Professional Documents
Culture Documents
Sr61e PDF
Sr61e PDF
OFFICERS
R. R. BARTELSMEYER, Chairman C. D. CURTISS, First Vice Chairman
WILBUR S. SMITH, Second Vice Chairman
FRED BURGGRAF, Director WILLIAM N. CAREY, JR., Assistant Director
Executive Committee
REX M. WHITTON, Federal Highway Administrator, Bureau of Public Roads (ex officio)
A. E. JOHNSON, Executive Secretary, American Association of State Highway Officials
(ex officio)
LouIs JORDAN, Executive Secretary, Division of Engineering and Industrial Research,
National Research Council (ex officio)
PYKE JOHNSON, Retired (ex officio, Past Chairman 1960)
W. A. BUGGE, Director of Highways, Washington Department of Highways (ex officio,
Past Chairman 1961)
R. R. BARTELSMEYER, Chief Highway Engineer, Illinois Division of Highways
E. W. BAUMAN, Director, National Slag Association, Washington, D. C.
DONALD S. BERRY, Professor of Civil Engineering, Northwestern University
MASON A. BUTCHER, County Manager, Montgomery, County, Md.
J. DOUGLAS CARROLL, JR., Director, Chicago Area Transportation Study
C. D. CURTISS, Special Assistant to the Executive Vice President, American Road
Builders' Association
HARMER E. DAVIS, Director, Institute of Transportation and Traffic Engineering, Uni-
versity of California
Duxs W. DUNBAR, Attorney General of Colorado
MICHAEL FERENCE, JR., Executive Director, Scientific Laboratory, Ford Motor Company
D. C. GREER, State Highway Engineer, Texas State Highway Department
JOHN T. HOWARD, Head, Department of City and Regional Planning, Massachusetts
Institute of Technology
BURTON W. MARSH, Director, Traffic Engineering and Safety Department, American
Automobile Association
OSCAR T. MARZKE, Vice President, Fundamental Research, U. S. Steel Corporation
B. MCMORRAN, Superintendent of Public Works, New York State Department of
Public Works
CLIFFORD F. RASSWEILER, Vice President for Research and Development, Johns-Manville
Corporation
GLENN C. RICHARDS, Commissioner, Detroit Department of Public Works
C. H. SCHOLER, Applied Mechanics Department, Kansas State University
WILBUR S. SMITH, Wilbur Smith and Associates, New Haven, Conn.
B. WOODS, Head, School of Civil Engineering, and Director, Joint Highway Research
Project, Purdue University
Editorial Staff
Report 5
Pavement Research
By the
Special Report 61 E
I'
NATIONAL ADVISORY COMMITTEE
This committee was appointed by the Highway Research Board to advise the
Board and its project staff in relation to administrative and technical matters.
K. B. Woods, Chairman
Head, School of Civil Engineering, and
Director, Joint Highway Research Project, Purdue University
W. A. Bugge, Vice-Chairman
Director, Washington Department of Highways
iv
Preface
The AASHO Road Test was conceived and
sponsored by the American Association of
State Highway Officials as a study of the per-
formance of pavement and bridge structures of
known characteristics under moving loads of
known magnitude and frequency. It was ad-
ministered by the Highway Research Board of
the National Academy of Sciences—National
Research Council, and was considerably larger
and more comprehensive than any previous
highway research study.
This is the fifth of a series of major reports
on the AASHO Road Test. The first four re-
ports are "History and Description of Project",
"Materials and Construction", "Traffic Opera-
tions and Pavement Maintenance", and "Bridge
Research". Additional reports in the series in-
clude "Special Studies" and "Final Summary".
This report is. a comprehensive summariza-
tion of all research related to the test pave-
ments. It describes briefly the experiments and
experiment designs, test traffic, measurement
programs and data collection procedures,
methods of analysis, and the results of the
analyses.
It is presented in three major parts. The
first is general in nature, describing the back-
ground and objectives of the research, facilities
and methods of operation employed, and the
"present serviceability" concept of pavement
performance evaluation. The second and third
parts describe the flexible and rigid pavement
experimental designs and research and report
the findings.
Acknowledgments
Personnel from many organizations, indus-
trial firms, and institutions of higher learning
assisted in carrying out the AASHO Road Test.
- It is impractical to list the names of all indi-
viduals who participated in this report. How-
ever, the efforts of the following are particu-
larly acknowledged:
The Bureau of Public Roads of the U. S.
Department of Commerce, together with the
V Department of Defense, for technical advice
and services in a great many areas.
The Illinois Division of Highways for tech-
nical advice and services of personnel from its
headquarters in Springfield, Illinois, and Dis-
trict 3 office in Ottawa, Illinois, and for provid-
ing its Resident Task Force.
The Minnesota Department of Highways and
-: the Indiana State Highway Department for ex-
tensive cooperation and assistance to the Per-
formance Rating Panel.
• Purdue University, the University of Illinois,
and Lehigh University for technical advice and
services.
The Portland Cement Association, The As-
: phalt Institute, and the several States for par-
ticipation in materials testing programs.
V The following organizations for the services
of resident observer-consultants: The Asphalt
Institute, the Portland Cement Association, the
American Trucking Association, the Canadian
Good Roads Association, the Department of
V
Highways, Province of Ontario, Canada, and
the German High;ray Research Board.
The American Petroleum Industries for tech-
nical advice and services.
V V V
V
General Motors Corporation for equipment,
personnel and technical advice in performing
skid resistance experiments.
V
Shell Oil Company for equipment, personnel
V
and technical advice in dynamic testing of flexi-
ble pavements.
Definitions of Terms and Symbols
The following definitions are for terms and symbols that appear fre-
quently or that may have special connotations in this report. Terms in
parentheses represent abbreviated or alternative versions that are some-
times used for the defined term. Wherever appropriate, symbols for the
term are given after the definition.
Axle Load Applications (Cumulative load ap- 21 to 81 lb per 100 sq ft, the amount depend-
plications, load applications, applications) : The ing upon slab thickness (see Table 38). No
total number of axle loads that have crossed the expansion joints were provided except adjacent.
pavement sections from Day 836 to any later to structures.
day. .................................. W
Nonreinforced Pavement: Plain portland ce-
Axle Load Applications on an Index Day: ment concrete pavement with doweled trans-
Wt verse contraction joints spaced at 15 ft (see
Table 38). No expansion joints were provided.
Weighted or Unweighted Applications: If W
has been determined with the use of a seasonal Base (base course in Road Test usage)
weighting function, W represents weighted ap- Crushed stone, gravel, cement-treated or as-
plications; if the seasonal weighting function is phalt-treated sand-gravel material (subbase
not used, W represents unweighted applica- material) immediately under the surfacing ma-
tions. Unweighted applications can be con- terial. (At the Road Test base was used only
sidered to be determined by a seasonal weight- in the flexible pavements.)
ing function whose value is always 1. Reference
to applications implies that the applications are Base Thickness (base): The thickness of
unweighted unless the context specifies weighted base, expressed in inches ........... D2 (Flex.)
applications. (Where no base is specified the material im-
mediately under the surfacing course is either
subbase or embankment soil.)
