1st Conference

Performance of Flexible Pavement
W. N. Carey, Jr., Assistant Director for Research
Paul E. Irick, Research Statistician, Highway Research Board, AASHO Road Test Staff

Performance of Treated and Untreated Aggregate Bases
A. C. Benkelman, Flexible Pavement Research Engineer
Harold M. Schmitt, Assistant Flexible Pavement Engineer, Highway Research Board, Aasho Road Test Staff
R. Ian Kingham, Engineer-Observer, Canadian Good Roads Association, AASHO Road Test Staff

Application of AASHO Road Test Results to the Design of Flexible Pavement Structures
Wallace J. Liddle, U.S. Bureau of Public Roads

Thickness Design Relationships for Asphalt Pavements
F. N. Finn and J. F. Shook, The Asphalt Institute, College Park, Maryland

Analysis of AASHO Test Data by The Asphalt Institute
L. J. Painter, The Asphalt Institute, College Park, Maryland

The pavement serviceability-performance concept used to analyze Road Test data, is accepted as the basis of the design procedures that have been evolved. Also, the equations, that have been reported as the findings of the Road Test are accepted and used without modification as to basic form. It has been necessary, however, to make certain assumptions in applying the Road Test equations to mixed traffic conditions and to those situations where soils, materials, and climate differ from those that prevailed at the test site. Each adjustment is explained as it is encountered in the development of the design procedure.

The primary and first stated objective of the AASHO Road Test was:
“To determine the significant relationships between the number of repetitions of specified axle loads of different magnitude and arrangement and the performance of different thicknesses of uniformly designed and constructed asphaltic concrete . . . surfaces on different thicknesses of bases and subbases when on a basement soil of known characteristics.”

Another objective was:
“To develop instrumentation, test procedures, data, charts, graphs, and formulas, which will reflect the capabilities of the various test sections; and which will be helpful in future highway design, in the evaluation of the load-carrying capabilities of existing highways and in determining the most promising areas for further highway research.”

The basic relationship between design and load applications was derived from AASHO Road Test data. Extension of the basic relationship required consideration of the scatter in the data and the properties of the relationship in the region beyond Road Test experience. Current state highway design methods were relied upon to verify the extrapolations presented.

It was shown that the structural components (surfacing, base and subbase) could be treated as a linear combination of equivalent thicknesses of each layer. It was concluded that one inch of asphalt concrete surfacing or asphalt base was equal to two inches of a good crushed stone base and 2.67 inches of subbase.

Subgrade support was evaluated with the laboratory-soaked CBR test procedure. It was demonstrated that the AASHO Road Test subgrade soil had a CBR value of 2.5. Extension to include other CBR values was made using the current Asphalt Institute design method.

Mixed traffic was evaluated in terms of equivalent applications of an 18-kip single-axle load. Conversion factors were developed from the basic relationship derived from the AASHO Road Test data.

01008

A moderated discussion of the papers presened in Session II.

British Full-Scale Pavement Design Experiments
A. R. Lee and D. Croney
Road Research Laboratory, Harmondsworth, Middlesex, England

Pavement Evaluation Studies in Canada
Special Committee on Pavement Design and Evaluation,
E. B. Wilkins, Chairman, Canadian Good Roads Association, Ottawa 4, Canada

Michigan Pavement Performance Study for Design Control and Serviceability Rating
W. S. Housel, The University of Michigan, Ann Arbor, Michigan

Controlled Tests of Mixed Loads on Flexible Pavements
A. A. Maxwell, R. G. Ahlvin, and D. N. Brown
U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi

An Alternate Analysis of the Present Serviceability Index
L. J. Painter, The Asphalt Institute, College Park, Maryland

The procedure followed has been similar in all cases. Each site is used to investigate the performance of one or more base materials under various types and thicknesses of bituminous surfacing. A series of short sections of length 100-200 feet long are laid, in which the thicknesses of both base and surfacing are varied above and below the values expected from previous experience to be satisfactory. The performance of the sections is judged primarily on the deformation caused by traffic, and the degree of cracking which develops.

The paper summarizes the results from five of the larger experiments at present under observation. The main conclusions reached are as follows:

(1) Pavements surfaced with rolled asphalt wearing courses deformed less than equivalent pavements surfaced with open-textured coated macadam. In this respect 1-1/2 inches of rolled asphalt was equivalent to about 4 inches of coated macadam when laid on unbound or lightly bound bases.

(2) Rolled asphalt bases performed better than unbound, cement-bound or open-textured coated macadam bases. For heavily-trafficked trunk roads, a thickness of base of at least 8 inches was required when any of the latter base materials was used under a 4-inch asphalt surfacing. The thickness of base could be reduced to about 5 inches when rolled asphalt was used as the base material.

(3) With bound bases beneath an asphalt surfacing the over-all thickness of pavement required was less than the thickness indicated by the C.B.R. method of pavement design. Maintenance at periods of about four years was found to be necessary on pavements surfaced with open-textured bituminous materials, even where the over-all thickness of pavement was greater than that indicated by the C.B.R. method.

The performance of soil cement, lean concrete and bituminous-bound bases appears to depend critically on the grading of the aggregate and on the binder content, and particularly on the way in which these factors determine the density of the compacted materials. Insufficient evidence has so far been obtained on this matter, and further experiments are now in progress to investigate it further.

Part I contains an outline of the pavement investigation program. It describes the pavement design problem in Canada, the formation of the Committee and the objectives of the Committee’s work. It indicates the philosophy and approach adopted in the investigations, wherein the basic data is developed through a study of existing rigid and flexible pavements in service on the main highway systems. The three-stage investigation plan, together with the standardized procedures for data collection, testing and data processing, are described. The unique features of the program which include the method of rating pavement performance and the method of evaluating the strength of a section of pavement are presented in some detail.

Part II of the paper deals with pavement deflection testing procedures. The standard CGRA Benkelman Beam rebound deflection testing procedure which is used in all Canadian flexible pavement investigations is described. Data supporting this procedure is discussed and significant factors to be considered in carrying out such tests are noted. Several modifications to the standard Benkelman Beam are described. Both theoretical considerations and practical field data are used to substantiate that the CGRA Benkelman Beam rebound deflection test results are a valid measurement of pavement strength. Data are presented to show the influence of wheel load and pavement temperature on rebound deflection measurements.

Part III deals with the results of some special studies and the application of certain findings. A method of design for strengthening existing flexible pavements is described. Data are presented to show the influence of shoulders on flexible pavement strength. Data from the evaluation of the load-carrying capacity of special base course materials are presented. Information is discussed showing the use of the Benkelman Beam in evaluating pavement failures for maintenance purposes. The influence of frost action on pavement strength and performance is discussed. Data on the seasonal variation in the load-carrying capacity of flexible pavements are presented.

Part IV of the paper deals with the analysis of pavement performance data collected during 1959 and 1960. Pavement performance, design, age, loading, strength and environmental data from several thousand sections of flexible pavement in all parts of Canada have been collected under stage 1 of the investigation program, The information for each of these sections has been recorded on punch cards and stored at a central data processing center. Multiple regression analyses have been performed on a” electronic computer to relate flexible pavement performance with the principal variables affecting it. Results are presented and discussed which illustrate some of the Committee’s findings to date, while plans for further work are indicated. The volume of pavement inventory data collected during 1961 is far greater than the combined data for 1959 and 1960. The completed 1961 data was received by the central computation center in February 1962 and will be included in the analyses to be carried out during the spring of 1962. It is hoped that further results will be available at the time of the International Conference in August 1962.

1) The data may be used to isolate and evaluate the relative effects of the principal variables on the performance of pavements. This may be accomplished through multiple regression analysis with performance as the dependent variable and such factors as rebound deflection, age, traffic and climate as the independent variables.

2) Blocks of data may be selected from the pavement inventory for special evaluations. As the pedological soil classification of the basement material defines the subgrade strength quite accurately, sections constructed on the same pedological soil type may be selected to compare rebound deflection and thickness or to determine the effects of shoulder type on the performance of pavements of equivalent design.

3) The pavement inventory data provide a permanent, concise record of the design of all pavement sections in the highway system as well as of the performance and rebound deflections values at a given time. The inventory can thus be repeated periodically to trace the actual performance and rebound deflection histories of sections.

4) The inventory data may be used to estimate pavement life.

et cetera. . .

The objective of the study has been to develop quantitative measures of pavement adequacy for the primary purpose of correlating design and performance and to provide a basis for gauging serviceability. Pavements are built for the purpose of providing durable riding quality for the safety, convenience and comfort of the highway user. Major factors which affect the life and serviceability of a highway pavement are its structural characteristics as related to the vehicle loads to which it is subjected and the uncontrolled variables of environment. It was felt that the integrated result of all these factors, both controlled and uncontrolled, could be measured by changes in the pavement profile (roughness) and structural continuity (cracking pattern). These two factors were selected as the quantitative counterparts of riding quality and durability, the two most important attributes of a highway pavement.

A truck-mounted profilometer was built to accurately record pavement profiles; provision was made to record pavement cracking. In the past four years that the equipment has been in operation, some 9500 miles of pavement profiles have been recorded and analyzed. As a result of this investigation, a number of significant relationships have been developed.

Test results indicate that occasional application of overload traffic will shorten the useful life of flexible pavements in proportion to the magnitude of the overload traffic; occasional drastic overloads will not necessarily result in immediate pavement failure.

Some information is also included relative to the relationship between the Bureau of Public Roads roughometer measurements and slope variance measurements of the AASHO Road Test-type profilometer.

Alternate present serviceability equations have been developed for both rigid and asphalt pavements. The measurements required are identical to those for the Carey and Irick equations except spalling has been added to the rigid equation.

Note on a Method of Analysis for Pavements
J. Bachelez, Ingenieur en Chef de L’Aeroport de Paris
G. Jeuffroy, Societe Anonyme Pour la Construction et 1’Entretien des Routes, Paris, France

Fatigue Characteristics of Bitumen and Bituminous Mixes
P.S. Pell, University of Nottingham, Nottingham, England

Model Study of Stresses in Asphalt Pavements
Dr. Bh. Subbaraju, Central Road Research Institute, New Delhi, India

Continuation of Study on the Theoretical Design of Flexible Pavements Based on Shear Strength
William L. Hewitt, Cornell University, Ithaca, New York

Behavior of Asphaltic Concrete Diaphragms to Repetitive Loadings
R.A. Jiminez and B.M. Gallaway, Texas Transportation Institute, College Station, Texas

The Bearing Capacity of Asphaltic Concrete Carpets Surfacing
E. Shklarsky and M. Livneh, Israel Institute of Technology, Technion City, Israel

Shear Loads on Pavements
E.S. Barber, U.S. Bureau of Public Roads, Washington 25, D.C.

