indicating what still has to be done in flexible pavement design research to establish design criteria for the service loads of roads.
03009	The Transient and Long Term Performance of Pavements in Relation to Temperature N. W. Lister At the present stage of development of a structural method of design applicable to British conditions it is possible to model the transient stress-strain behaviour of several types of flexible pavement with some confidence whereas progress in solving the more complex problem of the long term behaviour of pavement materials and subgrades under repetitive loading is understandably slower. It is therefore of value to try to relate levels of transient stress and strain generated within the road directly to their observed long term behaviour. Most British pavements tend to deteriorate by deformation of the road surface followed by cracking at a relatively late stage. Elastic analysis of a pavement with a rolled asphalt surfacing and base indicates the critical nature of stress-strain conditions at high temperature and these are related to the development of deformation in an experimental road section under traffic, nearly all of which took place under high temperature conditions. Full-scale repeated loading tests carried out in a Road Machine under controlled conditions of wheel load and temperature on a pavement with a crushed stone base are described. The results again demonstrate the critical importance of temperature in determining deformation behaviour and indicate the existence of a critical value of vertical stress in the subgrade above which subgrade deformation occurs.
03010	A Model Utilizing Climatic Factors for Determining Stresses and Deflections in Flexible Pavement Systems Charles R. Marek, Barry J. Dempsey In this paper, a model for determining the stresses and deflections at various locations in a multi-layered flexible pavement system as a function of load, climatic exposure conditions, and material characteristics is described. Application of the model to a flexible pavement section at the AASHO Road Test and comparison of theoretical stresses and deflections with measured stresses and deflections at various times during the life of the section is made. The comparisons relate the applicability of the model for pavement analysis. The model combines a heat-transfer model for evaluating frost conditions and temperature-related effects, with an elastic layer model for stress and deflection determination. The heat-transfer portion of the model was derived from one-dimensional, forward-finite-difference, heat-transfer theory. This portion of the model was used to generate temperature profiles and frost-line locations in the pavement system at a specified time, It was designed to include many input parameters such as short-wave radiation, long-wave radiation, convection, and air temperature. Other factors considered are physical properties and thermal properties of the pavement materials including unit weight, moisture content, material classification, thermal conductivity, heat capacity, and latent heat. The model was developed so that appropriate thermal properties of the pavement materials are used depending on whether or not an unfrozen or frozen state exists. In addition, it can be easily expanded to include newly developed parameters. The model has been programmed for computer solution. The stresses and deflections at various locations in a multi-layered flexible pavement system were determined based on Burmister's elastic layer theory. A computer routine, developed by Chevron Research Corporation and modified to work on the University of Illinois' computer system, was employed for solution. Modular values and layer thicknesses were established based on layer interface location, frost-line location, and layer condition (frozen or unfrozen). Stiffnesses of the asphaltic concrete surface and binder layers were determined by use of the procedure developed by Van der Poel and revised by Heukelom and Clomp. A correction for air content when in excess of three per cent, as suggested by Van Draat and Sommer, was also employed. Average layer stiffnesses for the surface and binder layers were determined and utilized for computation of stresses and deflections in the system at a specified time. Using the model, theoretical stresses and deflections for Section 581 from the AASHO Road Test were computed for the period of July 1, 1959 to June 30, 1960. Comparisons were then made between theoretical and measured stress and deflection values to substantiate the applicability of the model for theoretical analysis of pavement systems.
03011	German Experiences with the Replacement of Granular Frost Blankets by Other Types of Construction H. Proksch During the years 1954 - 1970 numerous research projects and experiments were carried out in Germany in order to investigate by which means granular frost blankets customarily used for 40 years may be substituted by other types of construction. The studies being undertaken involved comparative experimental field projects using asphalt membrane envelopes, full-depth asphalt pavements and bituminous and plastic thermal insulating layers as well as extensive laboratory investigations, The paper describes the experimental projects in detail and reports the results already available of measurements of subsoil moisture content, frost penetration depth, thermal properties of the bituminous and elastic materials used and of load bearing characteristics applying plate bearing tests, dynamic load tests and the Benkelman Beam.
03012	The Effect of Climatic Factors on Benkelman Beam Deflections in the Melbourne Area of Victoria, Australia A. Ratnarajah A programme of testing to evaluate the effects of seasonal climatic variations on pavement deflections as measured by the Benkelman Beam was initiated in March 1969 jointly by the Australian Road Research Board and the Country Roads Board, Victoria. The areas selected for testing were all in the vicinity of Melbourne. In order to minimize the effects of variations in subgrade type, groups of sites in geologically similar areas were selected. Five such areas were chosen. The frequency of testing was approximately monthly for the first twelve months and two monthly thereafter. Results to date indicate that ambient temperature as such has no effect on the subgrade that is measurable in this investigation. Rainfall however has a measurable effect on the subgrade. Its effect on Benkelman Beam deflections is tied in with the effects of the permeability of the subgrade and with the surface drainage conditions at the site. The effects of evaporation of water from the ground surface would be to reduce the amount of water falling on the surface that would otherwise percolate into the pavement and subgrade. No attempt has been made to evaluate this effect. The tests have given some basis on which a correction factor to the observed deflections may be applied in certain situations when the deflection testing is not carried out at the time the subgrade is at its weakest.
03013	Prediction and Observation of the Performance of a Flexible Pavement on an Expansive Clay Subgrade B. G. Richards, R. Gordon For an experimental section of road pavement in the Darling Downs, Queensland, investigations carried out earlier on a highly expansive clay subgrade suggested an ultimate equilibrium subgrade suction of the order of 350 p.s.i. or 6 per cent dry of optimum moisture content (OMC) for standard AASHO compaction. During and after construction, rapid drying in the vicinity of the edges of the sealed pavement caused serious longitudinal shrinkage cracking at the edge of the seal in less than 12 months. Subsequent seasonal rainfall entering these cracks led to serious deformations and loss in surface shape and, in many cases, to shear failure. Practical methods are suggested which may overcome these problems. A laboratory programme was devised to investigate the volume change characteristics and repeated loading 'resilient' moduli as functions of subgrade suction and initial compaction. The results indicate that for an expansive clay subgrade, the volumetric and vertical strain can be linearly correlated with the logarithm of the suction over the suction range considered. The compacted material is also isotropic with the vertical strain being exactly one third of the volumetric strain. The repeated loading 'resilient' moduli are very sensitive to suction and the logarithm of the moduli can also be linearly correlated with the suction logarithm. From the "Shell Rational Method" design charts for lightly trafficked roads, these correlations are found to have a variation of pavement thicknesses from 15 in. at the equilibrium suction to over 4 in. at OMC depending on compaction conditions. Finite element analyses have been made for deformations due to volume changes under changing subgrade moisture conditions and the deformations due to an 18,000 lb axle loading with moisture and stress dependent material properties. These theoretical predictions have been compared with preliminary field observations obtained from the experimental sections. Both the predicted and observed behaviour are in general agreement and clearly indicate the important influence of subgrade moisture on the performance of a pavement built on expansive clay subgrades in a semi-arid environment.
03014	Effects of Environment on Pavement Temperatures R. H. Williamson Climatological considerations are of considerable importance in the design, construction and maintenance of both rigid and flexible pavements. This paper describes some of the deleterious effects of warm environments on road pavements and presents ways in which both empirically obtained long-term temperature data and theoretically evaluated temperature variations in multilayered structures may he related to practical engineering problems. The motivation for, and evaluation of, a simulation model which will predict pavement temperatures using finite difference techniques is presented, and the advantages of using such a model for the acquisition of relevant data on pavement temperatures enumerated. Pertinent areas in which the mechanisms of heat transfer are not well understood are outlined and some existing concepts questioned.