Pavement Structure
Subbase (subbase course in Road Test
Pavement Structure (pavement): One or usage).: The layer of graded sand-gravel mate-
more layers of specially processed materials rial between the surface of the embankment
overlying the embankment soil. soil and the base course (or surfacing course
when there is no base course).
Flexible Pavement: A pavement structure
generally consisting of asphaltic concrete sur- Subbase Thickness (subbase) : The thickness
facing, base and/or subbase. of subbase, expressed in inches. ..... D3 (Flex.
or Rigid)
Rigid Pavement: A pavement structure con-
sisting of portland cement concrete surfacing, Embankment (embankment soil): The pre-
with or without subbase. pared soil underlying the pavement structure.
Surface: The visible portion of a pavement. Pavement Design (design, structure de-
sign): The specifications for materials and
Surfacing (surface) : The layers of asphaltic thicknesses of the pavement components (for
concrete or portland cement concrete material flexible pavements usually abbreviated D1 , D2,
upon which traffic operates.
Surface Course: The uppermost layer of as- Thickness Index: A linear combination of
phaltic concrete surfacing. pavement components (flexible pavements)
that expresses pavement design as a single
Binder Course: The layer of asphaltic con- number ................................ D
crete underlying the surface course. D = a,D1 + aD2 + a3D3 (Coefficients a1, a2
Surfacing Thickness (surface thickness, sur- and a, depend on the analysis in which the
face, slab thickness (rigid)) : The thickness thickness index is to be used.) In certain cases
of surfacing material, usually expressed in a subscript has been used with D; e.g., a crack-
inches. .........................D1 (Flex.) ing index..............................
D., (Rigid)
Pavement Sections, Experiment Designs
Reinforcement (surfacing reinforcement)
A qualitative variable for rigid pavement used Structural Section: A two-lane section of test
to distinguish between plain portland cement pavement of the same design on both sides of
concrete and surfacing reinforced with wire the centerline.
mesh. .......................... D1 (Rigid)
Test Section (section) : A one-lane section of
Reinforced Pavement: Portland cement con- test pavement that has the same load assign-
crete pavement with doweled transverse con- ment for its full length, and the same design
traction joints spaced at 40 ft and containing throughout (except in the case of wedge sec-
welded wire fabric in amounts varying from tions and flexible paved shoulder sections).
viii
DEFINITIONS OF TEEMS AND SYMBOLS (Continued)
Subsection: One of four 40-ft segments of Cracking Index, Flexible Pavement: The
160-ft test sections in the special flexible pave- number of axle load applications before Class 2
ment experiments. cracking appeared (see Section 3.2.3.1). . . . W
Wedge Section (special base section): A flex- Cracking Index, Rigid Pavement: The total
ible pavement structural section whose base cracking. in the surface of a test section. . . . C'
course varies in thickness at a uniform rate Patching: The area of a test section, patched
throughout its length. with asphaltic concrete, expressed in square
Replicate Sections: Two test sections or sub- feet of patching per 1,000 sq ft of surfac-
sections having the same pavement design and ing..................................... P
load assignment. Pumping: The ejection of water and subbase
material or embankment soil from beneath the
Experiment Design: A set of test sections pavement surfacing.
that form the basic units for controlled varia-
tion in pavement design or load factors. Pumping Index: An index used classifying
the severity of edge pumping, expressed in
Factorial Experiment (Design 1, factorial cubic inches of material per inch of pavement
sections) : A set of test sections (flexible or edge. (The term "pumping score"—Equivalent
rigid) for which, in each loop, all possible com- to 100 x the pumping index—was used for con-
binations of selected levels appear for the three venience in record keeping.)
pavement design factors, D,, D2 and D.
Rut Depth (rut) : The maximum vertical dis-
Paved Shoulder Studies (Design 2): A set of placement of a point of the surface measured
test sections (flexible pavement) in which as- from the center of a 4-ft transverse straight-
phaltic concrete shoulder paving decreases in edge................. ................ RD
width uniformly from 8 ft to zero from end to (The mean rut depth, RD, taken in both wheel-
nd in the direction of traffic. paths at several points longitudinally along a
test section was used in most analyses).
Paved Shoulder—Sub baseStudies (Design
3): A factorial experiment (rigid pavement) Slope: A profilometer estimate of the angle
in each main loop for which the design vari- between a horizontal plane and a line joining
ables (and their respective levels) are paved two surface points 9 in. apart longitudinally in
shoulders (either present or absent), a 6-in. the wheelpaths.
subbase (either present or absent), and sur-
facing thickness (at two levels). Slope Variance: The variance (mean square
deviation) of a set of slopes about the mean
Special Base Studies (Design 4): The set of slope.................................SV
wedge sections (flexible pavement) where the (The mean slope variance in the two wheel-
thickness and type of base material were paths of a section, SV, was used in most analy-
varied. ses).
Subsurface Studies (Design 5): a set of sec- Faulting: The vertical displacement of the
tions (flexible and rigid) in the non-traffic loop surface of a portland cement concrete pave-
provided for determining seasonal changes in ment at one side of a joint or crack relative to
subsurface conditions. the slab surface on the other side of the joint
or crack measured at the center of each wheel-
Surface Treatment Study (Design 6): The path and expressed in tenths of inches per
set of flexible pavement sections whose surfac- 1,000 ft of wheelpath.
ing was a bituminous treatment.
Frost Heave: The vertical displacement of
the surface of the embankment soil or of any of
Accumulated Behavior of Pavements the structural pavement components associated
with volume change due to freezing.
Cracking: The amount of cracking in the
pavement surface that existed at any given Serviceability Rating: The judgment of an
time, expressed in square feet of cracked area observer as to the current ability of a pave-
per 1,000 sq ft of surface area (flexible) or in ment to serve the traffic it is meant to serve.
linear feet of projection (see Section 3.2.3.1) PSR
per 1,000 sq ft of surface area (rigid). Usually
classifled according to severity, as follows: Serviceability Index: An estimate of the
mean of serviceability ratings made by a panel
Cracking, for serviceability indexes. ..... C of judges. A present serviceability index
lx
DEFINITIONS OF TERMS AND SYMBOLS (Continued)
formula is used to determine the estimate of Rebound Procedure Deflection: The dif-
the serviceability rating of a section . ....... p ference between the probe elevation after re-
moval of the load and the probe elevation when
Serviceability Index on an Index Day. . . . the load is opposite the probe (flex.)