Analyses of Road Test Data Using Procedures Developed in U.S. Army Corps of Engineers Accelerated Traffic Tests
C.R. Foster, National Bituminous Concrete Association, College Station, Texas

General Analysis of Stresses and Displacements in Layered Elastic Systems
R.L. Schiffman, Rensselaer Polytechnic Institute, Troy, New York

Fatigue tests on sandsheet specimens carried out under constant amplitude bending stress at various temperatures between -13.5 deg C and +25 deg C show that the material exhibits fatigue properties over wide ranges of stress and that for a particular temperature and speed of loading the relationship between the logarithm of the stress and the logarithm of the number of cycles of loading to cause failure is linear between 10^4 and 10^8 cycles. The life under constant stress amplitude tests is highly dependent on the temperature, a low temperature giving a longer life at a particular stress; it is also dependent to some extent on the speed of loading. However, taking into account the stiffness of the material which depends on temperature, speed of loading, rheological characteristics, and composition of the mix, it has been found that when the logarithm of the strain, calculated as the stress amplitude divided by the stiffness, is plotted against the logarithm of the number of cycles to failure, all experimental results at different speeds and temperatures for one mix lie with a certain amount of scatter about one straight line. It appears, therefore, that the fatigue life is primarily controlled by the magnitude of the applied strain and not by the stress, and that the effects of temperature and speed of loading can be accounted for by their effect on the stiffness of the specimen.

The results of fatigue tests on sandsheet specimens under constant amplitude torsional strain at different temperatures between -20 deg C and +40 deg C confirmed the bending results, but at the higher temperatures under this type of loading the fatigue life includes a considerable crack propagation time, the rate of propagation depending on the stress at the tip of the crack. Examination of the fatigue cracks and failure surfaces showed that in nearly all cases failure originated on the principal tensile plane.

Similar results have been obtained for mixes containing different amounts of aggregate but as the quantity of aggregate in the mix is reduced so the life for a given strain increases, suggesting that the criterion of fatigue crack initiation in bituminous mixes may be one of tensile strain in the bitumen present in the mix.

Tests on bitumen alone at various temperatures both in bending and shear also gave comparable results on the basis of tensile strain, but under certain conditions, particularly at low stresses, the measured fatigue life Includes a considerable length of time necessary to propagate the crack or cracks sufficiently to terminate the test. Unlike sandsheet specimens, bitumen alone showed beneficial effects of rest periods particularly at higher temperatures.

The load was applied to the model pavement slab either through a 3-3/4 inches diameter by 1/2 inch thick circular steel bearing plate or through a hard rubber wheel of 4 inches diameter attached to the head of a Universal testing machine.

Strains were measured at different parts of the slab using strain gauges of the equiangular rosette type. The strain data obtained from about 700 strain measurements were reduced to principal stresses and shear stresses.

The results obtained in this study indicate the presence of rather large tensile stresses in the slab and must therefore, be given careful consideration in the design of asphalt pavements. Under the conditions of the experiments, the maximum stress in the asphaltic concrete pavement could be computed very closely by an equation given in this paper.

In comparing the theoretical design procedure developed by the writer with that used by the Texas Highway Department, It is found that the basic equation gives thickness of pavement suitable for only moderate roadway traffic and that a traffic factor or safety factor must be applied to adjust for traffic volume. The range of traffic factors is from 1 to 2, corresponding to traffic volumes of moderate to heavy.

An indirect comparison with the Hveem Stabilometer method of thickness design indicates that the theoretical procedure gives reasonable values for moderate traffic, indicating the need for a traffic factor for heavy traffic. The indirect comparison is made possible through the development of a relationship between angle of internal friction and resistance value, R.

Design curves are developed for each of four wheel load categories with a traffic factor of 1.0. For the 10,000 pound dual wheel load design curves are given also for a traffic factor of 2.0. It is hoped that the convenience of the design curves will stimulate interest in the design of flexible pavements on the basis of the shear strength properties of the pavement components.

Several loading and mixture variables were investigated to determine their effects on the flexibility and endurance to repeated loads of coarse sheet-asphalt mixtures. The results of the investigation show expected findings that (a) resistance to repetitive loads is a function of specimen thickness, (b) different types of asphaltic concrete have different degrees of endurance to repetitive loads and (c) thicker specimens are not capable of bending as much as the thinner ones.

In this paper, the basis for the bearing capacity calculation is the author’s suggested modification of Coulomb’s rupture theory in which the internal friction angle is isotropic and the cohesion anisotropic. In the light of this modified theory, and with the aid of the appropriate laws of the theory of plasticity, bearing capacity formulae can be derived for the following three cases: (a) thick carpets; (b) thin carpets over a non-rigid base; (c) thin carpets over a rigid base. These formulae take into account the friction between the contact surfaces of wheel and asphaltic concrete and between the asphaltic concrete underside and the upper surface of the base, and utilize the values of the anisotropic cohesion and isotropic angle of internal friction of both carpet and base material.

The main conclusions arrived at from the theoretical considerations in this paper, show the following factors affecting the bearing capacity of an asphalt carpet (a) Anisotropy factor. In all cases where it is greater than unity the bearing capacity is reduced. It is clear that disregarding this factor leads to overestimation of the bearing capacity. (b) Braking stresses. Although the cohesion of the material is higher in braking than at rest, the former state remains the most critical, and practical experience supports this. (c) Strength of the base in carpets over a flexible base. A small increase in the cohesion of the base material greatly increases the bearing capacity, as does increasing the thickness of the carpet. (d) Coefficient of friction between the carpet and a rigid base. When the coefficient of friction is small, this case may be more critical, than that of a flexible base. Here, too, increasing the thickness of the carpet increases the bearing capacity for inclined loading but reduces it for vertical loading.

Naturally, only practical tests in the field can provide final confirmation of the conclusions; pending such confirmation they can, it is hoped, serve as a satisfactory means of estimation.

When a vertical load is applied to a pneumatic tire on a horizontal surface, shear stresses acting toward the center of the loaded area are produced on the surface of contact. By numerical combination of concentric rings of shear stress, the vertical normal stresses produced by a uniform distribution of inward acting shear stresses over a circular area are calculated and graphed.

The vertical normal stress under the center of a circular area to which various distributions of inward acting shear stresses are applied is also shown. For vertically loaded pneumatic tires, measured stresses are generally somewhat greater than those calculated from the vertical load. The effect is most important at shallow depths and should be considered in studies of effect of pavement thickness on transmitted stresses.

Shear stresses in one direction between tire and pavement surface may be produced by various means, such as longitudinal grades, super-elevation, starting, stopping, and horizontal curves. Interior vertical normal and horizontal shear stresses caused by a uniform shear stress in one direction applied to a circular area are calculated and related to similar stresses from applied normal loads. The horizontal shear stress is especially important in the stability of layered systems, where the bond between layers may be critical.

The effect of horizontal load components in reducing bearing capacity is evaluated. When horizontal and vertical stresses are applied together, their combined effect must he considered in any analysis of bearing capacity.

Under static loads, which are often critical for design because of creep, an exact analysis of stresses must include the inward acting shear stresses peculiar to pneumatic tires. When acceleration or deceleration produce strong horizontal loading components, critical stresses are greatly increased and bearing capacity is greatly decreased when compared to vertical loading over the same time. Either of these conditions is more critical than a vehicle moving at constant velocity along a horizontal tangent.

In the analysis of the several accelerated traffic tests conducted by the U.S. Army Corps of Engineers on flexible pavements it was necessary to separate the effects due to shear deformation and compaction and to make separate analyses. Also, separate procedures were developed to provide designs against shear deformation and compaction.

Shear deformation occurs when a layer is overstressed and for any given loading condition the primary variables are the strength of the layer being considered and the thickness above it, All existing flexible pavement designs are concerned primarily with thickness. In the analysis of thickness requirements as indicated by the Corps of Engineers accelerated traffic tests and by actual airfield performance it was found necessary to consider the in-place strength of the layer being studied. Also, since the available data for any one condition of loading were limited it was found desirable to develop methods of expressing the loading conditions in dimensionless values so that all the available data on thickness versus strength could be compared simultaneously.

This paper reviews the dimensionless methods developed in the analysis of the Corps of Engineer accelerated traffic tests for thickness and illustrates the use of these methods with the WASH0 test results, The compaction that develops in a given layer in a flexible pavement structure under load is primarily a function of: (1) the soil type, (2) the loading, and (3) the depth from the point of load application to the layer. The variability produced by differences in soil types has traditionally been treated in a semi-dimensional manner by expressing the degree of compaction in terms of the maximum unit weight obtained in a standard laboratory compaction test. In the analysis of the compaction requirements as indicated by Corps of Engineer accelerated traffic tests, methods of expressing loading conditions in dimensionless numbers were developed so that simultaneous comparisons could be made of the compaction-thickness relationships for all available data. These are reviewed in this paper.

The paper points out that although the developments were based on airfield pavements they are applicable to road test data.

This paper presents a general theory for the analysis of elastic layered systems. The current state of knowledge is reviewed. General formulations of the effect of layering are developed. Analytical procedures are developed for the consideration of general distributions of normal and tangential surface loads, and axi-symmetric and slightly inclined rigid plates.

The Explicit Solution of the Equations of the Elastic Deformations for a Stratified Road Under Given Stresses in the Dynamic Case
Andree Bastiani, Dr. in Mathematics, Paris, France

A Fundamental Approach to the Design of Flexible Pavements
K. R. Peattie, Shell Research Limited, Thornton Research Center, Chester, England

Theoretical Concepts Applied to Asphalt Concrete Pavement Design
F. N. Finn and E. L. Skok, Jr., The Asphalt Institute, College Park, Maryland

Applications of Layered System Concepts and Principles to Interpretations and Evaluations of Asphalt Pavement Performances and to Design and Construction
Donald M. Burmister, Columbia University, New York 27, New York

The Response of Linear Viscoelastic Materials in the Frequency Domain with Emphasis on Asphaltic Concrete
Hratch S. Papazian, Ohio State University, Columbus 10, Ohio

A Structural Design Procedure for Pavements
R. F. Baker, Ohio State University, Columbus 10, Ohio

Viscoelastic Behavior of Asphalt Concrete Pavements
Carl L. Monismith and K. E. Secor, University of California, Richmond, California

The Application of Elastic Theory to Flexible Pavements
A. C. Whiffin And N. W. Lister, Road Research Laboratory, Harmondsworth, Middlesex, England

Analysis of Viscoelastic Pavements Subjected to Moving Loads
K. S. Pister and R. A. Westmann, University of California, Berkeley 4, California

Use of Galerkin’s Method for the Study of Static and Dynamic Behavior of Road Structures
R. Lattes, J. L. Lions, J. Bonitzer, Laboratoire Central des Ponts et Chaussees, Paris (XVe), France

Designing Flexible Road Pavements
G. Schnitter and R. Jenatsch, Swiss Federal Institute of Technology, Zurich, Switzerland

First, we solve the problem of finding the solution of the system in the i-th layer, knowing displacement or stress components at the upper or lower surfaces. For this, the notion of function is generalized using the Ehresmann local structures and the differential system is considered as a system of equations for the new objects; by an integral transformation the system is reduced to a system of linear equations. By solving this system, we obtain relations between the integral transforms of the displacement and stress components at the upper and lower intersurfaces of the i-th layer as well as at the upper surface of the last layer.