03015	Laboratory Evaluation of Rutting in Base Course Materials Richard D. Barksdale A method is presented for evaluating the relative performance of unstabilized base course materials with respect to rutting and is then used in the evaluation of a number of materials. A general method is also proposed for calculating rut depth occurring in flexible pavements, The proposed methods make use of the plastic axial strains obtained from the repeated load triaxial test. Cylindrical specimens 6 in. in diameter and 12 in. in height of crushed stone and soil-aggregate mixtures were placed in a conventional triaxial cell and subjected to 100,000 load repetitions using a constant confining pressure and a triangular stress pulse. Stress-strain curves giving the relationship between deviator stress, confining pressure and plastic axial strain were constructed for each material studied using the repeated load test results. The concept of a rut index was proposed which can be calculated making use of the plastic stress-strain relationship, and is approximately proportional to the rut depth that will occur in the base after a desired number of load repetitions, The rut index appears to offer a practical laboratory method for evaluating the relative performance of base materials used in pavements having similar structural configurations. An evaluation of the test results using the rut index approach indicates that under good conditions of drainage and proper maintenance of the pavement surface, carefully selected blends of 20 percent soil and 80 percent stone should perform satisfactorily. Soil aggregate blends having properties similar to the materials tested should probably not be used at all under poor drainage conditions, and 40-60 blends should not be used even under good conditions of drainage. The results further indicate that only a sufficient amount of fines should be used in a crushed stone base to permit proper compaction if the amount of rutting in the base is to be minimized. Furthermore, even though the specified gradation and density may be the same, bases constructed from aggregates obtained from different sources may exhibit different rutting characteristics. A general engineering method for estimating the rut depth in a flexible pavement after a desired number of load repetitions was proposed which utilizes nonlinear Layered theory, the plastic stress-strain response of the component materials, and a hyperbolic, plastic stress- strain law. Field verification is now needed of both the proposed rut index and the general method for predicting rut depth.
03016	Stiffness of Pavements Characterized by the Dynamic Modulus Evd - Definition of Dynamic Plate Loading G. Baum The deformation properties of unbound pavement layers are studied by means of static plate loading tests. Attempts to use static plate loading on the surface of bituminous pavement layers failed on account of the properties of viscoelastic bituminous road materials, This paper describes that static plate loading can be replaced by defining a dynamic plate loading test. This is done by substituting sinusoidal forces for the static step forces produced by the load plates. The sinusoidal forces are generated by an unbalance vibration machine and set a road surface in vibrational motion (and/or, depending on the test to be made, any surface of a pavement layer). The generated vibrations are related to various parameters. To characterize the vertical deformations at a point of measurement the theory of simple linear vibration systems is used. The investigations have brought out that only one of the constants derived from the technique of vibration gives sufficient technical information to be a valid test data. This is a spring constant which is applicable to frequencies exceeding 20 to 40 Hz (higher frequency domain). By means of the dynamic theory of the elastic half-space the dynamic deformation value, E(sub vd), is calculated from the spring constant. Pavements are considered as one-layer systems. In addition, by measuring the velocity of surface waves at low frequencies the effective modulus, E(sub 2), of the subgrade of pavements is determined. And by the use of the Burmister theory the effective modulus of elasticity, E(sub 1,b), of the "upper layer" (pavement) with a thickness of h is calculated from E(sub vd) and E(sub 2). The systematic procedure of the investigation is thoroughly described. For lack of space, however, references are used to complement the description. Finally, full particulars on the procedure can be found in the description of 26 test sections of the road experiment "Grunbach" (Germany) especially designed for base course testing.
03017	Fatigue Tests on Pavements by Pulse Generators H. Behr For the study of the Optimum thickness design Of pavements and the selection Of suitable road materials many approaches have been used ranging from purely theoretical calculations to observations of road experiments under heavy test traffic with special trucks. The paper describes the basic considerations which have led to finding still another way by the use of simulative fatigue testing roads in a laboratory. The equipment (pulse generators) installed at the Bundesanstalt fur Strassenwesen (BAST) - Federal Road Research Institute - is explained in detail. The results of an analysis of the pattern of loading produced by truck tires and pulsed loading are compared to prove the similarity of the most important load criteria. Seventeen fatigue tests have been made by means of the pulse generators described in the paper. Each comprised 600,000 to 1,000,000 passages per track (30 cm) on the test sections built. The simulative loadings correspond to moving wheel loads of 3 - 10 Mp (approx. 30 - 100 N). The results obtained hitherto are relative to 9 selected pavements with bases of dry-bound crushed stone and asphalt-coated gravel varying in thickness, as well as to special designs with heat insulating layers of plastic foam. The bearing properties of these structures under dynamic loading are defined and the behaviour of roads subjected to fatigue tests described. Special attention is paid to permanent surface deformations, since, in a way, this characteristic is a very interesting criteria for the description of riding quality and rutting. An attempt is described to establish an approximative law for permanent deformations in relation to the magnitude of loading and the number of load repetitions. The dependency on the load is expressed by deformation coefficients. With the reservations that the laws still have to be placed on more secure foundations by results of further fatigue tests, load equivalents have been calculated. The load equivalents are a first step towards calculating the effect of mixed traffic, i.e., various load quantities in random sequence. In addition the real test acceleration produced by fatigue tests can be determined on the basis of load equivalents relative to a road under traffic. Before more detailed data can be stated, results from further fatigue tests must be waited for. The future program of research is briefly outlined.
03018	Assessing the Properties of Materials for the Structural Design of Pavements J. Bonnot At the present time, knowledge concerning the mechanical properties of materials has not progressed as rapidly as pavement design methods, and the conventional tests used for selecting the composition of pavement mixes are not directly related to the mechanical characteristics affecting their performance in the pavement. The Laboratoires des Ponts et Chaussees have undertaken the development of a series of tests allowing the basic properties of pavement materials to be fully defined. For the gravel treated with hydraulic binders used in the base course, the modulus of elasticity, which is required for calculating the distribution of stresses and strains in the pavement, is determined from the stress-strain curve obtained during direct tensile testing. The values obtained are characteristic of very rigid materials. The risk of fatigue failure is examined by means of fatigue bending tests; the observed behaviour is very different from that of asphaltic materials and this has major consequences on pavement design. The risk of cracking by shrinkage is examined from strain failure under direct tension, or by means of special tests in which the prevented thermal shrinkage process is reproduced in the laboratory. The spacing between the cracks may be estimated from the breaking strength under direct tension. For asphaltic materials, the moduli values necessary for calculating the distribution of strains and stresses in the pavement are given by the complex modulus test. Other modulus values may be obtained through the relaxation test or from stress-strain curves obtained during the direct tensile test. Master curves of these different moduli are compared. The risk of fatigue failure is examined by means of fatigue bending tests with the imposed strain. Failure strength and strain data under direct tension are also given, and this test is compared with the Brazilian test. The tensile and compression test allows the determination of variations in cohesion and internal friction with the loading rate and with temperature. Resistance to rutting in asphaltic materials is also examined by means of creep tests under compression, or by a test based upon the wheel tracking test.
03019	Applied Rheology of Asphalt Mixes - Practical Application G. Chomton, P. J. Valayer The rheological properties of asphalt mixes are usefully studied when they emphasize the durability of a road. This durability is related to: - Stress distribution in the road, for which rheological properties must emphasize "stiffness" of the material. - Failure criteria related to cracks, for which rheological properties must emphasize "fatigue". - Failure criteria related to large deformations for which the rheological properties must emphasize creep under repeated shears. A brief description of the equipment used in Mont Saint Aignan is given. The stiffness is expressed in terms of complex moduli and is related not only to frequency of stress application and temperature but also to mix characteristics such as filler and bitumen content and also grain size distribution. General laws are checked for fatigue and the expression of fatigue versus stiffness is very attractive ; also, new laws are given, involving the energy consumption during a fatigue test. Finally, a vibrating creep test is described and criteria are issued from this test to select the materials which will not develop rutting in the roads. The laboratory results are then processed in terms of road performance by the use of strain distribution calculated from a computer. The modulus is an essential parameter: therefore all the results are presented versus the modulus of the layer examined. Both fatigue and vibrating creep tests are processed this way; the fatigue test by means of the conventional tensile strain at the lower part of the layer considered and the vibrating creep test by means of the maximum shear strains calculated in the layer. Finally, an important discussion is given on the limitations of this method : examples (rotating stress tensors under the moving load) illustrate the criticism. In conclusion, fundamental mechanical testing can be used for the selection of materials on roads : examples of practical applications are given.