Smoothed Serviceability Index: An average Static Rebound Deflection: The differ-
of serviceability values for several weeks be- ence in elevation of the probe when the load is
fore and after a given index day. ......... Pt static and opposite the probe and when the load
Serviceability Trend: A continuous graph of is removed. (Used in all rigid pavement tests
smoothed serviceability values plotted against with the beam reference on the shoulder.)
axle load applications. Static Rebound Edge Deflection: A test
Roughness Index, Flexible: The logarithm of with the probe midway between 15-ft trans-
(1 + slope variance). verse joints in nonreinforced pavement and
10 ft from jointsin reinforced pavement, and
Performance: The serviceability trend of a with the load (center of the dual wheels of the
test section with increasing number of axle ap- test vehicles) 20 in. from the pavement edge.
plications. -
Performance Data: Selected pairs of coor- Static Rebound Corner Deflection: A
dinates from a serviceability trend, usually test with the probe on a rigid pavement edge at
equally spaced on either a time basis or on a a joint and with the load 20 in. from the edge
serviceability basis. ................ p, W or (see Fig. 145) ......................... d'
p,logW
Dynamic Deflection: Deflection caused by
Out-of-Test Section (failed section): Any a moving vehicle and measured at a point with
section whose serviceability trend fell to 1.5 a linear variable differential transformer. . . . d
before the end of the test traffic (1.0 for De-
sign 6). Partial Deflection: The deflection meas-
ured under a 2-ft longitudinal chord at the
In-Test Section: Any section whose service- bottom of the deflection basin (flex.)
ability index was greater than 1.5 at the time
of observation (1.0 for Design 6). Corner Deflection: Deflection of a rigid
pavement edge 6 in. from a joint with the
center of the wheel load 20 in. from the edge
Transient Behavior of Pavements (see Fig. 138) ..........................
Corner Movement: Maximum vertical dis- Temperature: Unless otherwise specified,
placement, in inches, of a point on a rigid pave- temperature refers to temperature, in °F, of a
ment edge at a joint, occurring as the air tem- point within a pavement structure as measured
perature changes from a maximum to the next with a thermocouple.
minimum, or from a minimum to the next
maximum.............................. mc Air Temperature: Temperature of air, in
°F, at a point 5 ft above the ground........ U
Curling (in Road Test usage): Changes in
the shape of a rigid pavement surface over a Temperature Differential (in Road Test
period of maximum movement found to be usage) : The difference in temperature at the
about 12 hours. top and bottom of a 61/2 -in, concrete slab. . . . T
Deflection: The difference in elevation of a Strain: The unit change in length of an ele-
point on or in the pavement before and after a ment of the pavement surface between the
specified condition of loading. loaded and unloaded conditions. Strain may be
tensile or compressive. It is usually expressed
Beam Deflection: Deflection taken with a in microinches per inch ...................
Benkelman beam whose probe is at the point of
measurement. Pressure (embankment pressure): The unit
pressure, in lb per sq in., transmitted through a
Creep Speed Deflection: Deflection re- pavement structure under dynamic load to a
corded when the load approaches and leaves cell located in the surface of the embankment
the probe point at creep speed (2 to 3 mph) soil. ................................... p
(flex.)
Normal Procedure Deflection: The dif- Data and Analysis
ference between the probe elevation before Data System: A collection of data that may
loading and the probe elevation when the load contain initial observations, summarized data,
is opposite the probe (flex.) or results from analyses. Data in a particular
x
DEFINITIONS OF TERMS AND SYMBOLS (Continued)
system are generally in one or more of three tion less than one-half or more than twice the
forms: analog records, IBM cards and print- corresponding value calculated from the equa-
outs, or folders containing basic data and sum- tion.
maries. A four-digit code is, used to identify If only one mean log residual is used to
each Road Test data system. establish the error band, it can be expected that
the band will contain about 60 percent of the
Mathematical Model (model): An assumed residuals in log Y.
algebraic form for the relationship among par-
ticular experimental variables. The model in- Effects (explained effects): Changes, or
cludes constants whose values are to be deter- average changes, in an observed variable that
mined by analysis procedures, and may also are attributed to changes in one or more of the
include constants whose values are assumed. controlled factors of an experiment.
In many cases the models involve logarithms.
Throughout this report logarithms are always Residual Effects (unexplained effects, resi-
to the base 10. dual variation): Changes in an observed vari-
able that are attributed to changes in unidenti-
Residual: The difference between the ob- fied variables. Usually expressed in terms of
served value of an experimental variable and average residuals.
the value computed for this variable from a
model in which all constants are determined Statistical Significance: An (explained)
and specific values are assigned to all remain- effect is said to be statistically significant if its
ing variables. magnitude relative to an appropriate average
residual is so large that there is little risk that
Root Mean Square Residual: The square the explained effect is actually a residual effect.
root of the average squared residual. In gen- The assumed risk that the explained effect is
eral the divisor for this average is equal to the really a residual effect is called the significance
number of residuals less the number of con- level, usually selected to be no more than '5
stants determined for the model by the data. percent.
rmsr or Linear Model: A model whose general form
rms error is A0 + A 1X1 + A 2X2 ... + AkXk = 0, where
Average Residual (mean residual): The X1, X2,.. ., X. are functions of one or more ex-
average of the absolute values of all residuals. perimental variables, and at least one of the
In general, the divisor for this average is equal constants A0, A1 ,. . ., A k is to be determined
to the number of residuals................. r through data analysis.
Least Squares Linear Regression Analysis
Mean Log Residual: Whenever an equation (regression analysis) : A mathematical proce-
is derived for the logarithm of a variable, Y, dure for evaluating the Undetermined con-
instead of for Y itself, residuals and mean resi- stants in a linear model when it is assumed
duals represent discrepancies between observed that the best coefficients are those which lead to
and calculated logarithms of Y. Because it is a minimum for the sum of squared residuals
generally assumed that about 90 percent of in- for a particular experimental variable.
dividual residuals are less than twice the mean
residual, nine-tenths of the scatter of the ob- Total Variation: For a particular experi-
servations log Y, around the calculated values, mental variable, the sum of squared deviations
log Y, is contained in the band whose limits are of the values for the variable from the mean
log Y = logY + 2, where log'Y is calculated value of the variable.
from the derived equation and where is the
mean log residual. Squared Correlation Coefficient: For linear
It is often useful to express the error models the difference between unity' and the
limits in the original units of Y. By taking ratio of total squared residuals to total varia-
antilogarjhms these limits are given by Y = tion for a particular variable. Sometimes said
(102r) (Y), Thus nine-tenths of the scatter of to be that fraction of the total variation which
Y around Y is expressed as two percentages of is explained by the regression equation. ....r2
', (10) (Y) and (10'2) (p). For example,
if 7 = 0.16, then 10 032 = 0.48 and 10032 = Correlation Index: For the case of nonlinear
2.09, so that approximately nine-tenths of models the correlation index serves to indicate
the observed values for Y are found between the degree of correlation between observed
0.48 and 2.09 times the corresponding calcu- values of an experimental variable and corre-
lated values for Y. Roughly speaking, it would sponding values predicted from a derived equa-
be unusual, in this example, to find an observa- tion. Thus, the correlation index is a general-
xi
DEFINITIONS OF TERMS AND SYMBOLS (Continued)
ization of the squared correlation coefficient If /3 = 1 the trend is linear; if /3 > 1 the trend
that is used for linear models. The correlation decreases more rapidly with increasing appli-
index is computed by subtracting from one the cations; if /3 < 1 the trend decreases less
ratio of sum of squared residuals to the total rapidly with increasing applications. ....... /3
variation in the observations for the variable.