Assuming the displacement components known on the free surface, the integral transforms of the stress and displacement components are computed on the second intersurface as functions of these parameters. These quantities are also computed as functions of the displacement components at the last intersurface. By writing that the solution is the same in both cases, we are led to a system of linear equations, the solutions of which are the integral transforms of the displacement components at the free surface and at the last intersurface. From these, we deduce the solutions of the initial equations.

The solutions are obtained in the form of ordinary integrals containing the given stress components. These integrals can be computed with a computing machine or approximated by elementary functions (the approximation depends on respective sizes of the parameters).

The regularity conditions imposed for the given stresses are practically not restrictive; in particular, strain and stress components are not supposed to be harmonic.

Examples: stresses produced by a vibrating machine or by the movement of a vehicle.

The increasing use of unconventional road structures, such as those incorporating greater thicknesses of bitumen-bound layers, emphasizes the need for a better understanding of the contribution which each layer makes to the strength of the whole structure and for the development of thickness design procedures based on fundamental considerations. Most existing methods for determining the thicknesses of roads are empirical and do not take into account the different load-spreading abilities of base materials and cannot be extended to cover new conditions of loading without extensive field trials.

The basis of the theoretical treatment is the assumption that a real road structure may be represented by a series of elastic layers lying on a semi-infinite elastic mass. Traffic loads are assumed to be applied to the layered system as stresses uniformly distributed over a circular contact area.

The stresses and strains in such a structure may be obtained by solving the general elastic equations applying to the behaviour of elastic layered systems. Rigorous solutions of the stress and deflection equations have now been published for a wide range of the parameters involved in three-layer systems.

When calculating stresses, strains and deflections it is necessary to know the elastic properties of the materials in each layer. The moduli of soils and granular base materials may be determined in the field by vibrational techniques. An approximate relationship is available connecting the values of Young’s modulus of such materials with their California Bearing Ratio values. The moduli or stiffnesses of bitumen-bound road materials may be measured in the laboratory, and are dependent on temperature and traffic speed.

The critical quantities in a flexible road structure are considered to be the vertical compressive stress at the surface of the subgrade and the horizontal tensile strain at the bottom of the bituminous layer.

A relationship between the permissible value of the vertical stress in the subgrade and the CBR-value of the soil has been developed from an analysis of road structures which are known to be satisfactory in practice. Because of the predominantly repetitive nature of the loading applied to roads considerable attention has been given to the fatigue behaviour of bituminous road mixes. It has been suggested that the principal tensile strain is critical in this respect and a relationship connecting the value of this strain with the number of load applications causing failure has been obtained.

A suitable design for a flexible pavement may be obtained by assuming the layers to be of certain thickness and calculating the values of the critical stresses and strains developed in this structure by the design load. These thicknesses are then adjusted, to bring the values of the critical stresses and strains within the permissible limits. The construction of design charts which can be used more directly and rapidly is described.

It was decided to use the equations presented by Burmister and developed by others for the two-layer and three-layer system because they represented solutions for layers which are “fully elastic.” In order to use the elastic layered system theory for analysis, the components of a pavement were assumed to be elastic (have a constant modulus of elasticity). Consideration of the magnitude and duration of the stresses imposed on a pavement system by normal highway loads leads us to believe that stresses calculated from the elastic theory are proportional to the actual level of stress in a pavement system, With one application of a load, an adequate pavement will rebound almost completely, although in many cases permanent deformations occur after a number of applications. We feel, therefore, that a section can be treated elastically with each load application. When the visco-elastic theory is made more workable, it may provide a better approximation to the actual stresses and permanent deformations occurring in a pavement system.

The paper reviews the use of present methods for determining strength coefficients of the various components in a pavement section. These methods are used to establish proper “working” moduli of the materials. If the materials are tested at a stress-strain level close to that imposed on a pavement structure (which would be near the origin of stress-strain curve), the stress-strain ratio can be used as a modulus of elasticity, and the determination of theoretical stresses and deflections of various load and design conditions can be based on that value. In most cases for an adequate design, the strain level was low; therefore, a close approximation to an elastic condition can be assumed. “Working” moduli are determined for the pavement components at the WASH0 and AASHO Road Tests using plate load tests, deflection tests and some vibratory triaxial data on the pavement materials. These “working” moduli were then used to compute stresses in the various sections of the WASH0 and AASHO Road Test.

For correlation of stresses with field performance, it was assumed that accumulated transverse permanent deformation would be correlatable with the vertical stresses imparted to the subgrade (ZZ2) and shear stresses in the layers. Also, because cracking occurs primarily in the surface layer, the tensile radial stresses in that layer (RR1) were considered to correlate with cracking or disintegration in the surface layer. The serviceability concept as used at the AASHO Road Test because of its definition incorporates both cracking and accumulated permanent deformation: therefore, it was correlated with the vertical stress on the subgrade which was considered to be a measure of the general stress level in the pavement system. Mathematical formulae were developed to show these various relationships.

Although the relationships in the paper cannot as yet be used to design an asphalt concrete pavement section, we have attempted to show that the elastic theory can be used as a possible performance model. To set up the proper relationships, it will be necessary to work further to establish proper “working” moduli for the pavement components. We do not propose that the use of the elastic theory will result in a completely rational design, but by using the elastic theory in this manner, it is possible to relate the load on the pavement to the strength of the components in a more rational way than is now possible by most existing design procedures.

It is hoped that the advantages of using a theoretical approach to the design of an asphalt pavement have been shown and that further planned studies will be directed toward verifying these relationships.

The distribution of vertical stresses in a layered system discloses the increased load-spreading capacity and reduction in stressed imposed in the subgrade layer with increase in the ratio of the effective moduli of the component layers. But increased vertical stress gradients through the reinforcing layers bring into action a shear stress build-up through the reinforcing layers, which may become critical to the point of initiating a breakdown of a pavement structure. The critical nature of shear stress conditions is due to the fact that they are essentially deflection-dependent and that they increase with stiffness of the pavement system. These shear stress conditions can be alleviated by multi-layer construction with smaller jumps in moduli ratios and by somewhat increased thickness of reinforcing layers, so as to maintain the desired deflection performances of the layered pavement system.

The only practical, effective, and satisfactory method of evaluating the “in-place” layer moduli of the component layers, and the deflection performances of a layered pavement system is by means of systematic comparable series of prototype load bearing tests. The principles and methods of evaluation of layered pavement systems are presented, making use of deflection and shear stress influence curves. Such evaluations provide the basic information for design of multi-layer pavement systems and over-all thickness to limit deflections and shear stresses sufficiently below critical values to insure the integrity and long service life of layered pavement systems under repeated load conditions.

Two cases have been analyzed – the Hybla Valley Test Track Data, and the WASH0 Road Test Data. The results are presented in two graphs showing the “in-place” layer moduli, the deflection performances, and even more important the degree of constructional excellence and actual uniformity achieved. These findings are discussed and they have important implications with regard not only to necessary modifications of design criteria, but also to essential modifications in construction methods and sequences, in order to attain and to insure the necessary high standards of excellence in construction demanded for modern expressway systems. The spread in the quality of construction disclosed by these evaluations is too large to be considered acceptable, and the average of construction is too low compared with the potential excellence evident from these analyses, which could be consistently achieved in construction. The evaluations have shown that inferior deflection performances were due principally to local constructional conditions of inferior shear deformation characteristics in granular subbase and base course construction. Methods for “proof testing” each pavement component before acceptance should be developed and used as a planned supervision part of construction in order to insure a uniformly high standard of constructional excellence.

The general stress-strain equations of linear, viscoelastic materials are defined in the frequency domain in terms of algebraic coefficients which are functions of frequency. These coefficients are complex numbers whose magnitude and phase, at any given frequency, depend on the properties of the material. They are named the complex moduli of the material, and are shown to be fundamental material constants which are independent of testing procedures or boundary conditions. Analytical and graphical procedures are presented for obtaining the complex moduli of a material experimentally. These are based on a series of dynamic tests covering a wide range of frequencies, or a single static test covering a wide range of time. The first gives magnitude and phase of the complex modulus at each frequency used, while the second method yields an analytical expression for the modulus as a continuous function of frequency.

These methods are applied to determine the complex elastic modulus E*, and the complex transverse modulus T*, of asphaltic concrete, which, at levels of stress sufficiently low compared with its ultimate strength, and for small strains, is shown to act as a linear, viscoelastic material. Both sinusoidal-stress dynamic tests, and constant-stress static tests are performed on several mixes at various temperatures, and a good correlation is obtained between the results of the two types of tests.

In the case of an isotropic, linear viscoelastic material, it is shown that two independent complex moduli, such as E* and T*, are sufficient to describe its response in full. However, to relate the mean normal stress to the mean normal strain, and the deviatoric stress tensor to the deviatoric strain tensor, two other complex moduli are defined, namely, the complex bulk modulus K*, and the complex shear modulus G*; and formulas are presented for obtaining them in terms of the other two independent complex moduli. By means of a mechanical analogy, these moduli are interpreted as the impedances of mechanical models, made up of elastic and viscous elements, whose constants are easily found.

The general stress-strain equations of the linear, viscoelastic body in the frequency domain are shown to be identical in form with the classical stress-strain equations of the elastic body in the time domain and to include them as a special case. Methods are presented for transforming the relationships from the frequency domain back to the time domain.

Variations in the components of the elastic moduli are studied in terms of response of the material under stress, and as a means for the evaluation and comparison of pavement materials. Finally, applications to pavement design and performance studies are outlined.

The behavior of the components under stress is outside the elastic theory insofar as measurable permanent deformations or cracking occurs. The term “failure” of a material is ambiguous and the term “response” is preferred because the “failure” of a pavement structure is a serviceability-economic decision not based upon any specific level of material behavior.

The importance of the ratio of elastic moduli in defining stress levels is emphasized according to the theory, moving vehicles tend to develop higher flexural stresses and lower subgrade pressures than do static loads. Thus, moving loads are a more severe test of the pavement layers, while static loads produce a more rigorous test of the subgrade.

The load distribution effects of a pavement component will not be greater than indicated by the elastic modulus. Thus, the quality of a surface or base course from a load distribution perspective is defined by its modulus of elasticity (assuming Poisson’s Ratio is constant). However, since stress is defined by the ratio of moduli, the “quality” of a base course in terms of load distribution varies with the stiffness of the overlying and underlying materials.