03020	Strength of Bases and Subbases Charles R. Foster The results of in-place CBR tests made at various heights above the subgrade are reviewed to determine if the CBR at a given level in an untreated material is a function of the CBR of the underlying material and the height of the test above the underlying layer. The data presented show conclusively that the above postulation is correct; however, material characteristics affect the CBR and a general relationship could not be developed. The data do show that the thin layers of base do not develop significant strength. The author believes the water these thin layers collect and hold weakens the subgrade with the result that a pavement made with a treated base course may be structurally weaker if a thin cushion course is used than if the treated base course had been placed directly on the subgrade. The data also show that the modulus of an untreated base course will increase with thickness. Thus, thickness equivalencies of treated materials which have fairly constant moduli with thickness, will decrease with an increase in thickness of base. This pattern has been reported for the data developed at the AASHO Road Test.
03021	The Behaviour of Bituminous Mixtures in Laboratory Tests and Under Road Conditions R. Guericke, F. Weinert The fatigue behaviour, the expansibility and the stiffness of bituminous mixtures are important parameters for the dimensioning of road structures. After work of about five years informative values for the various bituminous mixtures can be given. The determination of flexibility in the fatigue test has remained problematic. Not sufficiently known are the effects of the frequency of load cycles and the influence of the intervals between the load cycles. As long as these influences are not exactly studied it is recommended to test the fatigue of bituminous mixtures with frequencies of at most 10 c.p.s. (Hz). Moreover, it was found that the results of fatigue tests with constant deformation amplitude are in better agreement with the field conditions than the results of fatigue tests with constant power amplitude. The determination of flexibility in the fatigue test generally requires much time and work. It was for this reason that in parallel with the fatigue tests the stress and strain were determined in axial tensile tests. As can be concluded from the results obtained so far there exists a close connection between the flexibility in the fatigue test (i.e. number of load cycles until rupture at a given deformation) and of the expansibility in the axial tensile test. The axial expansibility at low temperatures is, moreover, an important parameter with a view to the danger of the formation of cracks due to cooling in bituminous road surfacings. In parallel with the laboratory tests the behaviour of road pavements under standing and rolling loads has been investigated. In this connection the determination of the radius of curvature of the deformation trough has proved particularly effective and informative. The curvature meters developed by Franz Muller gives exact and excellently reproducible measured values, it is smaller and easier to handle than the Benkelman beam. One measurement takes only one or few minutes.
03022	Relations between Mix Design and Fatigue Properties of Asphaltic Concrete J. M. Kirk A great deal is known about the fatigue properties of bituminous materials, but this knowledge is based on tests carried out under various conditions. Consequently, it is difficult to obtain a general picture of how the fatigue properties are influenced by the most important parameters such as binder content, void contents, grading, temperature etc. Therefore, in this paper an attempt is made to generalize as far as possible and to isolate the influence of the most important parameters. The first step is to show that the influence of the three parameters, the penetration grade of the binder, the temperature and the loading time can be combined by use of one single parameter, the stiffness of the binder. Then is shown how the fatigue properties of a mix is influenced by this parameter. Next is shown that the fatigue properties improve with increasing the maximum size of the aggregate and how to isolate this factor. The next step is to show the effect of the binder content and how it is influenced by the content of voids, and a correction curve for this effect is shown. Furthermore, it is shown that a mix must contain a certain minimum of filler in order to obtain good fatigue properties. This leaves only one parameter, the shape of the gradation curve, which is of importance in so far as it determines the number of large voids in the mix, and a large number may lead to poor fatigue properties, While the fatigue properties are important in thickness design, when a bituminous base is used, they are only part of the problem in mix design, where other important points as the stiffness of the mix, which determines the load spreading capacity, the stability, the workability and the economy must be considered. This will often fix a limit to how good fatigue properties it is possible to obtain, but it is hoped that this paper will enable the mix designer to get close to that limit, which will allow the thickness designer to calculate with a high permissible strain in the material.
03023	Relationship Between Pavement Structural Integrity and Hardness of the Asphalt Cement Norman W. McLeod In colder climates, it is of questionable value to concentrate a great expenditure of effort on the design of a conventional, deep strength, or full depth asphalt pavement structure, if the integrity of the structure is to be impaired or destroyed by low temperature transverse pavement cracking. Several years of research on the problem of low temperature transverse pavement cracking in Canada, where this is currently the most serious pavement performance problem, have indicated that it can be most easily and inexpensively remedied by using softer grades of asphalt cement. Field and theoretical evidence are presented to support this conclusion. Finally, to bridge the gap between research and practice, a chart is provided to enable an engineer to select a grade of asphalt cement that will preserve the integrity of an asphalt pavement structure by avoiding low temperature transverse pavement cracking throughout the pavement's service life, provided the asphalt pavement has been properly designed and constructed.
03024	Tensile Behavior of Asphalt-Treated Materials Under Repetitive Loading Raymond K. Moore, Thomas W. Kennedy This paper summarizes the findings of an experimental program designed to evaluate the tensile and behavioral characteristics of high quality asphalt-treated materials subjected to repeated tensile stresses by means of the indirect tensile test. The objectives of the study were: (1) to determine whether the indirect tensile test can be used for the study of the behavior of asphalt-treated materials subjected to repeated tensile stresses, (2) to define the general nature of the relationship between applied tensile stress and fatigue life and to evaluate the effect on fatigue life of certain mixture and compaction variables, and (3) to investigate the possibility of estimating the fatigue life of asphalt-treated materials subjected to repeated applications of a tensile stress either by developing a predictive equation or establishing a correlation with other material characteristics. Results of the study indicated that the indirect tensile test can be used satisfactorily to evaluate the fatigue characteristics of high quality asphalt-treated materials under repeated tensile loadings. In addition, the general nature of the relationship between applied tensile stress and the fatigue life was determined along with the inherent variation associated with fatigue life. The tensile stress-log fatigue life relationships were essentially linear with failures occurring at tensile stresses ranging from 8 to 40 psi, which were approximately 6 to 30 percent of the static indirect tensile strength. Significant variation in fatigue life occurred, and it was found that the standard deviation varied linearly with the mean of fatigue life, with the coefficient of variation ranging from 30 percent to in excess of 75 percent. The tensile fatigue characteristics were found to be affected by asphalt content, type of asphalt cement, compaction temperature, and mixing temperature. Within the range tested, it was found that fatigue life was increased by using a more viscous asphalt cement, higher compaction temperature, and higher mixing temperature. It was also concluded that there is an optimum asphalt content for maximum fatigue life. In addition, a simple predictive equation was developed which adequately described the fatigue life of the specimens tested. Fatigue life was found to correlate with initial stiffness, initial tensile strain, and tensile stress-strength ratio, but these correlations were associated with a large amount of variation. No correlation was found to exist between fatigue life and percent air voids.
03025	Applicability of a Linear Viscoelastic Characterization for Asphalt Concrete Keshavan Nair, Wayne S. Smith, Chin-Yung Chang The objective of this investigation was to determine the validity of using a linear viscoelastic constitutive equation to characterize asphalt concrete in the design of pavement systems. The investigation was conducted in two phases. In the first phase, creep tests in compression, tension and torsion, and repeated loading tests were performed on cylindrical specimens of asphalt concrete for various axial loads at different confining pressures and different temperatures to determine the response functions, and establish the degree of linearity and evaluate the time-temperature equivalence of the response. On the basis of these tests creep functions and complex modulus values for the asphalt concrete were determined. The second phase of the investigation was conducted to check if the characterization of the asphalt concrete obtained in the first phase of the program could be used to predict the behaviour of asphalt concrete under stress states which are similar to those that might exist in actual pavements by testing beams and slabs and a Winkler foundation. In this paper only the results of the first phase of the investigation are presented and discussed. It was found that the type of test (e.g. uniaxial, triaxial, torsion) influences the magnitude and nature of the viscoelastic functions. For one type of test, the results obtained are consistent. Based on the results of tests under hydrostatic stress states it was observed that the samples exhibited a substantial degree of anisotropy. It is hypothesized that this is due to the method of compaction utilized in preparation of the samples. It was also observed that the volumetric response of asphalt concrete was tine dependent and that the usual assumption of incompressibility was questionable. Thermorheological simplicity was found to be a satisfactory assumption for uniaxial compression tests. In order to reduce the influence of stress level, stress state, anisotropy and other effects to a level that will make linear isotropic viscoelasticity an acceptable characterization of asphalt concrete, it has been suggested that a limit be placed on the strain level that can occur in the asphalt concrete. This was done using the experimental data on the basis of a subjective evaluation. It is strongly recommended that sensitivity studies on the basis of pavement performance be conducted to determine the acceptable variations for ideal material characteristics and hence establish the degree of refinement required in the characterization of naterials for the design of pavement systems.