Rho: In the Road Test model for perform-
Beta: In the Road Test model for perform- ance, a function of (D1, D2, D,, L1 , L2 ) that
ance, a function of pavement design and load equals the (computed) number of load applica-
variables (D1, D2, D 3, L1, L2 ) that determines tions required to bring the serviceability level of
the shape of a computed serviceability trend. a pavement to 1.5.........................p
xii
Contents
Acknowledgments..................................................................vi
Definitions of Terms and Symbols .................................................vii
Chapter 1. General Information
1.1 Background and Objectives ........................................ 1
1.1.1 History ................................................... 1
1.1.2 Intent of the AASHO Road Test ............................. 2
1.1.3 Objectives ................................................. 2
1.1.4 Objectivity of Findings ...................................... 3
1.1.5' Applicability of Findings ................................... 3
1.2 Facilities and Operations .......................................... 3
1.2.1 Site Location .............................................. 3
1.2.2 Test Facilities ............................................. 4
1.2.3 Construction ............................................... 7
1.2.4 Test Traffic ................................................ 7
1.2.5 Measurement Programs...................................... 8
1.2.6 Pavement Maintenance ...................................... 8
1.2.7, Environmental Conditions ................................... 8
1.3 Pavement Serviceability and Performance .......................... 10
1.3.1 Relation to Objectives ........................................ 10
1.3.2 Rating of Pavements in Service ............................. 10
1.3.3 Present Serviceability Index ................................ 12
1.3.4 Pavement Performance Data ............................... 15
1.3.5 Procedures for Analysis ..................................... 17
1.4 Needed Research—General ......................................... 18
1.4.1 Modifications of Performance Relationships ................... 18
1.4.2 Generalization and Extension of Relationships ................ 18
1.4.3 Serviceability of Pavements .................................. 19
Chapter 2. Flexible Pavement Research
2.1 Description of Flexible Pavements ................................20
2.1.1 Experiment Designs and Layout .............................20
2.1.2 Materials and Construction .................................20
2.2 Pavement Performance ...........................................23
2.2.1 Serviceability Index for Flexible Pavement ...................23
2.2.2 Performance as a Function of Design and Load .................27
2.2.2.1 Main Factorial Experiments (Design 1) ...............27
2.2.2.2 Paved Shoulder Studies .............................43
2.2.2.3 Special Base Type Experiments .....................46
2.2.3 Structural Deterioration ......................................53
2.2.3.1 Transverse Profile Changes ...........................58
2.2.3.2 Cracking ...........................................79
xlii
xiv CONTENTS
3.3.5 Dynamic Corner Deflection as a Function of Design and Load .... 192
3.3.6 Static Edge Deflection as a Function of Design, Load and Tem-
perature .......................................................193
3.3.7 Static Corner Deflection as a Function of Design, Load and Tem-
perature ...................................................... 194
3.3.8 Strain and Deflection as Functions of Speed .................... 194
3.3.9 Variation of Deflection Across Loops ......................... 196
3.4 Prediction of Performance from Strain or Deflection ................ 198
3.4.1 Performance from Dynamic Edge Strain ..................... 199
3.4.2 Performance from Static Edge Deflection ..................... 200
3.4.3 Performance from Static Corner Deflection .................... 201
3.5 Auxiliary Studies ................................................ 203
3.5.1 Overlays ................................................... 203
3.5.2 Subsurface ................................................. 204
3.5.2.1 Strength and Condition Data, Loop 1 .................. 204
3.5.2.2 Trenching Program, Loops 3-6 ....................... 206
3.5.3 Curling of Concrete Slabs..................................... 206
3.5.3.1 Changes in Internal Temperature Distribution with Time 211
3.5..3.2 Instrumentation of Corner Movement and Curling of Con-
creteSlabs ............................................... 214
3.5.3.3 Movement at Corner with Changing Temperature Condi-
tions...................................................... 216
3.5.3.4 Typical Curling of Slabs in Corner Region ............. 222
3.5.4 Load Stresses in Surface of Concrete Slabs ................... 228
3.5.4.1 Instrumentation ..................................... 229
3.5.4.2 Field Procedures and Data Processing ................. 231
3.5.4.3 Typical Stress Distributions .......................... 232
3.5.5 Moisture and Temperature Coefficients of Expansion ........... 237
3.5.6 Serviceability Changes, Non-Traffic Loop ..................... 239
3.6 Summary of Findings and Needed Research ......................... 239
3.6.1 Summary of Findings ........................................ 239
3.6.2 Needed Research ............................................ 242
Appendix A. Pavement Performance Data ........................................ 243
Appendix B. Miscellaneous Data .............................., .................... 280
Appendix C. Creep Speed Deflection Data .......................................... 281
Appendix D. Test Procedures and Equipment ........................................ 283
Appendix E. Formulas from Elastic Theory Used in Connection with Section 3.5.4 289
Appendix F. The Pavement Serviceability-Performance Concept ..................... 291
Appendix G. A Rationale for Analysis of Pavement Performance ................... 307
Appendix H. Embankment Soil Test Correlations ................................... 323
Appendix I. Data Systems ....................................................... 326
Appendix J. Committees, Advisory Panels, and Project Personnel ..................... 346
Regional Advisory Committees ....................................... 346
Region1 ...................................................... 346
Region2 ....................................................... 346
Xvi CONTENTS
Chapter 1
General Information
1.1 BACKGROUND AND OBJECTIVES test and published the results as HRB Special
Report 4.
1.1.1 History Concurrently, the Committee on Highway
The events leading to the three most recent Transport of the American Association of State
large-scale highway research projects, Road Highway Officials recommended that additional
Test 1—MD, the WASHO Road Test and the road tests be initiated by the regional members
AASHO Road Test, are described in detail in of the Association. As a result, the Western
AASHO Road Test Report 1, "History and De- Association of State Highway Officials spon-
scription of the Project" (HRB Special Report sored the WASHO Road Test, consisting of a
61A). The following is a summary of these number of specially-built flexible pavements in
events and the activities of the AASHO Road Idaho tested in 1953-54 under the same loads
Test. used in the Maryland test. The results of this
test, also conducted by the Highway Research
For many years the member states of the Board, were published as Special Reports 18
American Association of State Highway Offi- and 22.
cials had been confronted with the dual problem In March 1951, the Mississippi Valley Con-
of constructing pavements to carry a growing ference of State Highway Engineers had
traffic load and establishing an equitable policy started planning a third regional project. How-
for vehicle sizes and weights. The Association ever, the idea of another regional project of
recognized the common need for factual data limited extent was abandoned in favor of a
for use in resolving the problem. Therefore, in more comprehensive road test, to be sponsored
September 1948, it set up a procedure for initi- by the entire Association. In October, comply-
ating and administering research projects to ing with a request by the Association, a High-
be jointly financed by two or more states. way Research Board task committee submitted
In December of the following year a meeting a report, "Proposal. for Road Tests," after
was held at Columbus, at the request of the which the Association appointed a working
Governor of Ohio, to consider the problem of committee to prepare a prospectus. on. the proj-
vehicle weight and its effect upon existing and ect. By December it had been decided to in-
future pavements. The conference was attended clude bridges in the research.
by representatives of the Council of State Gov- In June 1952, the Working Committee pro-
ernments and highway officials of 14 eastern duced a report, "AASHO Road Test Project
and midwestern states. The need for more Statement." In July it selected a site for the
factual data concerning the effect of axle loads project near Ottawa, Ill. In January 1953, it
of various magnitudes on pavements was con- submitted a second report, "AASHO Road Test
firmed. Project Program," and in August 1954, a third'
As a result, Road Test 1—MD was conducted entitled "Project Program Supplement." In
in 1950. An existing concrete pavement in May 1955, this committee produced its fourth
Maryland was tested under repeated applica- and final report "Statement of Fundamental
tion of two single- and two tandem-axle loads. Principles, Project Elements and Specific Di-
The Highway Research Board administered the rections."