Continued work in evaluating stress and performance, and additional solutions to the elastic theory layered system will improve the suggested approach. The method is considered a higher level approximation than is currently in use and an evolutionary step in the ultimate use of the viscoelastic approach.

Based on analysis of data of the various investigations, two simple linear viscoelastic models were selected and their suitability evaluated by four types of triaxial compression tests on one asphalt concrete mixture, namely: (1) creep, (2) stress relaxation, (3) constant rate-of-strain, and (4) repeated axial load. Initially it was hoped to eliminate the effects of lateral pressure in these tests. However, data for deviatoric stress and strain are presented which preclude the use of such an approach.

From the triaxial compression tests, it was ascertained that the simplest model for asphalt concrete required at least four elements in order for it to demonstrate instantaneous elastic deformation, retarded elastic deformation and viscous flow. Data are presented illustrating that this model will, in the majority of cases, predict with a reasonable degree of accuracy the time-dependent behavior of asphalt concrete in the four types of tests over a range in load conditions, lateral pressures (0 – 250 psi) and temperatures (40 deg F – 140 deg F). Variations of the viscoelastic properties for these conditions are also presented together with a discussion of their significance. Thus this phase of the research produced a basic system of data relating the stress-strain-time characteristics of a particular asphalt concrete.

Using the data developed from the triaxial compression tests, solutions were developed for the time-dependent deflections of a viscoelastic plate on an elastic foundation for static loading. Static plate loading tests were performed on 4 ft by 4 ft slabs of the same asphalt concrete mixture resting on an elastic foundation composed of 1600 springs over a range in temperatures (40 deg F – 140 deg F) and at different levels of stress. Comparisons of the predicted and measured time-dependent deflections are presented. In general, the measured and computed deflection profiles had the same general shape and time-dependence, but the measured values had magnitudes considerably greater than those given by theory with the deviations between the two sets of values increasing with increased temperature. A discussion of the probable sources of difference between the slab test data and the theoretical predictions is included.

One of the major discrepancies between theory and actual behavior appeared to be the assumption of equal properties in tension and compression of asphalt concrete for slow rates of load application. Thus data are presented for tension and compression creep tests at 77 deg F which illustrate differences in the time-dependent properties. Data are also presented for bending creep tests on the same specimens which indicate that the divergence in tensile and compressive properties is time dependent.

Using the results of the bending creep tests, a modified theory is presented and the slab test results at 77 deg F are reanalyzed using this theory accounting for the differences in tensile and compression properties. Considerable improvement between the theoretical predictions and actual slab test data is obtained. Thus it would appear that for static loading, at least, viscoelastic theory may be required to predict the behavior of asphalt concrete pavements.

The present methods of pavement design are ad hoc in character and based upon experience of the behaviour of different types of road over a wide range of traffic and soil conditions. A more reliable method of design might be developed on the basis of the dynamic stresses produced in the road by moving vehicles, the stress/deformation characteristics of the layers in the road under repetitive loading, and the variation of these properties with time. The present paper reviews the position, but it is not yet possible to present a technique for designing roads from this information.

Several theories for computing stresses and deflections in multi-layered elastic systems have been developed and are reviewed. That devised by Burmister appears to fit the conditions applying in a flexible road. Diagrams are given showing how the various stresses in a multi-layer road vary with the dynamic elastic moduli of the layers and their thicknesses, the curves being based on an analysis of the computations performed by Acum and Fox. The vertical stress at the soil/base interface under a moving wheel rises with increase of the elastic modulus of the soil. This stress falls with increase of the elastic modulus or thickness of the road base or surfacing. A road base of high elastic modulus has good load-spreading properties, but this is accompanied by high horizontal tensile stresses in the base near to its interface with the soil. Some road bases, although having the high elastic modulus necessary for good load-spreading characteristics, do not have sufficient tensile strength to withstand the stresses generated within them and failure occurs.

A review is given of the information available concerning the dynamic elastic moduli of road-making materials and shows that much more information is needed, It has so far proved difficult to devise in situ tests for determining the moduli of bitumen or tar-bound materials at the rates of loading applying to traffic conditions.

Values of the dynamic elastic moduli have been obtained for some of the materials in roads where measurements have been made on the dynamic stresses generated in the subgrade. Reasonable agreement then occurred between the measured and computed values of the dynamic stress. The measurements of stress showed that, of the base materials tested so far, rolled asphalt appeared to have the best long term load-spreading properties. When any of the layers of the road contained tar or bitumen, the dynamic stress applied to the soil rose with increase of the temperature of the road. The stresses were found to be proportional to wheel load.

Before data on stresses and deflections can be used in pavement design, detailed information is required on the behaviour of the layers under the conditions of repeated stress to which they are subjected in a road. Asphalts have been studied, but little work has been done on mixes of high void content such as bitumen macadam or tarmacadam, while no information is available on the fatigue properties of road bases. Measurements of the deflections of roads under moving vehicles will eventually lead to information concerning the elastic moduli of the layers. The Benkelman Deflection Beam is being used in several countries to assess the quality of a road, the measurements being made when subgrades are normally in their weakest condition. Results obtained so far indicate that the magnitude of deflection criteria depend on the intensity of the traffic using the road, the type of subgrade and the type and thickness of base and surfacing. The most important factors seem to be the type of the base material and the traffic intensity. The elastic approach outlined in the paper may never give a complete design method, but it is already being used to analyze road failures and a typical example is outlined in detail.

Stationary loads (for example, parked vehicles) produce time-dependent deformation in a vlscoelastic pavement. A single load or a series of load applications causes an accumulated displacement or strain condition that may produce permanent deformation or failure by cracking. The analysis for such problems can be carried out in a manner parallel to the elastic analysis by uslng the elastic viscoelastic correspondence principle. Typical examples of such analysis for stationary loads applied to viscoelastic plates on elastic as well as viscoelastic foundations may be found in (3) and (4). The deformation produced by a number of loads distributed over an area can be obtained by superposition as in the corresponding elastic problem. It should be noted that the relationship between temperature and viscoelastic mechanical properties, as described in (1) and (2) is of fundamental importance in this type of problem.

In the sequel two factors which have not received much attention in the application of elastic or viscoelastic theory to pavement analysis will be discussed. The first concerns the effect of moving loads on a vlscoelastic pavement, while the second Involves improving the analysis by including the possibility of different mechanical properties in tension and compression.

Galerkin’s method presents many advantages:

1. It is well fitted to use of electronic computers.

2. It is applicable to dynamic loading cases as well as to static ones.

3. It allows stresses and deformations to be computed in a great many points (one hundred for instance) of the road structures without excessive increase of the cost.

4. Cost of computations does not increase very much as the number of layers of the structure increases.

5. Calculations are made in a Cartesian system of coordinates which allows simple calculation of stresses and deformations even when loaded areas are not of simple forms. Also it is expected the method could be extended to cases of horizontally limited structures.

6. Viscoelastic properties of certain layers can be taken into account. Presented model uses simple Voigt models, but more complex models, giving better approximation of viscoelastic properties of materials, could be as well used.

Finally, it seemed useful to compute not only stresses and deformations, but also their derivatives relating to elastic and viscoelastic parameters characterizing road layers.

The design of the rigid pavement can today rely on uniform methods. The basic factors, such as the behaviour of the building material and the mechanics of the pavement acting as a slab, are governed by relatively simple rules and can be regarded as fairly settled.

As regards the design of the flexible pavement, however, so far no uniform method has become established. The fact that the properties of the bituminous binders vary with temperature and loading period, and also the wide variety of the products available have so far defeated a clear determination of the key factors essential for the design of flexible structures. Though there are various designing methods, these are largely based on the personal experience and preference of their inventors.

Accordingly, the present report chiefly gives a survey of the terms and key factors used. Moreover, it gives a description and a critical comparison of the principal designing methods at present applied to flexible pavements.

Reference is made to the effect of dynamic factors on design. Finally, the method of design used by VAWE is described.

Time-Dependent Deformation of Clay Soils Under Shear Stress
Manrique Lara-Thomas, Ohio State University, Columbus 10, Ohio

Basic Material Properties for the Designof Bituminous Concrete Surfaces
Emil R. Hargett, South Dakota State College, Brookings, South Dakota

Resilience Characteristics of Subgrade Soils and their Relation to Fatigue Failures in Asphalt Pavements
H. B. Seed, C. K. Chan, and C. E. Lee, University of California, Berkeley, California

A Study of the Repeated Load Strength Moduli of Soils
H. G. Larew And S. B. Ahmed, University of Virginia, Charlottesville, Virginia

The Effect of Resilience-Deflection Relationship on the Structural Design of Asphaltic Pavements
F. N. Hveem, E. Zube, R. Bridges, and R. Forsythe, Division of Highways, Sacramento 19, California

Dynamic Testing as a Means of Controlling Pavements During and After Construction
W. Heukelom and A. J. G. Klomp, Koninklijkc/Shell-Laboratorium, Amsterdam, Netherlands

Analysis of Flexible Airfield Pavements by Surface Plate Loading
Philip P. Brown, Bureau of Yards and Docks, U.S. Navy, Washington 25, D. C.

The Structural Behavior of Flexible Pavement – An Analysis of Rigid-Plate Bearing Tests on Full-Size Test Sections
A. C. Benkelman and S. Williams, Phvsical Research Division, U. S. Bureau of Public Roads, Washington 25, D. C.

Dynamic Testing at the AASHO Road Test
L. W. Nijboer, Koninklijke/Shell Laboratorium, Amsterdam, Netherlands
C. T. Metcalf, Shell Oil Company, Martinez, California

Resiliency of Base Rock
O. A. White, Oregon State Highway Department, Salem, Oregon

Finnish Road Structures and the Use of Wedges Against Frost Heaving
0. A. Taivainen, Faculty of Technology, Oulu University, Oulu Finland

The two most noticeable factors associated with the weakness of the pavements observed have been (1) inadequate compaction within the structure and (2) inadequate subgrade support. From the first experimental project it appears that high deflections can be minimized more effectively by improving compaction of the structural components than by increasing the thickness of asphaltic concrete in the same total structural thickness. But on more recently constructed projects an even more striking reduction in deflection has resulted from substituting subgrade stabilization for select borrow as the lowest course in the pavement structure.

Some of the deflection data on which the above statements were based are summarized graphically in this paper. Advantages of subgrade stabilization are listed along with certain practical considerations to be made before choosing this method of design. Also, recommendations for the application of the research findings to the design of future pavements are shown in a table, These recommendations, if adopted, should result in major savings in initial construction costs in comparison with rigid slab designs and with some recent flexible designs involving greater thicknesses of black base than those tabulated.