03026	Dynamic Structural Properties of Asphalt Pavement Mixtures Charles A. Pagen An approach to the structural design of pavements is advanced, based on the utilization of the in-service mechanical properties of road materials in the computerized viscoelastic and elastic multi-layered theories. Design and performance of roadways are discussed in light of the fundamental structural properties of the component materials. The primary goal of this program was to evaluate the significant changes in the dynamic mechanical properties of asphaltic concrete mixes used in highway pavement systems. This study has quantitatively evaluated the effect of laboratory and roadway compaction on: the rheological creep and dynamic moduli, the elastic properties and the ultimate unconfined compressive strength of asphaltic concrete mixtures. The influence of compaction on the loading time and temperature-dependent strength and deformation properties of asphaltic mixtures was investigated in laboratory experiments on the phenomenological level. Such laboratory and roadway correlations are essential in order to evaluate the influence of construction techniques on pavement performance, asphalt mixture design procedures, thickness equivalencies and the structural design of road and airport pavements. The data indicate that highway test studies are required to supplement laboratory and theoretical research work. The experimental phases of the study involved the testing of three dense asphaltic concrete mixes prepared in the laboratory by gyratory compaction and comparable field core test specimens obtained from a test section in U.S. Route 42 near Delaware, Ohio. Laboratory experiments were performed to validate the application of material science concepts to roadway asphalt mixtures and to determine the limitations of such approaches to highway design and strength evaluation programs. Experiments have quantitatively evaluated the influence of method of compaction, loading time or frequency, age of test specimen and temperature on the failure, and rheological complex and elastic moduli of the materials investigated. The field and laboratory compacted test specimen data have indicated that the application of the linear viscoelastic theory and mechanistic models to asphaltic concrete is valid, as well as that the time-temperature superposition concept is applicable for the research data obtained. Long-term creep experiments and dynamic tests have provided additional and independent checks of the concepts employed. Significant differences in the mechanical properties of the essentially identical laboratory and field compacted test sample were noted for the complex moduli and other strength parameters investigated. Discussions on the utilization of the research procedures for evaluation of pavement thickness equivalencies, and suggestions for potential highway engineering utilization of the results in the structural design of flexible pavement systems, future studies of asphaltic concrete mixture properties, and acceptance of construction work are presented.
03027	Deformations in Asphalt Concrete Wearing Courses Caused by Traffic W. D. O. Paterson The transient and permanent deformations caused in dense-graded asphalt concrete wearing courses by traffic loads have been measured and are discussed in this paper. The study is concerned mainly with wearing courses ranging in thickness from 2 cm to 10 cm and tested on a concrete-based testing track with controlled conditions of load and temperature. A comparative study made on a local motorway provides some correlation to real conditions. The deformations were measured using pairs of induction coils 25mm diameter with a high speed chart recorder for readout. Deformations were measured vertically, longitudinally and transversely to give comprehensive results. A technique of coil placement is described. The asphalt concrete materials used were of high stability and had a fixed dense grading of crushed aggregate and two maximum aggregate sizes of 16mm and 9.5mm. The results of trafficking showed the quantitative effects of load, temperature and pavement thickness on the densification of the asphalt concrete. As the density increased the apparent dynamic modulus and the stability of the mix increased. The apparent dynamic modulus also increased markedly as the pavement thickness decreased towards twice the maximum aggregate size, leading to a quantitative assessment of the role of pavement thickness. Vertical and horizontal strains decreased as the density increased and at lower temperatures vertical strains can be less than middepth horizontal strains. Vertical strains are greatly increased by a rise in temperature while horizontal strains increase only slightly.
03028	The Characteristics of Materials for the Design of Flexible Pavement Structures P. S. Pell, S. F. Brown An important part of the development of a structural design approach to the design of flexible pavements is adequate characterisation of the constituent materials in the context of the part they play in the pavement structure. This characterisation should take the form of the determination of elastic constants and failure criteria so that linear elastic theory may be used to compute critical stresses and strains and the acceptability of these may be assessed in terms of the anticipated life of the pavement. The characteristics of bitumen bound, cement bound, unbound materials and also cohesive soils are discussed against the background of current knowledge. Some results from recent research projects on bituminous materials and soils are briefly presented. The importance of reproducing in situ stress conditions in laboratory tests is emphasised and present test methods are critically reviewed. Some of the more important areas for future research are outlined as a result of this review of the current state of the art. The most important of these is considered to be correlation of laboratory determined results and performance in the road, so that the results of laboratory materials testing can be used with confidence in design.
03029	Mechanical Response of Bituminous Mixture Under Various Loading Conditions Teruo Sugawara This paper describes the test results concerning the mechanical response of bituminous mixtures which are required in rational designing and performance studies of pavement structure. A stress-strain relationship of bituminous mixture was investigated under various loading conditions at various temperatures. A relaxation modulus under the constant rate of strain was computed from the stress-strain curve as a function of loading time. In the flexure test, mode of fracture, relation between flexural strength vs. temperature, strain at rupture vs. temperature were obtained. As a result of beam flexure test, the movement of the transition point at which the mode of fracture shows a change in mode from brittle to ductile, was measured. A convergence of the strain at rupture in the ductile fracture zone (2 to 5 x 0.01), in the brittle fracture zone (1 to 2 x 0.001) and at the transition point (4 to 6 x 0.001) was obtained respectively, throughout this investigation. The effect of the rate of strain, binder content, type of mixture and type of binder on the rheological properties of the mixture was also investigated. At higher road temperatures, the resistance to deformation is one of the primary factors in the pavement performance. In this study, simulated wheel tracking test was carried out. The systematic relationships between the resistance to deformation and moving speed. tire contact pressure, temperature and binder property were obtained: And also the relationship between: the Marshall Stability and the resistance to deformation by the Wheel-tracking test was found. This result may shows a difference between-static and dynamic test on the flow properties of the mixture. A kneading action caused from the wheel passage was briefly discussed. An electro-hydraulic apparatus was developed to investigate the dynamic response and the fatigue properties of the bituminous mixture, and the dynamic response under the programmed strain wave was discussed.
03030	The Fatigue of Bitumen and Bituminous Mixes W. van Dijk, H. Moreaud, A. Quedeville, P. Uge We have carried out laboratory investigations into the phenomenon of fatigue both in bitumen ('asphalt cement') itself and in mixes of mineral aggregate bound with bitumen as used in asphalt pavements. The objectives are to determine the role of the binder and to attempt to establish a method of predicting road performance from laboratory fatigue tests. Our research into the fatigue of bitumen has been carried out on thin film specimens that have been subjected to sinusoidal shearing over a range of temperatures and frequencies. In tests with a constant stress amplitude, the initial strain for a given fatigue life has been found to depend strongly on the complex modulus of the bitumen, and the slope of the fatigue curve (log life against log initial strain) to be related to the phase angle between the stress and the strain. When plotting the slope of the fatigue curve against the phase angle there is a maximum which is more pronounced for bitumens of low temperature susceptibility. The fatigue life of bitumen can be interpreted in terms of energy: the energy dissipated per cycle of stress being related to the fatigue damage per cycle, and the fatigue resistance may be regarded as the capacity to dissipate this energy into heat. Films of filler/bitumen mixtures have also been tested in the same apparatus and it appears that a filler bitumen behaves like an unfilled one of harder grade. Tests have also been carried out on various asphalt mixes, including a dense wearing course type and an economical base course type with a low binder content. The object has been to correlate the fatigue behaviour of the mixes in bending with that of the bitumen films on the one hand, and with the development of cracking in a slab of mix loaded by a rolling wheel on the other. This last test simulates the tri-axial loading that occurs in practice in an asphalt pavement. Two types of loading have been used in the bending tests: sinusoidal to provide a link with the usual mode of laboratory testing, and intermittent to give strains whose wave-shapes are similar to those that have been recorded under moving traffic in asphalt layers of pavements. It has been found that the rest periods between successive groups of waves, corresponding to the spacing between vehicles, give fatigue lives that are longer than those with continuous sinusoidal loading. In this way, realistic data are being developed for incorporation into asphalt pavement design procedures.