THE AASHO ROAD TEST, REPORT 5
Meanwhile, in March 1953, AASHO had reviewed by the various committees, and sub-
formulated a plan for prorating the cost of the mitted to.the Board.
project among its member departments and, The field office .for the project was closed in
later, had received assurances of 'participation January 1962. . However, the Highway Re-
from the States, the Automobile Manufacturers searchBoard agreed to continue certain studies
Association, the Bureau of Public Roads and associated with the Road Test pavement per-
the Ameriëañ Petroleum Institute, while the formance analyses in its Washington office. The
Department of Defense had agreed to furnish results of these studies will be reported by the
military personnel for driving the vehicles. Highway Research Board.
On February 22, 1955, the Highway Re-
search Board with the approval of its parent 1.1.2 Intent of the AASHO Road Test
organization, the National Academy of Sciences The following formal statement of the intent
—National Researëh Council, accepted from of the Road Test was approved by the Execu-
the Association the responsibility to administer tive Committee of the Highway Research Board
and direct the new project. The Board opened January 13, 1961:
a field office at Ottawa, Ill., in July 1955;and The AASHO Road Test plays a role in the
in August a task force of the Illinois Division total engineering and economic process of provid-
of Highways moved to the site to undertake the ing highways for the nation. It is important that
preparation of plans and to prepare for the this role be understood.
construction of the test facilities. The Road Test is composed of separate major
experiments, one relating to asphalt concrete
In March 1956, the Board appointed the Na- pavement, one relating to portland cement con-
tional Advisory Committee as its senior ad- crete pavement, and one to short span bridges.
visory group and in April selected a project There are numerous secondary experiments. In
each of the major experiments, the objective is
director. to relate design to performance under controlled
In June 1956, the National Advisory Com- loading conditions.
mittee passed a resolution recommending that In the asphalt concrete and portland cement
concrete experiments some of the pavement test
the Executive Committee of the Highway Re- sections are underdesigned and others overde-
search Board consider the inclusion in the signed. Each experiment requires separate
facility of a fifth test loop to be subjected to analysis. Eventually the collection and analysis
light axle loads. This resolution, recommended of additional engineering and economic data for
a local environment are necessary in order to
by the Bureau of Public Roads, was based on develop final and meaningful relations between
the pending enactment of the Federal Aid pavement types.
Highway Act .of 1956. In July, the Executive All of the short span bridges are underde-
Committee of the Board approved this change signed. Each is a separate case study.
Failures and distress of the pavement test
and made. additional changes involving special sections and the beams of the short span bridges
studies areas. The final layout of thetest facili- are important to the success of each of the ex-
ties is described in Section 1.2.2. periments.
The Highway Research Board of the National
Construction of the test facilities began in Academy of Sciences—National Research Council
August 1956, and test traffic was inaugurated has• the responsibility of administering the proj-
on October 15, 1958. Test traffic was operated ect for the sponsor, the American Association
until November 30, 1960, .at which time 1,114,- of State Highway Officials, within the bounds of
the objectives of the test. The Board is also
000 axle loads had been applied to the pave- responsible for collecting engineering data, de-
ment and the bridges. veloping methods of analysis and. presentation of
A special studies program was conducted in data, preparing comprehensive reports describ-
ing the tests, and drawing valid findings and con-
the spring and. early summer of 1961 over some clusions. It is here that the role of the Highway
of the remaining test sections. Strains, defiec- Research Board ends.
tions and pressures were. measured in these As the total engineering and economic process
studies under a widevariety of vehicle types, of providing highways for the nation is developed,
engineering data from the AASHO Road 'Test
load suspensions, tires and tire pressures. and engineering and economic data from many
Special military vehicles, included at the re- other sources will flow to the sponsor and its
quest of the Army, as well as highway con- member departments. It is here that studies will
struction equipment, were included in these be made and final conclusions drawn that will be
helpful to the executive and legislative branches
tests. The results of the studies are presented of our several, levels of goveriiment and to the
in Road Test Report 6. highway administrator and engineer.
During 1961, the research staff concentrated
on analysis of the test data and the preparation 1.1.3 Objectives
of reports. Each of the major reports was ap- The objectives of the AASHO Road Test as
proved by a review subcommittee of the Na- stated by the National Advisory Committee
tional Advisory. Committee and later submitted were as follows:
to the entire National Advisory Committee and
the Regional Advisory Committees prior to its 1. To determine the significant relationships
between the number of repetitions of specified
publication by the Highway Research Board. axle loads of different magnitude and arrange-
All reports were completed by the project staff, ment and the performance of different thick-
GENERAL INFORMATION 3
nesses of uniformly designed and constructed of. these variables. In general, mathematical
asphaltic concrete, plain portland cement con- models were used to represent associations
crete, and reinforced portland cement concrete
surfaces on different thicknesses of bases and among experimental variables, then statistical
subbases when on a basement soil of known methods were employed to determine constants
characteristics. for the models as well as to describe the relia-
To determine the significant effects of speci- bility of the evaluated models. Thus experi-
fied vehicle axle loads and gross vehicle loads
when applied at known frequency on bridges of mental designs and analytical procedures were
known design and characteristics. developed in order to obtain unbiased estimates
To make special studies dealing with such of the effects (and the statistical significance
subjects as paved shoulders, base types, pave- of many of the effects) of controlled experi-
ment fatigue, tire size and pressure, and heavy
military vehicles, and to correlate the findings of mental factors. The designs and procedures
these special studies with the results of the basic did not, however, make it possible to obtain
research. effects for other factors that were either held
To provide a record of the type and extent constant or that varied in an uncontrolled
of effort and materials required to keep each of
the test sections or portions thereof in a satis- fashion, for example, embankment soil,
factory condition until discontinued for test pur- strength of materials, and environmental con-
poses. ditions. Although estimates were obtained for
To develop instrumentation, test procedures, the effects of axle load and axle configuration,
data, charts, graphs, and formulas, which will
reflect the capabilities of the various test sec- it was not possible to determine the statistical
tions; and which will be helpful in future high- significance of these effects because replication
way design, in the evaluation of the load-carrying of load or configuration was not provided.
capabilities of existing highways and in deter- Nevertheless, particularily in the cases of load
mining the most promising areas for further
highway research. effect on both pavement types and axle con-
figuration effect on rigid pavement the differ-
This report deals primarily with work done ences observed were so great as to leave
in connection with Objectives 1 and 5 and with practically no doubt that the effects were sig-
some of the special studies mentioned in Objec- nifieantly greater than zero.
tive 3. Material relating to Objective 2 will be Basic data will be made available to other
found in Road Test Report 4 and Objective 4 groups equipped to perform independent anal-
is discussed in Report 3. Other special studies yses. Further analyses are to be encouraged by
suggested in Objective 3 are discussed in Re- the Highway Research Board in the expecta-
port 6. tion that the over-all usefulness of the project
will be enhanced.