The viscoelastic approach is based on the general concepts of the science of rheology, or the science of flow. Briefly, the theory assumes that the mechanical behavior of real materials can be approximated by a certain combination of the behavior of two ideal bodies: the ideal elastic body and the ideal viscous body. Thus, in a very general sense, the mechanical behavior of any real material can be determined by a particular equation or set of equations giving a relation between the components of the stress and strain tensors, and also between the components of the rate of stress and strain tensors.

The purpose of the study was:
(1) To investigate the deviatoric rheological properties of certain cohesive soils under creep tests.
(2) To apply concepts of the viscoelastic theory in the analysis of the results.
(3) To develop procedures to fit rheological models to experimental data.
(4) To investigate whether the materials tested exhibited linear viscoelastic behavior.
(5) To study the effect of load repetition on the properties described above.

A new method of tension and compression testing is described and the test results explained in view of our needs for basic design data. This method of tension and compression testing enables the research and design engineers to evaluate the stability of bituminous concrete in terms of basic strength components. Shearing resistance and angle of internal friction may be obtained graphically from a Mohr diagram plotted from tension and compression test data. Averages of the test results obtained from a limited amount of testing are included. The application of these test data to a rational design approach is discussed briefly. This paper advocates the collection and evaluation of data regarding basic material properties as a means of upgrading the design methods now in popular use.

The marked increase in resilient deformations which may result from compaction in the field to a degree of saturation exceeding about 85 per cent and, for relatively inexpansive soils, the desirability of compacting to degrees of saturation not exceeding about 80 per cent, in order to minimize resilient deformations, are indicated. The marked reduction in resilient deformations resulting from slight increases in compaction density are also shown.

Evidence is presented to show that the resilience characteristics of a soil compacted with pneumatic rollers in the field are very similar to those of laboratory samples prepared by kneading compaction; however, the properties of field samples which attain a high degree of saturation by moisture absorption after compaction to a low degree of saturation are best simulated either by duplication of these conditions in the laboratory, or by compacting the soil directly to the final condition by static compaction.

The resilience characteristics of the subgrade soils in the WASH0 and AASHO Test Roads are evaluated. It appears that the AASHO Test Road subgrade is perhaps a particularly resilient soil at high degrees of saturation and that the conditions and method of construction may have led to somewhat higher resilience characteristics throughout the project then would result from more conventional construction procedures.

Finally, considerations in selecting a resilient modulus for incorporation in elastic theories for layered systems are presented.

The method used to obtain the repeated load, triaxial test, stress-strain curves is believed to be original. Each of these repeated load stress-strain curves was obtained by subjecting a series of identical soil samples to varying levels of repeated deviator stress and then measuring the resulting equilibrium deformation or strain for each level of repeated deviator stress. These values of repeated deviator stress and resulting equilibrium strain were then used to plot the repeated load stress strain curve.

Stress-strain curves were obtained by both conventional (gradually applied load) and repeated load triaxial tests and the resulting deformation moduli were compared at various levels of compaction, molding moisture content and density.

For the three soils studied, it was found that the repeated load deformation modulus was normally less than the modulus obtained in the conventional triaxial test. In the case of the sand-clay material this difference between gradually applied and repeated load moduli was as great as 100 per cent. For a given level of compactive effort both gradually applied and repeated load moduli decreased with increasing molding moisture content for all soils, although the rate of decrease was less for the sand-clay soil.

For a given compactive effort the ratio of the gradually applied load modulus to repeated load modulus was affected to varying degrees by changes in molding moisture content. The effect of molding moisture content upon this ratio was more pronounced in the case of the sand-clay soil where the ratio of Es/Er attained a value of 2 at higher moisture contents at both high and low levels of compactive effort. For the other two soils studied this ratio varied from about 1 to 1.5.

These results seem to indicate that a soil deformation modulus obtained from the conventional or gradually applied load triaxial test on even a nonrepeated load plate bearing test will not properly reflect the action of a soil under a pavement, for example, where the soil will be subjected to repeated loads. Some current methods of pavement design based upon a deformation modulus obtained from conventional triaxial tests or plate bearing tests may need to be restudied and revised to more properly reflect the action of repeated loads.

In 1955, based upon the results of deflection measurements from 400 electronic gauge units on 43 projects, limiting deflection values for the prevention of fatigue cracking in the design life of a bituminous surfacing were presented by the California Division of Highways for a variety of structural sections. Since that time, these criteria have been used as a guide for the determination of the adequacy of existing pavements and the magnitude of necessary reconstruction.

In order to introduce the deflection factor into design criteria, a device known as the resiliometer, has been under development by the Materials and Research Department since 1946. This instrument subjects a 4 inch diameter by 4 inch high soil specimen to cyclic dynamic loads of varying intensities of 0 to 50 pounds per square inch. The resilience test value is the net volumetric compression and rebound resulting from these loads. Considerable time and effort was expended in modifying the resiliometer in order to increase test sensitivity and reproducibility and in making qualitative resilience appraisals of different soil types from throughout the State.

In 1960, a program aimed at the correlation of laboratory resilience measurements with field deflection data was initiated as the first step toward the incorporation of the test soil resilience value into the flexible pavement design procedure. Twenty-five sampling locations from California highways and the AASHO Test Road were included in this study.

At each test area, field deflection measurements were made with the Benkelman Beam and truck (15,000 lb. single axle load). The sampling locations selected (based primarily on the range and consistency of the deflection measurements obtained) were sampled to a depth of 30 inches. Undisturbed samples obtained from each element of the structural section were tested in the resiliometer at the air pressures appropriate to the depths at which the samples were taken based upon a modification of the Boussinesq equation. Where undisturbed samples were not obtainable, the soils were recompacted in the laboratory and tested at field moisture and density.

A plot of the resilience summations for each sampling location vs. field deflection is shown in the report. The result reflects an encouraging trend toward correlation and, with future samplings should provide a tie with pavement performance suitable for the establishment of resilience design criteria. Utilization of this relationship will be based upon adjustment of a proposed structural section to reduce the pavement deflection, predicted from the results of laboratory resilience tests on preliminary samples to a tolerable minimum.

A design example involving resilience test data is shown.

The results of resilience tests on remolded “sensitive” soils after curing periods of varying length are also presented.

Various theories give the possibility to calculate the deflection of layered systems when the E-moduli are known. So the E-moduli derived from wave propagation measurements were found to tally with the deflections measured under the vibrations of the heavy machine.

The measurements can be carried out at any stage of construction. Measurements on soils can be regarded as the first step in road design. Test data give indications about the necessity to improve the subgrade. From the elastic properties an idea of the values of the permissible vertical soil pressure can be deduced.

From measurements on soil improvements and bases it was found that the degree to which layers of unbound materials can be compacted depends to a large extent on the reaction of the subsoil. In general the observed E-moduli are not higher than 1-1/2 to 2-1/2 times the E-values of the underlying material. On poor soils, however, there is a tendency for the modular ratio to be somewhat higher.

To explain these findings, equilibrium stress conditions for granular materials have been considered and compared with computed stresses under the load of traffic.

Bound materials show higher E-values, which are only slightly dependent on the nature of the underlying material. Dynamic testing of materials in the laboratory has yielded data on typical wearing-course materials such as asphaltic concrete.

Dynamic measurements on total road constructions can serve to check whether the finished road meets the requirements specified in the design.

Variations of the properties of the successive layers of the road construction, resulting from compaction, changes in water content, freezing, thawing and cracking could be observed.

Analysis of the data show that for those pavements loaded, which in all cases had been subjected to traffic for several years, the effective subgrade modulus computed from surface loading is considerably greater than that determined by loading directly on the subgrade. Conversely, the effective elastic modulus of pavement material (surfacing, base and subbase) is considerably lower than has previously been indicated.

The data also show clearly that the deflections of the pavement under small radius loadings are governed largely by the properties of the pavement materials, while deflections under large radius loadings are determined primarily by the subgrade properties. The importance of this factor to design for high pressure tires is discussed.

A method of using surface plate load data to determine the required thickness of flexible overlay to support increased wheel loadings is presented. This procedure utilizes the test plot of load versus radii (for a specific deflection) as an influence curve, and applies to instances where the overlay material has the same modulus as the existing pavement material. When superior overlay material is used, the procedure must be supplemented by the layered system analysis.

The test sections consisted of a 3, 6, or 9 inch thickness of bituminous concrete surface course on 6, 12, 16 or 24 inch granular base courses constructed on a uniform clay soil embankment; also, 3, 6, and 12 inch bituminous concrete courses laid directly on the clay soil.

Loads were applied repetitiously through circular steel plates of various diameters to the surface course, the base course, and the subgrade of each of the test sections. Most of the data were obtained by either of two procedures, namely, the Accelerated Test and the Repetitional Test. Briefly, the former test consisted of the application and release of three loads of increasing magnitude followed by the application of a continuously increasing load. The latter test consisted of the application and release of loads of 16, 32, 48 and 64 psi, followed by 75 applications of an 80 psi load.

The effect of the loads was obtained by measuring both the deflection and the recovery of each of the pavement structure components and the subgrade simultaneously.

Flexible pavements in service whose design is structurally adequate to carry the prevailing traffic must act essentially in an elastic manner. It was found in the Repetitional Test series that for certain limiting loads applied on the surface of the pavement, the supporting subgrade acted wholly elastic between about the 10th and the 75th or final application of load. It was considered that these loads were safe loads for the conditions of the test. For the criterion thus established, it was found that the thickness of pavement structure (3-inch bituminous concrete surface plus granular base) required to support a unit load of 80 psi, without producing detrimental or nonelastic movement of the subgrade, varies approximately as the total load to the 0.4 power.

Other findings of the investigation include:
1. The subgrade soil when load-tested directly through rigid plates never developed the same degree of elastic action as when the load was distributed to it through the overlying pavement structure.

2. A consistent and orderly interrelation of the effect of unit load, diameter of test plate, and pavement structure thickness was found.

3. For average surface course temperatures in excess of about 75 deg F, the granular base course appeared to be more effective in supporting load applied through a rigid plate than an equal thickness of bituminous concrete; for temperatures below this level the bituminous concrete surface was the more effective of the two pavement components.

4. The surcharge provided by the bituminous surface course appeared to have little effect upon the ability of the base course to support load applied through rigid plates. Likewise, the surcharge provided by the surface plus base course had little effect on the load supporting capacity of the subgrade.

5. The unit load supported by the subgrade soil at a given deflection decreased at a diminishing rate with an increase in size of loaded area up to that of a plate 64 inches in diameter, the maximum size tested.