03031	A Fundamental Structural Design Procedure for Flexible Pavements S. F. Brown, P. S. Pell The main aim of the conference is to work towards improved methods of structural design for flexible pavements wherein the empirical content of particular methods will be reduced and replaced by analytical methods backed by a knowledge of relevant material properties, details of traffic loading and pavement performance. Such approaches have come to be known as 'rational design methods', though an empirical procedure may in fact be no less 'rational'. 'Improved design methods' or the 'structural design approach' are better descriptions of the aim of this work. At the 1967 conference, the question was asked: how much more research must we do before such a method of design emerges? This paper is an attempt to bring together various research efforts and incorporate them in a structural design approach to flexible pavement design. The approach is based on treating the layered pavement system as a structure in the traditional civil engineering manner and designing it accordingly. A framework for the design procedure is presented and is outlined in a simple flow diagram. The design process involves the consideration of traffic loading, material characteristics under repeated dynamic loads, the computation of load induced stresses and strains and their comparison with the maximum allowable values for the various materials in the structure. An iterative procedure which modifies an initial estimate of layer thicknesses and materials is used in order to produce a pavement which is satisfactory both structurally and economically. The difficulty of applying this method lies in the complexity of the structure, materials, loading, environment and criteria of performance. However, certain simplifications can be made and the procedure presented is based on a simple fundamental approach to the problem. The basis of the method is the availability of computer solutions for the linear elastic stress analysis of the structure. More sophisticated programmes are being developed to deal more realistically with the complexities both of the structure and the materials but as yet sufficient information is not available on the material characteristics and performance criteria to warrant their use. It is the Authors' opinion that linear elastic theory may be used in the manner presented with some confidence and a design example of the procedure is given. One of the aims of the paper is to highlight those areas where more research effort is needed, such as the important aspect of pavement performance and its correlation with allowable stresses, strains and deformations in the constituent materials of the pavement.
03032	Calculation of the Deformations Caused by Vehicles to Flexible Pavements Paolo Ferrari The repeated flexural deformations that vehicles produce in flexible pavements are the most frequent cause of failure of these pavements. This type of rupture is due to a fatigue phenomenon that no asphalt pavement can escape, because of the nature of the materials of which it is constructed, so that even correctly executed pavements crack after a more or less long period. It is therefore important to design flexible pavements in such a way that they can withstand the flexural deformations caused by vehicles for a certain time without cracking. Recent experimental research has made it possible to know the relations existing between the flexural deformations caused to an asphaltic concrete and the number of repetitions of this deformation necessary to bring it to rupture. Starting from this relation, it is possible to forecast the life of a pavement subjected to a certain type of traffic, if the deformations caused by the vehicles are known. The calculation of the deformations is the subject of this paper. For this purpose, the pavement-subgrade system is schematised as a semi-infinite visco-elastic body, whose mechanical parameters can be determined by a special plate bearing test, described in the paper. The results of the calculations made in this paper clearly indicate the effects on the deformations in the surface courses of a bituminous pavement produced by weight, number and arrangement of vehicle axles and the speed of such vehicles. These results therefore make it possible, among other things, to more rationally assess the influence of the type of traffic and properties of materials in designing flexible pavements. They also make It possible to obtain the knowledge of maximum strains in the surface course, which is necessary for an essentially rational design of flexible pavements, considering the fatigue processes as a determinant cause of failures. However in the paper it is pointed that, before this design approach can be fully developed, much more theoretical and experimental work needs to be done, essentially in order to investigate the fatigue behavior of the bituminous concrete under the actual road conditions, where at any point of the surface course a series of strains take place, all different from each other and at irregular time intervals.
03033	Applications of Theory in the Design of Asphalt Pavements F. N. Finn, Keshavan Nair, C. L. Monismith Pavement design, like other aspects of engineering design, requires that engineers have the ability to analyze pavement structures in terms of significant system parameters. Moreover, it is necessary that such analyses incorporate essential features of observed pavement performance and appropriately measured values of the parameters to make the necessary quantitative evaluations required for design. It is generally recognized, however, that the parameters involved together with their interrelationships are complex. In recent years a number of attempts have been made to formulate, in a systematic manner, pavement design systems which bring these factors together as a part of the development of improved methods of pavement design, methods which will have the capabilities to: (1) accommodate the continually changing loading requirements; (2) better utilize available materials; (3) accommodate new materials which might be developed; (4) better define the role of construction; and (5) improve the reliability for performance prediction (or of the design estimate). While it is difficult to develop one pavement design system which, at this stage in time, will incorporate all desirable factors, it is possible to develop a series of subsystems the goal of each of which is to minimize a particular form of distress. To minimize the effects of various distress mechanisms, design frameworks (subsystems) have been established and specific formats are developed. These subsystems parallel the design approach widely used in Civil Engineering practice in which a structure is selected (designed), its behavior under anticipated service conditions analyzed, and its adequacy with respect to a specific distress criterion determined. A discussion of the most recent methods available to examine each of these subsystems is included. Considering this information, the authors recommend a specific technique to solve for each distress mode recognizing present limitations in materials characterization techniques and availability of solutions for boundary value problems representative of pavement structures as well as present limitations in knowledge of traffic, environmental, and construction effects. The concluding section is concerned with a discussion of factors not now included in the subsystems described in the report. Consideration is given to optimization and the applicability of dynamic programming techniques to optimizing the solution of the pavement design problem is discussed. Included in the discussion is the ability of dynamic programming to consider the following factors: (a) Adaptive System - the incorporation of new information gained through observations of performance to predict future performance, (b) Sensitivity - the influence of individual parameters on the total system, and (c) Stochastic Processes - the ability to evaluate the degree of uncertainty in the information that forms the basis for design.
03034	Prediction of the Resilient Response of Pavements Containing Granular Layers Using Non-Linear Elastic Theory R. G Hicks, C. L. Monismith A laboratory-field study to better define those properties of granular base materials which significantly contribute to the resilient response of pavement structures under repetitive loadings has been conducted. Results of the laboratory study, previously reported by the authors, indicated that the stress deformation properties of granular materials when measured in repeated load triaxial compression are non-linear with their modulus values and Poisson's ratios dependent on stress state to a considerable degree. Field studies included measurements of the response of a prototype pavement, 20 ft. by 20 ft. in plan, to repeated-load plate tests and of an in-service pavement in San Diego County, California to an actual truck load. Both pavements consisted of an asphalt concrete surface, a granular base course, and a clay subgrade. Instruments were placed at various positions in the pavement structures to measure deflections, stresses and strains. For the prototype pavement, the plate load tests were conducted at the pavement surface with the base course in both a partially saturated and saturated state. Loads were applied through rigid plates at durations representative of moving traffic and in sufficient numbers to assure that a reasonable measure of the resilient behavior could be obtained. These responses were then compared with computed values which were obtained from three different non-linear elastic analyses and laboratory determined stress-strain properties. The results indicated that the predicted stresses, strains and displacements compared reasonably well with field measurements; the responses yielded by the different numerical methods were similar; and the resilient responses of the system were only slightly affected by degree of saturation of the base course. The approach developed to predict pavement responses for the prototype pavement was extended to a full-scale test road in San Diego County, California, to verify its application to conditions representative of moving traffic on highways. Measurements of strains and displacements were obtained under a slow moving truck and compared with predicted values. As with the prototype pavement, results of the analyses indicated that the predicted and measured responses were in good agreement.
03035	The Modulus of Asphalt Layers at High Temperatures: Comparison of Laboratory Measurements Under Simulated Traffic Conditions with Theory A. Hofstra, C. P. Valkering In order to investigate whether elastic theory can be applied to the behaviour of flexible pavements under moving wheels, particularly at high temperatures, we have measured in our Laboratory Test Track the stresses and strains imposed by a rolling wheel. The test pavements consisted of an asphalt layer on top of a sand subgrade; the temperature of the asphalt layer has been varied between 20 and 60°C. The results have been compared with the theory. Apart from a certain degree of asymmetry of mainly the strain signals, the shapes of the experimental signals correspond well with those predicted by theory. The asymmetry has been found to be little dependent on temperature. The maximum values of the stresses and strains are also compared with those predicted by theory. In this comparison, for the asphalt layer a Young modulus is used which is dependent on the temperature of the asphalt layer and on the width of the signal (or the loading time). In this way the influence of wheel speed and penetration grade of the binder on the maximum strains as well as the difference in magnitude between the longitudinal and transverse strains in the wheeltrack could be explained. The comparison of the experimental results with theory indicates a condition of the interface between the asphalt layer and sand subgrade which is intermediate between slip end complete friction. It is concluded that elastic theory can be used to describe the behaviour of flexible pavements subjected to moving wheels even at the highest temperatures that are likely to be experienced in pavements.