1.1.4 Objectivity of Findings
Discussion of the results given in this report 1.1.5 Applicability of Findings
has generally been limited to specific relation- The findings of the AASHO Road Test, as
ships derived from the data. Restraint has stated in the relationships shown by formulas,
been exercised in expressing opinions, conjec- graphs, and tables throughout this report, re-
tures, and speculations. Conclusions have been late specifically to the physical environment of
drawn only when supported by data acquired the project, to the materials used in the pave-
during the tests. ments, to the range of thicknesses and loads
At the request of the National Academy of and number of load applications included in
Sciences a panel of statisticians was appointed the experiments, to the construction techniques
in 1955 so that professional advice was avail- employed, to the specific times and rates of
able for both the designs of the Road Test ex- application of test traffic,. and to the climatic
periments and for the procedures by which the cycles that occurred 'during construction and
experimental data would be analyzed. It was testing of the experimental pavements. More
not the function of this group to select vari- specific limitations on certain of the findings
ables nor levels for variables to be included in are given in the discussion of results in various
the Road Test. This was the responsibility of sections ,of this report. Generalizations and
the National Advisory Committee, acting. upon extrapolat'ions of these findings to conditions
the recommendations of the original AASHO other than those that existed at the Road Test
Transport Committee's Working Committee. should be based upon experimental or other
The Statistical Panel played an important role evidence of the effects, on pavement perform-
in influencing the experimental layout through ance of variations in climate, soil type, ma-
its recommendations for complete factorial de- terials, construction practices and traffic.
signs, randomization, and replication. Its rec-
ommendations, accepted by the Advisory
Committee; made possible effective studies of 1.2 FACILITIES AND OPERATIONS
the relationships sought by the objectives.
Within the space, time and funds avail- 1.2.1 Site Location
able, only a few variables could be studied The location of the AASHO Road Test was
thoroughly. The experiment was designed and near Ottawa, Ill., in LaSalle County, about 80
the test facilities built specifically for the study mi southwest of Chicago (Fig. 1). The test
4 THE AASHO ROAD TEST, REPORT 5
facility was constructed along the alignment of elevated turnarounds with 42-ft radii. Center-
Interstate Route 80. The site was chosen be- lines divided the pavements into inner and
cause the soil within the area was uniform and outer lanes, called lane 1 and lane 2 respec-
of a type representative of that found in large tively.
areas of the country, because the climate was All vehicles assigned to any one traffic lane
typical of that found throughout much of the of Loops 2 through 6 had the same axle
northern United States, and because much of arrangement-axle load combinations. No traffic
the earthwork and pavement construction could operated over Loop 1. In all loops, the north
ultimately be utilized in the construction of a tangents were surfaced with asphaltic concrete
section of the National System of Interstate and south tangents with portland cement con-
and Defense Highways. crete. All variables for pavement studies were
concerned with pavement designs and loads
1.2.2 Test Facilities within each of the 12 tangents. Each tangent
was constructed as a succession of pavement
The test facilities consisted of four large sections called structural sections. Pavement
loops, numbered 3 through 6, and two smaller designs, as a rule, varied from section to sec-
loops, 1 and 2. Test bridges were at four loca- tion. The minimum length of a section was
tions in two of the large loops. The layout of 100 ft in Loops 2 through 6, and 15 ft in Loop
the six test loops, the administration area and 1. Sections were separated by short transition
the Army barracks is shown in Figure 2. pavements. Each structural section was sepa-
Each loop was a segment of a four-lane rated into two pavement test sections by the
divided highway whose parallel roadways, or centerline of the pavement. Figure 3 shows
tangents, were connected by a turnaround at the layout of two typical test loops and loca-
each end. Tangent lengths were 6,800 ft in tions of the test bridges.
Loops 3 through 6, 4,400 ft in Loop 2 and 2,000 Details of the experiment designs are given
ft in Loop 1. Turnarounds in the major loops in Report 1 and are summarized in Sections
had 200-ft radii and were superelevated sO that 2.1.1 and 3.1.1 of this report. Details concern-
the traffic could operate over them at 25 mph ing all features of bridge research are given in
with little or no side thrust. Loop 2 had super- Road Test Report 4.
GENERAL INFORMATION 61
'1
FRONTAGE ROAD
PROPOSED
J us_I
LOOP
_!OOP 6
6
ROAD BARRACKS \ \
\\Ottawa
\1
S
'Utica" SCALE—MILES
I
-; z -:
Figure 2. Layout of AASHO Road Test.
Cll
PRESTRESSED
__________ CONCRETE
Test Tangent
-1
Rigid
_I1I
x
Test Tangent
STEEL I-BEAM
LOOP 5
REINFORCED
CONCRETE
Test Tangent
'-Z Rigid
r-'-
AL 4
-.
-
4
- -.. -
-... ...;
. . --
- •k- - -•'•: -. __ . ..• -..
414.
Figrure 6. driver iIlarti (Wallace lariarh-).
GENERAL INFORMATION fj
negotiated through normal Illinois contractual Figure 7. Typical test vehicle axle loadings.
channels. It was started 'in late summer 1956
and completed in time for test traffic to begin
in the fall of 1958. 'S. J. Groves'and,Sons'was in lane 1 of Loop 2 and eight vehicles in lane 2
the principal contractor in a joint venture with of Loop 2. In January 1960, the traffic was
Arcole Midwest,.Inc., in the, embankment con- increased to ten vehicles in each lane of Loops
struction and with Rock Roads, Inc., as a sub- 3' through 6, six in lane 1 and. 12, in. lane .2 of
contractor for asphaltic concrete surfacing. Loop 2. These vehicle distributions were , se-
Valley Builders, Inc., built the bridges. lected in order that axle load applications could
be accumulated at the same rate in each of the
1.2.4 Test Traffic ten traffic lanes.
A detailed description of the operation of the All lanes had identical ' specifications for
test traffic is presented in Road Test Report 3. transverse placement, speed, and rate of axle
As previously stated, Loop 1 was not subjected load accumulation. Tire pressure and steering
to test traffic. One lane of this loop was used axle, loads were representative of normal prac-
for subsurface and special load studies, the tice. Some of the vehicles' were 'gasoline and
other for observing the effect of environment others diesel powered.. Further information
on pavements not subjected to traffic. The re- concerning the vehicles is contained in Road
maining five loops, 2 through 6, were subjected Test' Reports 1 and 3.
to traffic for slightly more than two years. Whenever possible, traffic was operated at
Every vehicle in any one of the ten traffic lanes 35 mph on the test tangents. Traffic was sched-
had the same axle load and axle configuration. uled to operate over an 18-hr, 40-mm period
The assignment of axle loads and vehicle types each day, 6 days a week, except that during the
to the various lanes is shown in Figure 7. first 6 months of 1960 the schedule was ex-
The vehicles were loaded with concrete 'blocks tended to 7 days a week. The schedule was
that were anchored down with steel bands and maintained except .when pavement distress,
chains. Although the traffic phase was inaugu- truck breakdowns, bad weather and certain
rated on October 15, 1958, early operation other causes made it impossible. A total accu-
indicated the need to readjust the test loads. mulation of 1,1 14,000 axle-load applications was
This delayed full-scale traffic until November attained during the 25-month traffic testing
5, 1958. From November 1958 to January 1960 period. To accomplish this, soldiers of the U. S.
controlled test traffic consisted of six vehicles Army Transportation Corps.Road Test Support
in each lane of Loops 3 through 6, four vehicles Activity drove more than 17 million miles.