The Road Vibration Machine was recently employed to measure pavement properties of the AASHO Road Test in northern Illinois. Dynamic measurements were made on pavement sections ranging in thickness from 11 to 19 inches, over a frequency range from 8 to 2500 cycles/second. These measurements showed that:

1. The stiffness of asphalt pavements is greatly influenced by seasonal changes.

2. Softening of the subsoil is the main factor influencing the reduction of stiffness during spring breakup.

3. When the quality of base and subbase is impaired, pavement failure is imminent.

4. The critical stiffness level for satisfactory performance is approximately 100 ton/cm.

5. Seasonal recovery seems to take place, but its magnitude needs further investigation.

6. Recovery is speeded up by traffic, but some damage from repeated loading also occurs.

7. There is good agreement between measured stiffnesses and stiffnesses calculated from velocity measurements.

8. The stiffness of pavement sections containing black (asphalt-stabilized) bases is very high.

The thickness of base and subbase courses of gravel roads has been about 85-95 cm. since 1930-1950 and from 1954, 50-80 cm. For the 1930-1950 period when no consideration was made for the frost susceptibility of soils, roads generally showed considerable frost heaving, especially in the end of rock cuts and earth cuts. The detrimental effect of frost heaving has increased since modern Finnish roads have been covered with flexible pavements. The pavements have been broken and traffic problems have resulted. Since 1954, pavement has thickness of 5 cm.; base course, 20 cm.; and subbase course 25-55 cm. according to soils’ frost susceptibility.

Because of these facts modern Finnish roads use wedges for paved surfaces. In Sweden wedges have been used since 1949: In Finland the Administration of Roads and Waterways (ARW) in 1955 issued instructions for building of wedges, and in 1956 supplementary instructions were given.

The wedges are used in Finland on the border of rock cuts, where embankments and cuts meet and where soil changes and on the each side of
culverts, when the ground is frost susceptible soil. The measurements of wedges are shown on the figures 1, 2, 3, 4, 5 and 6.

The effects on the paved roads have been studied only during two winter periods. Observations of considerable frost heavings have generally
not been made, because the paved roads are situated in South Finland, where the winter freezing index has been relative low. But in North Finland, where the freezing index is high, a few observations have been made.

Structural Section Drainage
H. R. Cedergren, California Division of Highways, Sacramento 14, California
W. R. Lovering, The Asphalt Institute, Sacramento, California

The Extension To Practice Of A Fundamental Procedure For The Design Of Flexible Pavements
G. M. Dormon, Shell International Petroleum Company Limited, London, England

The Application of Pedology to Flexible Pavement Design in Wisconsin
Dr. R. H. Keyser, University of Santa Clara, Santa Clara, California

Structural Design of Flexible Pavements in North Carolina
L. D. Hicks, North Carolina State Highway Commission, Raleigh, North Carolina

An Investigation of Flexure Cracking on a Major Highway
G. L. Dehlen, National Institute for Road Research, Pretoria, South Africa

Design of Flexible Pavements Considering Mixed Loads and Traffic Volume
W. J. Turnbull, C. R. Foster, and R. G. Ahlvin, U.S. Army Engineers, Waterways Experiment Station, Vicksburg, Mississippi

German Experiences in the Construction of Hot-mixed Asphalt Bases
Dr. H. W. Schmidt, Esso A. G., Moorburger Strasse 16, Hamburg, Germany

Development and Structural Design of Asphalt Pavements In Germany
Dr. Walter Becker, Esso A. G., Hamburg, Germany

California Method for the Structural Design of Flexible Pavements
F. N. Hveem And G. B. Sherman, California Division of Highways, Sacramento 19, California

The continued serviceability of any pavement depends upon prevention of the accumulation of excess water at all levels within the structural section and in the underlying basement soil. Standard practice should provide balanced designs in relation to the porosity or permeability of the “roof” of the structural section (the wearing course), the permeability and seepage potential of the “basement” soil beneath the structural section and the capabilities of the section for removing water. In this paper some methods are presented that make possible the development of reasonably balanced designs with drainage layers “built-in” as an Integral part of the section. The importance of having drainage layers with adequate permeability is given special emphasis to demonstrate the value of layered drainage systems incorporating a protective layer of filter material and a drainage layer of coarse crushed rock or lean open-graded asphalt mix. The ability of open-grated asphalt mixes to transmit water may be verified by observing the flow of water which occurs at the edges of open-graded wearing courses immediately after rains.

It is pointed out that damage should be thought of in developing the over-all cross-sections of pavement. The paper discusses over-all design in relation to trench sections and full-width drainage.

The elastic properties which unbound soils and granular base materials develop in situ are limited by their ability to maintain equilibrium under load stress conditions (e.g., tensile stresses), and the moduli of successive layers are in consequence determined primarily by the “geometry” of the system. For mathematical convenience all granular base layers are considered to act as a single layer for which the “effective” modulus, assumed uniform over the whole thickness will generally vary only between about 1.5 and 4 times that of the subgrade. The stiffness modulus, and breaking properties of the bitumen-bound layers will vary with time and temperature of loading. For practical conditions, however, a limiting low value of the stiffness modulus exists in situ at high service temperatures, which is governed by the mix composition. Deformation of the structure is controlled by limiting the vertical compressive strain in the subgrade and granular base, This will usually be greatest at high service temperatures for constructions incorporating thick bitumen-bound layers, or at thaw temperatures if the subgrade is frost-susceptible. Brittle and fatigue fracture of the bitumen-bound layers are governed by the tensile stress at low temperatures and the tensile strain respectively in the bottom of the layer.

On the basis of these considerations, the relative influence of the thickness and properties of the construction layers on the critical stress conditions is assessed and it is shown that the over-all advantage of a dense bitumen-hound layer over a granular layer of the same thickness will increase as their thicknesses increase. The risk of cracking of the bitumen-bound layer is found to be influenced only to a small extent by the thickness of the granular layers, but to depend mainly on the properties of the subgrade and the thickness of the bitumen-bound layer. Fatigue fracture is for example, more likely to occur on weak subgrades (particularly during thaw periods) and a small increase in the thickness of a dense bitumen-bound layer will considerably reduce the risk in this respect. Provisional design curves have been drawn up which show, for any particular subgrade, alternative constructions with dense bitumen-bound and granular base layers, meeting specific design requirements for motorways. The design thicknesses are in reasonable agreement with practical experience and lend support to the applicability of the theory and its usefulness for indicating performance trends in both conventional and unusual constructions.

(1) The development of procedures and solutions to highway engineering problems, utilizing principles and data of both pedology and soils engineering, and,

(2) The development of principles and formulae relating the design of flexible pavements to the many variables associated with Wisconsin soil conditions and climate.

Approximately 360 Wisconsin pedological soil series are grouped under 124 major series units which are based on the relationships of their engineering properties, drainage, topographic and geologic features, pedological classification units, and engineering uses. The relationships were established by field Investigations as well as laboratory test results on approximately 420 soil samples. The Pedological soil series descriptions have been rewritten to stress the engineering aspects of each series.

Theoretical and empirical concepts of flexible pavement designs were utilized to develop a family of pavement design curves. The validation of the semi-empirical curves was checked by field investigations of pavement performance which attempted to relate pavement composition, thickness, and traffic volumes to subgrade conditions. Approximately 175 subgrade soil samples were tested to aid in the analysis of the subgrade conditions.

Based on tests on many soil samples from different locations, on field investigations, and on the sciences of Pedology and Soil Engineering, recommendations for solutions to highway engineering soil problems are summarized in tables and figures.

Flexible pavement design curves have been developed indicating effects of commercial traffic, soil series, and climatic influences. The curves are developed from data obtained from theoretical and practical concepts of flexible pavement designs, and from field investigations of pavement performance.

The design wheel load may be selected from a statistical analysis of traffic data, which has been expanded for future increase, or the selection may be made on a basis of an estimate of the daily average number of vehicles of a certain type and weight, based on traffic data, that will use the pavement over a given number of years. This latter method of selection is currently used in North Carolina.

The distribution of the pressure below the surface of contact induced by the design wheel load is a determining factor in designing the total thickness of a flexible pavement, as well as the thickness and quality of its components. The Boussinesq theory of pressure distribution was used to compute the vertical pressures beneath dual wheel assemblies designed to carry loads of 4,000, 5,000, 6,000, 6,500, 7,000, 8,000, 9,000, and 10,000 pounds. The computed pressures are shown in curve form to depths of 30 inches.

The strength of the subgrade soil, base course material, and bituminous pavement mixture can be determined by any reliable method. At the present time North Carolina is using the CBR test for measuring the bearing capacity of subgrade soils and base course materials. The CBR values are converted to pounds per square inch bearing capacity by a conversion curve included in the paper. A factor of safety of 2.5 is applied to the bearing capacity of subgrades to obtain the design bearing value. Bituminous paving mixtures are designed by the Marshall Method and the Hubbard-Field Method, depending upon the type, to specified stabilities.

The paper includes several examples of design to illustrate the method.

A number of possible remedial treatments were studied in full-scale trials on the road. Thin resurfacings studied to date have not been satisfactory, cracks reappearing soon after laying. Some types of 4 inch overlay have brought about a considerable reduction in flexure, and would probably prove satisfactory solutions in practice.

It does not appear that adequate warning of such failure is given by existing design methods. Studies have been begun to improve knowledge in this regard and, if necessary, to develop an additional complementary procedure for design in the future.

The Corps of Engineers has developed criteria relating loading and strength of material (CBR) to design thickness required. These criteria are used in design of airfield pavements for a reasonable use life. The Corps has also developed a relation between design thickness required (in percentage) and load repetitions. This relation is used in design of military airfields for short-term use, in evaluating effects of short-term use of existing airfields by overload aircraft, and in recent designs for the more intensive load repetitions attendant to channelized traffic of heavy, bicycle-type landing gear aircraft.

This paper reports rather extensive developments by which the principles and relations embodied in the criteria relating loading, strength, and thickness and those relating thickness and load repetitions have been further extended and combined to provide a revised method of design for highway pavements. The method provides means for combining any composition or distribution of loading and treating any intensity of use from that of a perimeter road, highway shoulder, or residence drive to that of the most heavily traveled turnpike or freeway. The method also provides means of designing for essentially light traffic with a mingling of very heavy traffic, such as might be experienced at remote missile sites, and it permits accommodation of any desired design life.

In summary, the revised pavement design method includes development of curves showing the number of repetitions of a basic 18,000-lb, single-axle, dual-tire load equivalent to one operation of any of a range of loads from passenger cars to very heavy trucks. It also includes development of design relations for the basic 18,000-lb single-axle load in terms of CBR, thickness, and load repetitions. The first-mentioned development permits the collective evaluation of the effects of any composition of traffic and for any desired pavement-use life. The second-mentioned development provides curves from which design thickness can be determined for the total equivalent operations of the basic load arrived at in the first development. The method provides a positive treatment of load repetitions and obviates the need for selection of a single design load to represent an entire range of light to heavy loads.