03036	A Comparison of Plate Load Testing with the Wave Propagation Technique D. A. Kasianchuk, G. H. Argue The repetitive static plate load test method has been used by the Canadian Department of Transport since its development in 1947 for the design and evaluation of the airport pavements within its jurisdiction. The vibratory loading wave propagation technique, of more recent development, offers some advantage over the plate load test in several aspects of the testing procedures. In order to determine the degree of correlation that could be expected between these two methods in the evaluation of the load carrying capacity of pavements at Canadian airports, a pilot test program was conducted during the summer of 1970. Thirty-five vibratory tests were performed using the Shell Canada Limited equipment and technique at six Canadian airports. The test sites were selected from among those included in the annual plate load test program to represent a wide variety of the asphalt pavement types found in Canadian airports. The wave propagation measurements were analysed to provide values of elastic modulus to be used in the prediction of pavement response by layered system elastic theory. These predicted values were compared to those measured at each site in the actual plate load test. Although the comparison obtained indicates some relationship between these approaches, further work is required to more accurately assess the predictability of plate load tests from vibratory test results.
03037	Design of Flexible Pavements for Major Highways A. M. Krivissky On the basis of comprehensive theoretic and experimental research carried out in the Soviet Union a method has been developed for the design of flexible pavements with asphalt concrete or similar surfacing. Structural design of this type of pavements should provide for their performance to be only in the phase of recovery (elastic strain). Special investigations have shown that evaluation of stress-strain condition of such pavements can be obtained by applying solutions of elasticity theory for lmred semi-infinite space. The following physically real conditions are used in design as limiting criteria: - Shear limit equilibrium in subgrade soil or in moderately cohesive materials of structural layers; - Maximum allowable bending-tensile stress in monolithic layers (asphalt concrete, cement-bound materials) for given loading conditions. The design procedure involves both strain (E,mu ) and strength characteristics of materials and soils the latter being shear resistance of soils and moderately cohesive materials (Y, C) and also bending-tensile strength of monolithic materials (R). Design characteristics are normal and can be obtained by means of testing samples, their state simulating the condition of materials or soils in pavement structure. Although the design procedure includes application of relationships which are rather complex for calculation it has been found possible to draw nomograms that bring complicated calculations to simple graphical solutions. The suggested method makes it possible to design not only a total pavement thickness, but also a required thickness of each structural layer, with detail consideration of properties of the materials used. The new structural method is well founded by the evidence of good performance of pavements under various climatic and service conditions.
03038	The In Situ Determination of the Elastic Moduli of Layered Pavements Using SH-Wave Propagation M. Kurzeme A method is presented of determining the in situ elastic properties of layered pavement materials through the excitation and observation of horizontally polarized shear waves (SH-waves). The method consists of generating continuous SH-waves at the frequency and the location desired, detecting and observing the wave motion over some distance from the wave generator, measuring the surface velocity of phase propagation at each imposed frequency, and interpreting the observed variation in surface phase velocity with change in imposed frequency of vibration to deduce the in situ elastic properties of the pavement materials. Theoretical dispersion relationships have been developed that describe the variation in surface phase velocity with change in imposed frequency of SH-wave motion in idealized layered pavement structures. The existence of modes of propagation predicted by the theoretical dispersion relationships were investigated on laboratory models of layered pavement structures. These tests confirmed the presence of the predicted modes. The propagation of SH-waves was then investigated on real pavement structures. A torsional vibrator, capable of generating predominantly SH-waves, was constructed and a phase velocity measuring system and a field procedure were developed. The relationships between the surface phase velocity of SH-waves and the imposed frequency as observed on a number of two- and three-layer pavements are presented. To deduce the shear wave velocities within the individual layer materials, the observed relationships have been interpreted by matching with the theoretical dispersion relationships of comparable idealized structures. A knowledge of the material densities allows the calculation of dynamic shear modulus of the materials. Using the theoretical relationships presented, it is concluded that in favourable cases observations on real pavements can be interpreted, to deduce the in situ dynamic shear moduli of the layer materials. The relevance of the material properties so obtained is dependent on how closely the theoretical model used approximates to the real structure.
03039	A Study of Stress and Strain in the Asphalt Pavement of Tomei Highway Yuji Miura The Tomei Highway of 346 km. was constructed in 1969, between Tokyo and Nagoya, serving for the heavy traffic of more than 25,000 vehicles per day, Having been designed by the CBR method adopting the Structural Number and Layer Equivalency that were the results of AASHO Road Test, the whole distance was paved with asphalt concrete of 10 to 15 cm. thickness, that included surface and binder courses, over asphalt stabilized base course of 15 to 22 cm. thickness laid on the various type and thickness of subbase. On the highway, 17 observation spots were settled, and a lot of electric resistance strain gauge, pressure cells, electrical deformation devices and thermocouples were laid into the pavement. The stress and strain of the pavement, including those in certain depth of the subgrade, were observed in the field under the controlled wheel loads. The purpose of this investigation is to know the behavior of the pavement and to obtain the data for rational design of asphalt pavement structure, In this paper, three out of 17 observation spots are taken for the study, and not only the details of the paving and measuring techniques but also the properties of the materials, that have been determined both in the field and in the laboratory, are elucidated. On the other hand, the measured values of stress and strain under the wheel loads are analyzed minutely, the measured data being compared with computed values according to the multi-layered elastic theory. The comparisons are made by the following two methods: 1) Assuming that each pavement is composed of three elastic layers and loaded with single circular load, the elastic modulus of each layers is arranged and determined so as to be equal approximately to the computed and measured values of stress and strain of the pavement, and then compared with the measured modulus taken in the field or in the laboratory. 2) Assuming that each pavement is composed of four or five elastic layers and loaded with dual circular load, the stress and strain of the pavement are calculated by using the elastic modulus obtained from the field measurement or the past data, and compared with the measured values, especially with the lateral strain of the asphalt treated layers. In addition, the results of the investigations and theoretical analysis are discussed concerning the center, longitudinal and traverse position of the loads. From the comparisons, made by the said two methods, some conclusions are drawn as for the propriety of analysis of the pavement structure by using the multi-layered elastic theory, and that the theory is able to predict not only the distribution of vertical stress and displacement of the pavement but also the distribution of horizontal strain in the asphalt treated layers under the traffic loads.
03040	A Stochastic Approach to Analysis and Design of Highway Pavements F. Moavenzadeh, J. F. Elliott A stochastic approach is proposed for the development of a rational method of analysis and design for flexible pavement structures. The approach utilizes a three-layer viscoelastic model, a cumulative damage theory, and systems simulation techniques. In this approach the environmental variables, properties of layered materials, geometry of the pavement structure, and loading variables are described in a stochastic manner utilizing the Monte Carlo simulation techniques. Through the use of these three interrelated models, the pavement structure and its structural integrity is simulated continuously throughout its design life. The results of the study is used to demonstrate (for a given pavement structure), the nature of damage, the manner in which damage accumulated, and the probable time that damage accumulation exceeds the allowable limits and the pavement requires reconstruction. For the design, several possible alternatives can be simulated, and the one which has the most desirable characteristics can be chosen.