THE AASHO ROAD TEST, REPORT 5
1.2.5 Measurement Programs from the bridge research are listed in Appendix
Each measurement program was designed to A, Road Test Report 4.
'accomplish one or more of the following pur- The text of this report contains many refer-
poses:. (1) to furnish information at regular ences to data systems whose contents are per-
and frequent intervals concerning the rough- tinent to the discussion. These references are
ness and visible deterioration of the surfacing explained in Appendix I. For example, a refer-
of each section; (2) to record early in the life ence to Data System 5121, or simply DS 5121,
of each section transient load effects that might is explained in Appendix I as containing all
be directly correlated with the ultimate per- routine Benkelman beam deflection data for
formance of the section; and (3) to furnish a flexible pavement sections on the traffic loops
limited amount of additional information which with an IBM printout of the data available on
might contribute to a better understanding of request.
pavement mechanics. Specific measurement programs are de-
Programs falling in the first category were scribed in the appropriate sections of Parts 2
and 3.
concerned with measurements of permanent
changes in the pavement profile along and 1.2.6 Pavement Maintenance
across the wheelpaths, as well as the extent of Detailed descriptions of maintenance criteria
cracking and patching of the surfacing. These and procedures are given in Road Test Report
measurements were given major emphasis since 3. Complete maintenance histories of each test
they were used to define the performance of section are available in DS 6300.
each section as required by the first Road Test The objectives of the Road Test were con-
objective. cerned with the performance of the test
Programs falling in the second category in- sections as constructed. Consequently, mainte-
cluded the measurement of strains and defiec- nance operations were held to a minimum in
tions which became the basis for estimating any section that was still considered under
pavement capability, as required by the fifth study. When the "present serviceability" (see
objective. Section 1.3) of any section dropped to a speci-
Finally, programs of the third category en- fied level the section was considered to be out
compassed such measurements as the severity of test and maintenance or reconstruction was
of pumping of rigid pavements, changes in performed as needed.
layer thickness in flexible pavements, pavement Since the prime objective of the maintenance
temperatures, subsurface conditions, and nu- work was to keep test traffic operating as much
merous other measurements. as possible, minor repairs were made when re-
In general, measurements were restricted to quired regardless of weather or time of day.
those variables that had been demonstrated by The use of pierced steel landing mats permitted
previous research to be related significantly to traffic to operate through a complete driving
pavement performance. A further restriction, period so that more conventional repairs could
applying especially to subsurface studies, was be made during the daily 5-hr, 20-min traffic
imposed by the overriding necessity to keep the break.
test traffic moving. All repairs were made with flexible-type
In spite of these restrictions, a formidable pavement material. Deep patches and recon-
amount of data was accumulated, and special struction consisted of compacted crushed stone
electronic systems were evolved to facilitate the base material surfaced with hot-mixed as-
storage and initial processing of the data. For phaltic concrete. Overlays consisted of asphal-
example, in the case of some programs, means tic concrete. Thin patches were made either
were provided to record automatically in the with hot-mix or cold-mix materials. Crushed
field the desired information directly on per- stone base material and cold-mix surfacing
forated paper tape, thus eliminating the task were stockpiled at several locations on the proj-
of the manual reading of analog records. In ect, and hot-mix asphaltic concrete was gen-
another case, an electronic device was used to erally purchased from a nearby contractor.
read field analog records and to punch the in- As a general rule, pavement maintenance
formation on paper tape for immediate trans- was done by project forces with project-owned
ference to an electronic computer. In general, equipment. However, in the critical spring
automatic data handling was used wherever periods of 1959 .and 1960, it was necessary to
possible and the majority of the data were augment the project maintenance forces with
stored on IBM cards. additional men and equipment.
Data from the various measurement systems
were classified into data systems, and a particu- 1.2.7 Environmental Conditions
lar system was identified by a four digit code. The topography of the Road Test area is
Appendix I lists major Road Test data systems level to gently undulating with elevations vary-
concerned with pavement research and notes ing from 605 to 635 ft. Drainage is provided
how the systems may be obtained from the by several small creeks which are tributaries
Highway Research Board. Major data systems of the Illinois River. Surface drainage, how-
GENERAL INFORMATION
00
AVERAGE MONTHLY MAXIMUM
80
20
I?O
inn-,
IQ'
flee
I.JO
I Ifl
macn I IuI
TRAFFIC PERIOD I
ever, is generally slow. Geologic information usually represented by samples exhibiting A-6
indicates that the area was covered by ice characteristics.
during several glacial periods and that the In the interest of uniformity, soil making up
subsurface soils were deposited or modified the top 3 ft of embankment directly under the
during these periods. Surface soils were sub- test pavements was taken from borrow areas
sequently derived from a thin mantle of bess near the project. This soil, underlying the sur-
deposited during a post-glacial period and were face stratum, was shown by tests to have a
reasonably uniform in the area of the project. plasticity index from 11 to 15, a liquid limit
Soil drainage is generally poor. Bed rock is from 27 to 32, and a grain size distribution of
found 10 to 30 ft below the surface. 80 to 85 percent finer than the 200 mesh sieve,
58-70 percent finer than 0.02 mm and 34-40
The upper layer of soil was from 1 to 2 ft percent finer than 0.005 mm. Maximum dry
thick and consisted generally of A-6 or A-7-6 densities were in the range 114 to 118 lb per
soil with similar characteristics. The adjacent cu ft and optimum moisture contents in the
underlying stratum was usually from 1 to 2 ft range of 14 to 16 percent when compacted in
thick and most of this material was fairly accordance with standard procedure, AASHO
plastic A-7-6 soil. Substratum layers were T99-49.
12
The climate of the Road Test area is temper- function of a pavement is to serve the traveling
ate with an average annual precipitation of public. Briefly, it was considered that a pave-
about 34 in. of which about 2.5 in. occurs as ment which maintained a high level of ability
25 in. of snow. The average mean summer to serve traffic over a period of time was
temperature is 76 F and the average mean superior in performance to one whose riding
winter temperature is 27 F. The soil usually qualities and general condition deteriorated at
remains frozen during the winter with alter- a more rapid rate under the same traffic. The
nate thawing and freezing of the immediate term "present serviceability" was adopted to
surface. Normally the average depth of frost represent the momentary ability of a pavement
penetration in the area is about 28 in. to serve traffic, and the performance of the
Summaries of climatological data observed at pavement was represented by its serviceability
weather stations on the project are given in history in conjunction with its load application
Figures 8 through 10 and frost depth informa- history.
tion in Figure 11. Depth of frost under the Though the serviceability of a pavement is
test pavements was obtained by means of * to be determined subjectively,
patently a matter
special instrumentation involving the measure- a method for converting it to a quantity based
ment of electrical resistance of the soil as de- on objective measuremeñts.is given in the next
scribed in Highway Research Abstracts, Vol. two sections. Since the Road Test was con-
27, No. 4. More detailed climatological and frost cerned only with the structural features of the
information is available in the form of IBM pavement, such items as grade, alignment, ac-
listings in Data Systems 3300, 3301, 3140 and cess, condition of shoulders, slipperiness and
3240. Figure 12 summarizes the observations glare were excluded from consideration in
made at the project on the elevation of the arriving at a value for pavement serviceability.