A short review is given on the structural design of asphalt pavements with an asphalt base course in Germany. Experiences gathered in the construction of hot-mix asphalt bases during the last 6 years are presented. It is shown how the favorable results obtained have influenced the establishment of standard design models for asphalt pavements.

Six different asphalt base course projects built from 1956 to 1958 are described in detail. These projects include the construction of new Autobahn stretches and of sections on expressways, city streets and highways for various traffic conditions. Details given on these projects comprise the structural design and cross-section applied, the location and length of the project, the subgrade conditions, the machinery used, the composition and properties of the different layers from frost blanket to wearing course, the preparation and laying of the asphalt mixes and the behavior of the finished pavement under the existing traffic loads. Information is also presented on the reconstruction of rigid Autobahn sections using asphalt pavements and hot-mix asphalt base courses. Cross-sections applied on the Koeln-Frankfurt Autobahn in 1958 are discussed. Reference is made to tentative specifications published in 1958 recommending two methods of reconstruction.

In conclusion, the application of hot-mix asphalt bases for rigid types of pavements on new Autobahn stretches and on airfields is briefly described.

Owing to the steady increase in the number of motor vehicles, primarily of heavy trucks, it soon became evident that a renewal of the wearing surfacings was not sufficient but often the roads had to be renewed entirely. In this connection it was recognized that the base is of special importance for the performance of the road. In particular, it was necessary to replace the handpacked stone base, until then preferably used, by other construction methods which are technically better and which can be carried out more economically by employing machinery.

In order to find suitable base construction methods a number of test roads were built in Germany during the last 10 years. The paper describes these test roads briefly and Indicates the results already available,

Subsequently the paper outlines according to what viewpoints asphalt roads for heavy traffic are presently constructed in Germany. There does not exist an officially recognized thickness design procedure. When designing a pavement the experiences gained with existing roads of various thicknesses and their behavior under traffic are taken as a basis. From this basis there developed rules as to how a pavement has to be designed to withstand a certain traffic load. All roads have to be provided with a frost blanket of an appropriate thickness, unless a frost-proof subgrade exists. On a frost-susceptible subgrade road structures in Germany are generally built at least 70 cm thick. The frost blanket is the balance remaining after the deduction of the thicknesses of the base course and the surfacing.

In recent years specifications for the subgrade, the base and the surfacings of asphalt pavements have been issued, which the paper explains in detail. These specifications also indicate the thicknesses considered necessary for each construction method under a specific traffic load. The total thickness of a road structure in Germany is the sum of these individual thicknesses requirements. Respective tables have been compiled and included in the paper.

The continuous efforts for standardizing road construction methods have resulted in a restriction of possible variants contained in the Federal Specifications for the various types of construction. For instance, it has been decided to use a standard design model for “Autobahnen”. Also some States of the German Federal Republic introduced standard design models for the Federal and State roads constructed by them; these standard design models are based on the experiences the respective States have meanwhile collected and take account of special local conditions. The paper furnishes information on such standard design models.

A method is shown for converting mixed truck traffic into a single number, called the Traffic Index, which indicates the relative destructive effect of traffic. The statistical approach is used to evaluate traffic. A large sample of trucks are weighed and average axle weights determined for trucks of various axle groupings. From these average axle figures, constants are determined which can be applied to various axle load groups to determine the Equivalent Value in terms of 5000-pound wheel loads. From the EWL (Equivalent Wheel Load) calculations based on the number and weight of trucks expected to use the road, a Traffic Index may be determined. The Traffic Index is directly proportional to the thickness of structure needed to carry the anticipated traffic.

The factor which measures the resistance of a soil or granular base to deformation is determined by the Hveem Stabilometer and is known as “Resistance Value.”

The paper discusses the effect of cohesion (tensile strength) upon the performance of the pavement. Included in this phase of the paper is a discussion of equivalency of asphalt concrete in terms of inches of gravel or stone base. Data from our study of the AASHO Test Road is shown which indicates that the equivalency varies with the load of the vehicle and the strength of the asphalt mixture.

A design formula is proposed, which is as follows:

Thickness of Structure = [0.080 (Traffic Index) (90 – Resistance Value)] / (Cohesiometer Value)^O. 2

The paper discusses modifications in the California Design Formula which might be made as a result of the study of the AASHO Test Road data. Charts showing correlation with the Test Road performance are included. The formula shows a possible coefficient of correlation of 0.97 and a standard error of estimate of +/- 1.5 inches.

The paper also briefly discusses the variability of correlation statistics depending upon which layer or layers are corrected to accommodate the error.

The primary and important advantages of the California formula are:

1. The California procedure utilizes numerical values derived from physical tests of the layered system consisting of the basement soil, the subbase, base and pavement.

2. The method recognizes the effects of both structural strength and surcharge effect of the pavement and base layers.

3. The California method recognizes the effect of load repetition, individually and in combination, as well as weight and provides a logical means for converting miscellaneous traffic wheel loads to a single number, the Traffic Index, which number bears a direct linear relationship to the thickness of pavement structure required.

4. The California method has been in use for approximately 13 years and has demonstrated that it can accommodate wide variations in the type of soil, type of base and type of pavement as well as variations in wheel loads and the number of load repetitions. It also shows excellent correlation with the Test Road data on flexible pavements.

Hot-mixtures of Calcareous Soil-Sand-asphalt Type
E. A. Gonella and J. J. Font, Vialidad National, Buenos Aires, Argentina

Development in the Design and Construction of Bituminous Surfaced Pavements in the State of Victoria, Australia
A. H. Gawith And C. C. Perrin, Country Roads Board, Victoria, Australia

Asphalt Pavement in Southwestern Santa Fe Province, Argentina, with Fine Local Materials
L. M. Zalazar, Southern University and University of Buenos Aires, Buenos Aires, Argentina

Design, Construction and Evaluation Criteria of Flexible Asphalt Pavement for Airports in Italy
Giorgio Moraldi, Libero Docente in Tecnica, Delle Fondazioni E Costruzioni in Terra, Roma, Italia

The Use of Pavement Deflections in Asphalt Pavement Overlay Design
E. Zube and R. Bridges, Division of Highways, Sacramento 19, California

Jet Aircraft Runway Resurfacing
Wayne A. Thompson, Dictrict Public Works Office, U.S. Navy, San Bruno, California

Effect of Surface Color on Thaw Penetration Beneath an Asphalt Surface in the Arctic
C. W. Fulwinder And G. W. Aitken, U.S. Army Cold Regions Researchand Engineering Laboratory, Hanover, New Hampshire

Research Future of The AASHO Road Test Facility
W. E. Chastain And J. E. Burke, Illinois Division of Highways, Springfield, Illinois

Soil Stabilization with Cutback Asphalt in Southern Brazil
L. M. Zalazar, Southern University and University of Buenos Aires, Buenos Aires, Argentina
P.C. De Castro, State Highway Department of Rio Grande Do Sul. Brazil

The Application of Engineering Economy Studies to Asphalt Pavement Design
H.E. Riggs, Stanford Research Institute, Menlo Park, California

Two main purposes have been studied. First, the design of the overlay thickness that must be used to restore the old pavement on the basis of the measurement of Benkelman deflections. Second, to correlate the behaviour under traffic of a base course built of these hot-mixtures with laboratory stability test.

Dealing with the first scope, some experimental sections have been built of calcareous soil-sand asphalt reinforcements on failed pavement and the Benkelman deflections measured before and after the overlay was laid.

The decrease of deflection has shown the relative effectiveness of the reinforcement. Taking into account the over-all conditions, efforts have been made to determine the required thickness of the overlay once a given decrease of deflection has been established from an initial deflection. Proceeding with the second purpose, the behaviour of a test road has been watched where a base of the mixture referred to had been built. The factors of structural design of this road were checked and the Benkelman deflections measured from time to time. The quality of the hot-mix base was checked with Hveem and Marshall Stability tests. It was found to be a very good base, with an R Stabilometer value of 82.

Finally, some comments on the characteristics of the mixture used are made. The successful use of this mixture must be pointed out because of the great savings that have been made considering the fact that stone aggregates were not economically available within short haulage distances.

Factors which were taken into account in estimating pavement requirements from the CBR of the subgrade included consideration of the total weight of traffic expected to be carried during the designed life of the pavement as the equivalent repetitions of a 5,000 lb. wheel load modified by factors for vehicle placement on the pavement, and local climatic conditions, The chart evolved in 1945 for pavement thickness design by this method is reproduced, together with a simplified chart produced in 1949 after assembling all the factors into a single formula in terms of the number of commercial vehicles using the roadway at the time, the average annual raInfall, and the CBR of the subgrade at 95% modified AASHO compaction. To facilitate the evaluation of more soil samples without carrying out the actual CBR tests, relationships have been established between the CBR of a soil, the simple Atterberg Tests and the Group Index. Formulae for estimating the CBR from these relationships are given in the paper.

The use of various “in situ” bearing or penetration tests, as an alternative to the CBR test for estimating the CBR of the soil is referred to. These include the Proctor Needle, the Static Cone Penetrometer developed in Holland, and a Dynamic Cone Penetrometer based on a Swiss original, and relationships between values obtained with these implements and the CBR are suggested.

Current design practice makes use of the curves produced by the Road Research Laboratory of Great Britain, and these are included in the paper. The curves suggest a pavement thickness for traffic falling within certain groups, for certain CBR values of the subgrade which have been either established in situ or estimated by some other method. Where CBR values are estimated in the laboratory they are modified to take into account the annual local rainfall.

A tentative formula for estimating the additional thickness required to strengthen an existing pavement, deduced from the deflection of the pavement as measured by the Benkelman beam, is suggested, The paper makes reference to the relative values of various materials used in the pavement, and tabulates suggested design CBR values for materials having certain characteristics of grading, plasticity, etc., together with the minimum cover required over these materials.

The paper concludes with a reference to construction practices in the State of Victoria and the predominant use of light bituminous surfacing, the success of which, on pavements carrying quite high traffic volumes, has been due not only to the detailed attention paid to the design and application of the surface treatment, but to the construction compaction, and subsequently, the traffic compaction of the pavement prior to its bituminous treatment.

We have described such a region in a previous publication. It belongs to zone 2 which is one of the five typical soil profiles in Argentina. The southwestern portion of Santa Fe, wherein we have developed the design described in this article, is the best portion of the Province and is shown in the shaded area on Figure 1. Although it is an area in which there are no coarse materials like the whole of Santa Fe, sand is available in a few localized river deposits and this sand generally is composed of round polished particles. The soil profile of southwestern Santa Fe is of the typical Chermoziem type. The first horizon is a black humus silty soil good for agricultural purposes but undesirable for pavement design, The second horizon is a heavy clay, a product of the illuviation process, and retains moisture tenaciously. The third horizon is composed of the best soil of the profile with a loess soil as a parent material.