03041	Applications of Computer Codes to the Analysis of Flexible Pavements Raman Pichumani The capabilities and limitations of three computer codes suitable for static analysis of flexible pavement structures are presented in this paper. One of these codes, WIL67, is used for the analysis of single-wheel loading; the other two, BISTRO and AFPAV, can analyze the pavement response due to the multiple wheels of very heavy landing gears of modern jumbo jets such as the U.S. Air Force transport aircraft C-5A and the Boeing 747 commercial aircraft. The WIL67 and AFPAV codes are based on finite element structural analysis technique; the BISTRO code makes use of Burmister's layered-system concept. The BISTRO code, developed by the Shell Oil Company, has made it easy to apply Burmister's general theory for calculating stresses, strains, and displacements at any point in a linear elastic multilayered pavement system due to normal surface loads. This computer program predicts the pavement response due to one load at a time at a designated point in any layer and then uses the principle of superposition to determine the multiple load response. The WIL67 code, developed at the University of California at Berkeley, is a two-dimensional finite element program which solves the layered-pavement problem by treating the pavement as an axisymmetric solid subjected to an axisymmetric load. Although there is a version of this code which can analyze the effects of multiple loads using the superposition principle, the program studied in this research effort is restricted to the analysis of a single wheel load. The AFPAV code is an extended two-dimensional finite element program capable of analyzing prismatic solids. This program idealizes the layered pavement system as a layered prismatic structure and expresses the applied loading as well as the resulting displacements in terms of Fourier series. Therefore, the accuracy of the solution and the computer time required by this code is governed by the number of Fourier terms used. However, unlike the BISTRO code, the linear elastic response of the pavement system due to multiple wheels can be determined by this code in one step without recourse to the superposition principle. Therefore, the AFPAV code is far more efficient and economical than the BISTRO code for analyzing layered pavements loaded by multiple wheels, particularly with increasing number of layers, and increasing number of points where stresses and displacements are required. Theoretical pavement response predicted by the AFPAV code is compared with field data from a full-scale flexible pavement test section which was constructed by the U.S. Army Engineers Waterways Experiment Station (WES) at Vicksburg, Mississippi and statically loaded by a 12-wheel assembly of a C-5A landing gear. The dependence of the pavement response on the elastic constants of various layers is demonstrated by a parametric study, thus emphasizing the importance of accurately determining these parameters. In conclusion, it is shown that it is necessary to continually update the capabilities of the AFFAV code to consider the more realistic characterizations of the pavement layers as they become available.
03042	Experimental and Theoretical Studies of Pavement Behaviour Under Vehicular Loading in Relation to Elastic Theory E. N. Thrower, N. W. Lister, J. F. Potter The paper describes a series of experiments to determine the degree to which multi-layer elastic theory can be used to predict stresses, strains and deflections of pavements under moving vehicles. Measurements of stress and deflection in flexible pavements of realistic design have been made under pilot-scale conditions over a wide range of wheel loads, speeds and temperatures, and similar tests have also been carried out on a simple concrete slab. Results are presented showing the effects of temperature and wheel load on the measured behaviour and this is compared with that predicted from multi-layer elastic theory using data derived from both laboratory and in-situ testing. The results show that elastic theory can be used to predict the dynamic behaviour of relatively stiff pavements, but that deviations occur under high temperature conditions, where it is difficult to establish an effective modulus for bituminous materials because of the rapid change in dynamic modulus with the timescale of the loading, and for pavements which derive a large part of their-structural stiffness from granular materials.
03043	Some Considerations on the Theoretical Estimation for Deflection of Pavement Structures K. Ueshita, T. Arakawa, Y. Watanabe Empirical relations between surface deflection and thickness factor of pavement structures show that the effect of thickness of pavement to reduce deflection is more than the one estimated from the theory of elasticity. Extreme case is shown by the soil-cement/subgrade two-layer system which can never been explained by the theory of elasticity. The authors tried to explain these phenomena by using the finite element analyses assuming non-linear elastic properties for subgrade materials. By these computations, experimental relations of the soil-cement pavement could be almost explained. Besides, a useful equation was introduced to formulate the relationship of deflection versus thickness of non-linear elastic pavements.
03044	Effects of Multiple Wheel Systems and Horizontal Surface Loads on Pavement Structures C. P. Valkering The theory of elasticity has been applied to layered systems representing flexible pavements. The distributions and the maximum values of the tensile strain in the asphalt layer as the compressive strain in the subgrade under multiple-wheel systems have been calculated. The concerted action of the wheels depends, not only on the wheel spacing, but also to a significant degree on the structure which renders weighting traffic for design purposes difficult. A thickness design method based on these strains as criteria is equivalent to one that is based on shear or deformation energy criteria, provided that the permissible values are obtained at the same stress condition as prevails in the pavement. Under normal end tangential surface forces the influence of asphalt layer thickness on the normal and shear stresses at the asphalt/base interface has been investigated: the stresses decrease with increasing thickness. At high asphalt temperatures the shear stresses might be critical for the adhesion between, particularly, a thin asphalt layer and a stiff base.
03045	Design of Full-Depth Asphalt Airfield Pavements M. W. Witczak A theoretical design procedure for Full-Depth asphalt concrete airfield pavements is presented. The design is based upon the use of multilayered elastic theory and utilizes the concept of limiting strains to prevent repetitive permanent deformation and/or shear failure within the subgrade layer and repetitive load cracking within the asphalt bound layer. Development of the allowable strain levels for both failure modes are presented. The method utilized to obtain limiting strains associated with the subgrade was to theoretically analyze flexible (granular base) pavement thickness requirements as defined by the newly revised U.S.A.C.E. thickness design method. These revisions are related to changes in thickness requirements as well as load repetition effect upon thickness from the previous method. Results indicate that a limiting vartical strain of 1460 microinches per inch, evaluated at a limiting asphalt concrete modulus of 100,000 psi is capable of withstanding 1,000,000 strain repetitions. Limiting strains associated with the asphalt bound layer have been established from results developed by Kingham from Full-Depth asphalt concrete pavements of the AASHO Road Test study. An allowable tensile strain of 76 microinches per inch will allow 1,000,000 repetitions when evaluated at a critical asphalt concrete modulus of 1,450,000 psi. Because of the extreme dependency of the stress and strain distributive characteristics of thick Full-Depth asphalt pavements to temperature, monthly cumulative damage techniques were used to develop thickness adjustment factors (TsubF) to adjust thickness requirements for both failure modes due to differing environments. The limiting or critical modulus utilized in the vertical subgrade strain analysis is related to an average annual air temperature of 75°F. For cooler environments, a thickness reduction may be used. The maximum suggested reduction proposed is 10% (TsubP = 0.90) for environments having a 50°F average annual air temperature. Maximum percentage thickness reductions of 13% are suggested for the tensile asphalt concrete strain analysis. This is equivalent to a maximum TsubF value of 1.00 at 40°F (average annual air temperature) and a minimum TsubF value of 0.87 for 60°F environments.
03046	Sensitivity Analysis of Various Cost Elements in Flexible Pavement Design Eldon J. Yoder, Farideh Ramjerdi, William L. Grecco This paper presents a method for predicting the optimum initial service life and optimum periods of resurfacing for flexible pavements. The method is based on consideration of total pavement costs including the cost of initial construction, routine maintenance and major maintenance and increased road user costs resulting from the maintenance operations. Standard economic analyses techniques were used for determining the average annual cost of alternate designs. A modification of the Radzikowski model was used for estimating routine maintenance cost of flexible highway pavements. The pavement design method developed by the Corps of Engineers was utilized in estimating initial design as well as required major maintenance (resurfacing). A method was developed which presented an estimation of road user costs due to maintenance and resurfacing operations. Variables evaluated in this paper included: (1) subgrade type; (2) initial traffic volume; (3) rate of traffic growth; and (4) rate of interest on the investment. Solutions were made for both 2-lane and divided 4-lane highways. The results of the study are presented in the form of graphs which indicate the initial design period which results in least cost for combinations of the variables given.
03047	Low-Temperature Pavement Cracking Studies in Canada Roads & Transportation Association of Canada, Soils & Materials Committee, and Pavement Design & Evaluation Committee Cracking of pavements at low temperatures has become a serious and extensive problem in Canada. During the past decade, various highway departments, producers and others have devoted considerable effort towards finding a solution. This paper summarizes their progress in finding the causes of the problem and their development of some practical, engineering solutions. Field inventories were initially conducted to determine the nature and extent of the cracking. From these, and other observations, the bituminous component seemed in most cases to be the major variable. Subsequent field sampling, plus full-scale experiments such as the Ste. Anne Test Road in Manitoba, the Alberta Test Road, the Saskatchewan Test Road and the Arkona Test Road in Ontario, and laboratory investigations confirmed that certain asphalt cements were primarily involved. Several design approaches were formulated from these findings, the earliest being that of modified specifications. Later, asphalt and mix stiffness or strain limits, fracture temperature calculation procedures, and most recently a cracking frequency estimation technique, were developed as design guides. The paper demonstrates that while this progress has been significant, some key aspects of the problem remain. One of these concerns the treatment of the many miles of existing, cracked pavements.
03048	The Fatigue of Flexible Pavements Bonner S. Coffman. George J. Ilves, William F. Edwards Five asphaltic concrete pavements, each of which measured 20x30 ft. and contained two test areas, were constructed on a 48" compacted clay subgrade. Certain of these pavement areas were fatigue loaded through concentric rings composed of truck tire rubber by superimposing one dynamic 10 Hz haversine pulse on a small static load every second to simulate a continuous line of wheel loads traveling in identical wheel paths 50' apart at 40 MPH. Surface tangential strain, surface deflection and temperature sensors were placed at a number of radii from the load plate centerline and recorded periodically throughout each test. The fatigue of four test areas, as evidenced by visible cracking, was observed closely and noted in a log book. Asphalt and compacted subgrade samples were returned to the laboratory for the determination of structural strength and physical properties. The results of these tests were entered into the Chevron n-layer program along with a number of hypothetical moduli for the natural subgrade underlying these layers. Theoretical strain and deflection profiles obtained from these calculations were compared to measured profiles to determine the best average apparent modulus for this semi-infinite layer. Trapezoidal specimens were sawed from asphalt pavement blocks and fatigue loaded with one 10 Hz haversine pulse per second over a wide range of temperatures and strain levels. An equation relating these quantities to fatigue life was developed and coupled with a theoretical pavement fatigue model to predict the time of initial surface cracking on the four test areas where visible cracks were observed.
03049	Failure Criteria for Flexible Pavements D. Croney The term 'design life' when applied to a road pavement implies a terminal or 'failure' condition beyond which the performance of the pavement will be regarded as unacceptable. For design procedures based on past experience a relatively loose definition of failure has been acceptable, but with the growing interest in structural design procedures, failure criteria expressed in more exact physical terms are essential. This paper discusses the definition of the 'failure' condition which has been accepted for flexible pavements in Britain. The 'critical' condition at which overlaying to extend the life of the pavement should be carried out, is also considered. Both these performance criteria have, for flexible pavements, been expressed in terms of permanent deformation either expressed as a rut-depth or as total deformation from the original pavement level. Observations made on normal in-service roads and on closely observed experimental roads have shown that the criteria ore not markedly different for pavements using lean concrete, bituminous macadam or unbound stone bases. The Present Serviceability Index concept is not regarded in Britain as very applicable to the structural design problem because of its heavy dependence on riding quality factors, not necessarily associated with traffic stresses. However an approximate correlation between the British approach and P.S.I. values is given.
03050	Permanent Deformation of Flexible Pavements Under Simulated Road Traffic Conditions A. Hofstra, A. J. G. Klomp In a circular laboratory test track the rutting of flexible pavements has been studied at various temperatures under well-controlled conditions. It has been found that temperature has a great influence on the depth of rutting, the increase in rutting over a temperature range of 20°C-60°C being much larger than the increase in calculated elastic deformation. The permanent deformation per wheel passage correlates with the stiffness of the asphalt binder used and decreases with increasing number of wheel passages. An increase in thickness of the asphalt layer leads to a distinct reduction in subgrade deformation. The change in thickness through rutting is not larger for a 20 cm than for a 10 cm asphalt layer. Proper mix design proves to be an important factor in relation with permanent deformation.
03051	Strain and Curvature as Factors for Predicting Pavement Fatigue Y. H. Huang The fatigue cracking of asphalt pavements is caused by the repeated applications of excessive tensile strains in the asphalt-bound layer. To predict fatigue, it is necessary to determine the maximum tensile strain at the bottom of the asphalt-bound layers. Two methods are suggested for determining the maximum tensile strain, based on the two-layer elastic theory. The use of two-layer theory, instead of the conventional three- or multiple-layer theory, is based on the fact that the tensile strains at the bottom of the asphalt-bound layer depend on the property of the asphalt-bound layer relative to that of the underlying layers. Any multi-layer systems can thus be reduced to a two-layer system, if an average modulus of elasticity is used to represent the combined effect of all the underlying layers, Fortunately, the moduli of untreated granular materials and soils generally fall within narrow ranges, and typical values can usually be assumed. To facilitate the application of the two methods, simple charts are presented for determining the maximum tensile strain under a set of dual tires. The first method, which can be used for pavement design, requires a knowledge of the elastic moduli of both layers and the thickness of the asphalt-bound layer. By entering these variables into the chart, the maximum tensile strain can be easily determined. The second method, which can be used for pavement evaluation, requires the measurement of curvature on the pavement surface. Knowing the curvature, the modulus ratio, and the thickness of the asphalt-bound layer, the maximum tensile strain can be determined from the charts. It was found that for asphalt- bound layers of 4 in. thick or more the curvature-tensile strain ratio is practically independent of the modulus ratio. By simply measuring the curvature on the surface, the maximum tensile strain can be estimated, and the adequacy of the pavement to withstand fatigue evaluated. Examples are given to illustrate the use of these charts for pavement design and evaluation.
03052	A Design System for Minimizing Fatigue, Permanent Deformation and Shrinkage Fracture Distress of Asphalt Pavements D. A. Kasianchuk, R. L. Terrel, R. C. G. Haas The design of asphalt pavements includes three major structural subsystems: load-associated fracture, load-associated permanent deformation, and shrinkage fracture. In order to accelerate progress towards a rational design system it is desirable to accelerate improvements in the technology of these three subsystem areas. This paper suggests a series of needed research and development tasks for each design subsystem. In addition, some discussion is devoted to briefly justifying the recommendations and to summarizing the current state of design knowledge for the three areas. The paper includes in flow chart form the interrelationships between subsystems and attempts to place the subsystems within an overall pavement design and management framework.
03053	Failure Criteria Developed from AASHO Road Test Data R. Ian Kingham Theoretical models of pavement deformation behavior such as elastic-layered theory can only be used for design purposes when failure criteria are specified. Although such models can be used to predict stress and strain states, they in no way indicate whether the material in the pavement can withstand the predicted deformations. For elastic-layered theory, limiting values of strain or stress need to be defined before the theory can be used to assist practicing engineers in the design of asphalt pavements. There is general agreement in the literature that horizontal tensile stress or strain at the bottom of a thick asphalt layer is the controlling criterion for design to prevent repetitive load cracking. Although such strains were not measured at the bottom of the asphalt layer at the AASHO Road Test, they can be inferred from a knowledge of the material characteristics and the measured deflections. Repetitive load cracking was observed to be the predominant mechanism of initial failure at the Road Test. Since the bituminous base sections provided a complete range of performance, from failures to survivors of over 1 million load repetitions it was possible to describe the strain history of these test sections in terms of performance. The bituminous base sections fell into three performance classifications, depending upon whether they failed the first spring of testing, survived the testing period with a low serviceability rating or survived the testing without any change in serviceability. The horizontal tensile strain, horizontal tensile stress and vertical strain on top of the subgrade data were computed for each test section in each performance classification. Asphalt moduli for a wide spectrum of deflection measurements were input into the stress and strain computations. Moduli values were determined from dynamic loading in compression. Subgrade moduli were inferred from the deflection measurements. The results of the elastic-layered computations showed that there were indeed large differences in horizontal tensile strain, horizontal tensile stress and vertical strain in the subgrade, depending upon the performance classification. Secondly, the level of strain or stress for each performance classification was a function of the asphalt base stiffness at the asphalt layer bottom. From the horizontal strain results it was apparent that asphalt pavements can tolerate higher strains at lower stiffnesses. The horizontal tensile strain and stress relationships with asphalt stiffness were converted into "load repetition to failure" relationships by relating two performance classifications to the number of load repetitions to failure. A log-log relationship was assumed. The resulting family of "fatigue-like" curves for a range of asphalt stiffnesses has been used by Witczak and is the subject of another paper to this conference.
03054	Deflection Criteria for Asphalt Pavements Sadao Nagumo, Minoru Tsukinari, Seiichi Tanimoto The structural design of asphalt pavements in Japan is based on the CBR method, but recently attempts have been m