Proceeding westerly, the C or third horizon improves very much in its properties, being a sandy loam at our starting point at about the middle of the Province and becoming a pure sandy loam at the border of Cordoba Province at the extreme west. Design for asphalt pavements in this area becomes a problem of soil stabilization using soils of the C horizon for bases and subbases and a blacktop of surface treatment or preferably a bituminous mat in view of local traffic conditions. The available fine sand would permit the granular stabilization of the loess as a first step, constituting some advantage in a place of many difficulties. However, the fine round sand particles cannot provide the required internal friction. Consequently, the granular portion of the mixture has been improved by blending the local sand with sharp angular coarse sand hauled from an average of 200 miles. This process has been considered unavoidable for the base course where internal. . .

et cetera

1. For the design of total pavement thickness both FAA. and CBR design diagrams are used, but they should be adopted with caution, as for Italian soils they may lead to very different results.

2. The prescriptions for the materials to be employed in the base and subbase courses differ according to their behaviour under compaction. To assure a convenient bearing capacity even in saturated conditions, not only a low plasticity index is requested, but also definite minimum soaked CBR values. Adequate compaction is controlled by means of density tests, and for coarse grained materials by means of the modulus of deformation value, assessed with plate bearing tests using small plates.

3. Research on the behaviour of subgrades, base courses, and finished pavements by means of plate bearing tests have been performed by late Prof. Maresca, by Mr. Quaranta and by the writer. Maresca proposed a new method of interpretation of repeated plate bearing tests, in which the elastic and the plastic components of the deflection are evidenced. He has also defined a critical bearing capacity which corresponds to a sudden rise in the curve of the increments of plastic deformations. Quaranta has shown that the bearing capacity of different base and surface course layers is not only a logarithmic function of the thickness of the layer, but for each material reaches an upper limit which depends on the bearing capacity of the underlying layer. He has also given a simple formula to evaluate the ratio between the bearing capacity of an asphalt pavement and the bearing capacity of the base course. The author has shown that the validity of the well-known linear relationship between the deflection and the number of load repetitions not always holds true, as there seems to exist a fatigue limit of the structure evidenced by a non-linear relationship. He therefore invalidates the possibility of extrapolating the results of repeated plate bearing tests, and advocates the adoption of a simple conventional test based on a single cycle of load application followed by other three-load repetitions.

4. Experience on resurfacing of rigid pavements with asphalt overlays has shown that the presence of an intermediate layer of open graded bituminous concrete 7 cm thick, placed between the old pavement and the new one, is very effective in avoiding reflection cracks.

This report describes an approach used in California on some 25 projects in arriving at the required thickness of asphalt blanket using the results from field deflection tests performed on existing pavements. A brief description of testing equipment is provided along with a testing procedure now in use.

Also provided in the report are limiting critical deflection values based upon a 15000 pound single axle load for different types of construction, and traffic conditions usually associated with the primary highway system. Experience on the deflection damping capabilities of different layer thicknesses of asphalt mixtures, cement treated bases and aggregate bases is discussed. Graphs of actual deflection results obtained from highway construction are provided which show the deflection attenuation properties of each material. Although the deflection results are not definitive, they do suggest that, for the conditions of test, crushed gravel bases reduce deflections in order of 0.001 to 0.002 inches per inch of base thickness, asphalt mixtures at 70 deg F to 100 deg F reduce deflections 0.002 to 0.004 inches per inch and cement treated bases reduce deflections 0.003 to 0.005 inches per inch of applied thickness. The initial deflection of an existing pavement is very important in determining how much one inch of any material will reduce the deflection. For example, at deflection levels of 0.100 inches each inch of overlay is capable of reducing the deflection by several thousandths of an inch. At deflection levels of 0.010 inches an inch of overlay has a minute and sometimes unrecordable effect.

From this study a method is developed for selection of type of overlay with a chart of equivalent thicknesses of gravel required versus initial deflections. Also provided are some examples of application of the method.

Plate load tests, penetration tests of the old pavement surface and a study of the original construction features governed the selection, use of materials and thickness of the overlay design.

Design features were worked out cooperatively by the Asphalt Institute and Navy personnel. Features that were particularly noteworthy and debatable are: The use of all crushed rock for coarse and fine aggregate; all limestone dust for mineral filler (passing no. 200); 40-50 penetration for grade of asphalt; 98% of Marshall for in-place density.

Special emphasis was placed on control during construction with representatives from the Asphalt Institute participating by invitation. A field laboratory especially constructed for the job and assistance from specialists in the field of pavement construction with added inspection help were special features to insure adequate control. Pre-construction conferences and training also contributed to the over-all emphasis on insuring a good job. Batching operations, laydown operations, and rolling operations were carefully inspected and regulated by those concerned and experienced in Airfield Pavement Construction.

Continuing research is a part of the over-all study to determine the feasibility of using asphalt concrete for similar projects relative to reinforcing old pavements and construction of new pavements to withstand heavy loads and high tire pressures. The Naval Civil Engineering Laboratory, Port Hueneme is conducting the surveillance program assisted by the Twelfth Naval District, Public Works Office. The tests included:
(a) Unconfined compression tests at 77 deg F.
(b) Density.
(c) Penetration of Asphalt Cement.
(d) Air permeability.
(e) Maintenance costs.
(f) Traffic.

The paper concludes that although the construction work was very successful and the pavement is now functioning very well under the loads and pressures imposed it will take several years to determine the rate of hardening of the asphalt and whether or not it is possible to build a flexible pavement which will remain sufficiently flexible to resist cracking caused by small deflections of the pavement surface under the loads of small high pressure tires.

This study was one phase of an over-all program conducted by the Corps of Engineers to study the design and construction of pavements in the arctic and subarctic. A description of the site, complete with meteorological and soil data, is furnished and the investigational program conducted to study the pavement at Thule is outlined. The results obtained in 1953-1954 from a small, white-painted test area on an asphalt taxiway at Thule indicated that a reduction in depth of thaw of up to 1.8 feet was effected by painting the surface.

A thirteen hundred foot section of the main runway, where thawing into the subgrade had caused pavement subsidence, was painted white in July 1959 to permit further study of a white-painted asphalt surface and to reduce thaw penetration. The paint used was Prismo R/W white traffic paint with ground pumice mixed with the paint as a texturizing material. The paint was applied by 36-inch-wide, truck-mounted spray bars and was dry to the touch in 30 minutes.

Comparison of the subsurface temperature observations obtained under the white-painted runway area with those obtained under unpainted sections of the runway show that a reduction in thaw penetration of 1.5 feet was obtained in 1959 under the white-painted area even though almost one-third of the thawing season had elapsed prior to completion of the painting. Analysis of the 1960 subsurface temperature data showed the reduction in thaw penetration attributable to the white-painted surface to be approximately 2 feet. In both 1959 and 1960, thaw did not penetrate into the subgrade beneath the white-painted pavement.

Additional factors evaluated in the course of this investigation included the effects of normal traffic on the durability of the paint, the reactions of flight and operations personnel to the white-painted surface and any unusual maintenance problems which could be attributed to the white-painting program.

It is concluded that painting the runway surface prevented occurrence of further settlement, thereby eliminating the costly maintenance formerly required in the subsidence area. Further observations will be made to determine the continued effectiveness with time in reduction of thaw penetration.

“To determine the significant relationships between the number of repetitions of specified axle loads of different magnitude and arrangement and the performance of different thicknesses of uniformly designed and constructed asphaltic concrete, plain Portland cement concrete, and reinforced Portland cement concrete surfaces on different thicknesses of bases and subbases when on a basement soil of known characteristics.”

This first objective asked for relationships between pavement performance and pavement design variables for various axle loads.

A new concept was developed during the Road Test to define pavement performance, and the test results with respect to this objective were expressed by formulas that show the relationships that were found to exist.

The AASHO Road Test pavement experiment design was based on scientific statistical principles. Test sections were selected and located in a manner to provide unbiased estimates of the effects of traffic of known loading and frequency of application on the performance of pavements of specific designs. The structural sections of the major experiment relating to asphalt pavements comprised a a complete factorial experiment wherein the design factors were surfacing thickness, base thickness, and subbase thickness.

Within the space and funds available, only a few variables could be studied thoroughly. The experiment was designed to investigate these particular variables. In general, the variables selected for study were those of prime importance which could not be studied conveniently by other means.

et cetera

We will consider it as the stabilization process, i.e. a process that provides stable structures with a low percent of bitumen content through a cutback asphalt, an asphalt emulsion or, following very modern developments, using Asphalt Cement in a “foamed state.” If the asphalt content increases to values of the order 6% plus in terms of pure asphalt cement we are in the field of “asphalt mixes” (sand-asphalt, sand shell-asphalt, etc.), which may serve as pavement base courses when they have the required internal friction for such structure. The lubricating action of asphalt combined with the possible water content gained through time in service reduces the internal friction in terms either of CBR (Porter), “R” value (Hveem), Shearing Strength (McDowell), etc.

In conducting economy studies, it is necessary:
(1) to define the alternatives between which a decision is to be made;
(2) to identify and define those factors that may cause differences in the costs of the alternatives and to eliminate from consideration all factors that will be the same no matter which design is ultimately chosen; and
(3) to convert all factors to a common dollar basis and make a comparison over time, considering the time value of money through the use of compound interest.

The economic comparison of designs that are expected to provide equivalent pavement service Over time is relatively easy; bids or estimates will provide factual information regarding the cost of constructing the pavement structure specified by each design. The more usual circumstance–and the more difficult one to analyze–arises when different pavement designs can be expected to provide different service experience. For example, one design may permit a reduction in maintenance cost, or a lengthening of the time period until resurfacing is required; another design may require a second stage of construction at some future date. Thus, alternative designs may involve not only different expenditures for initial construction but also differing patterns of expenditures over time-and it is both the present and future expenditures that must be compared, recognizing that money has a time value.

With specific reference to asphalt pavement designs, engineering economy techniques are singularly appropriate for use in evaluating the economic implications of stage construction and of over-design of the asphalt pavement structure. A graphic method of analysis employing hypothetical numerical examples is used to illustrate the application of engineering economy to these two specific design situations.

Presentation of Summation Reports by Moderators

Session I – AASHO Road Test and Performance Criteria
W.J. Turnbull

Session II – Road Tests, Field Studies, and Performance Criteria
A.C. Benkelman

Session III – Theoretical Developments Related To Structural Design of Asphalt Pavements
William H. Goetz:

Session IV – Theoretical Developments Related To Structural Design of Asphalt Pavements
Robert G. Hennes

Session V – Strength Evaluation Of Pavement Structure Elements
H. Bolton Seed

Session VI – Design and Construction Influence on Structural Behavior of Asphalt Pavements
Gordon D. Campbell

Session VII – Design and Construction Influence on Structural Behavior of Asphalt Pavements
William S. Housel

Floor Discussion

Closing Statement: Session Chairman