\.vater table under the test pavements and adja- The serviceability of each test sectin was
cent natural ground. determined every two weeks during the traffic
testing phase, and performance analyses were
1.3 PAVEMENT SE]. VICEABILITY AND based on the trend of serviceability with in-
PERFORMANCE creasing number of load applications. The
serviceability-performance concept is described
1.3.1 Relation to Objectives in detail in Appendix F.
The first objective of the Road Test (see
Section. 1.1.3) asks for relationships between 1.3.2 Rating of Pavements in Service
the.performance of the pavement and the pave- Serviceability was found to be influenced by
ment design variables for variousloads. In longitudinal and transverse profile as well as
cept was
order to define performance, a new don the extent of cracking and patéhing. The
evolved founded on the principle that the prime amount of weight to assign to each element in
lOG
8C
AVERAGE MO NL HE MAXINUM
60
D
I
0 1
-Jo
V
LiZ
- 10
I- Li
o 20 _IL _
0 Frost in Embank ment*
U.
U 30
* No Pavement * Under Subbos
U 0- Structure Material Only
40 I I
q,: NV
hV\
2 30
0.
40
ow
-II
WO
co z 0 11~ 7i7l
I
I-U
0-0
OIL
c 20
(flu)
0
Cr
UZ
30
4U
U-I>
Frost Under Rigid Pavement
40 I L
I'igure 11. Frost (lepth.
12 THE AASHO ROAD TEST, REPORT 5
C I I I I 0 I I I I I
I - I I I
I I I I
I
PAVEMENT SURFACE PAVEMENT SURFACE
RFACE - Loop2 Loop3 -
I-
w 4
w
II.
z
I-
a
Lu
0
8
I I I
12 2
lQ5R ioso iqsn 1401 1458 050 loOn CCL 050 1Q50 OCA CCI
0 0
N
PA EMENT SLI,,RFACE,,-,-~'Loo~ PAVEMENT SURFACE
Loop 6 -
I-
uJ4 4 4— -
Lu
LA.
rL
8
2 I I I I I I
I'
lOOP ICOC
12 I
1958 1959 000 1001
the determination of the over-all serviceability members was taken as the section's present
is a matter of subjective opinion. Furthermore, serviceability rating. Some of the sections were
the degree of serviceability loss to be associated rated more than once in order to determine the
with a given change in any one of these ele- ability of the panel to repeat itself. Road Test
ments depends on subjective judgment. To ob- field crews then measured variations in longi-
tain a good estimate of the opinion of the tudinal and transverse profiles, as well as the
traveling public in these subjective matters a amount of cracking and patching of each sec-
Pavement Serviceability Rating Panel was ap- tion.
pointed. This panel included highway designers,
highway maintenance men, highway adminis- 1.3.3 Present Serviceability Index
trators, men with materials interests, trucking Through a conventional statistical procedure
interests, automobile manufacturing interests (multiple regression analysis) it was possible
and others. These men made independent rat- to correlate the present serviceability rating
ings of the ability of 138 sections of pavement, with the objective measurements of longitudi-
located in three states, to serve high speed, nal profile variations, the amount of cracking
mixed truck and passenger traffic. Both rigid and patching and, in the case of flexible pave-
and flexible pavements were included, and cer- ments, transverse profile variations (rutting).
tain sections were selected for rating in each of For either type of pavement this analysis re-
five categories ranging from very poor to very sulted in a formula that used pavement
good. The members were instructed to use measurements to compute a "present service-
whatever system they wished in rating each ability index" which closely approximated the
pavement and to indicate their opinions of the mean rating of the panel.* The necessary
ability of the pavement to serve traffic at the measurements and serviceability index compu-
time of rating on a scale ranging from 0 to 5
* A detailed discussion of the work of the Rating
with adjective designations of very poor (0-1),
poor (1-2), fair (2-3), good (3-4), and very Panel, including the ratings, the data obtained in the
measurements of the sections that were rated, and the
good (4-5). For each section the mean of the derivation of the present serviceability indexes is pre-
independent ratings of the individual panel sented in Appendix F.
GENERAL INFORMATION 13
--
Z .' A-7
--
SV = i—I
n-1
in which
SV slope variance;
X1 = the ill, slope measurement; and
n total number of measurements.
The slope variance for each section was cal-
culated by the digital computer directly from
the tape output of the chart reader. For use by
other agencies, the Road Test staff has devel-
For Smooth Pavement oped a simplified proIllometer (Fig. 17),
Figure 15. Typical longitudinal profilometer record. designated the C1-ILOE Profilometer, whose
output is slope variance. Thus, neither a chart were required to make the measurements on all
reader nor a digital computer is required when sections for determining p,'.)
the CHLOE Profilometer is used. The serviceability history of each section was
It was found that of the several types of converted to a "smoothed serviceability his-
measurements used in the serviceability index tory" by a moving average that included at
formulas, longitudinal profile variation of a least three (generally five) successive index
section of pavement when represented by the values except that the end values for the
logarithm of the slope variance correlated most history were sometimes taken as end values for
highly with the rating of that section by the the smoothed history. Typical serviceability
panel. data and smoothed serviceability histories are
shown in Figure 18.
1.3.4 Pavement Performance Data The number of axle applications applied
As stated in Section 1.3.1, pavement per- (luring the t index period, averaged over the
formance analyses were i)ased on the trend of ten traffic lanes, was represented by nt, and the
the serviceability index (determined at inter- total number accumulated through that period
rals of two weeks, or more often when re- by N, thus,
quired) with increasing axle applications.
Prior to use in the analyses, performance data (2)
were identified and processed. It was observed early in the traffic phase of
Each 2-week period was termed in "index the Road Test, confirming experience else-
period'', and the last day of each period was where, that for sections of insufficient design
called in ''index day''. Index days were num- relative to load, the rate at which pavement
bereci sequentially from I to 55, the first oc- damage accumulated with applications of load
curring on \ovember 3, 1958, and the fifty-fifth was affected by seasonal changes, especially in
n had
o n \ovember 30, 1960. Because all section.-, the case of flexible pavements. The design of
been subjected to almost the same number of the Road Test experiment did not permit a
applications of axle loads on any given date, clearcut comparison of the damage rate in the
the pairing of all index value with all index day various seasons since sections which failed in
was equivalent to specifying the serviceability one season were not available for observation
index corresponding to a given number of axle during subsequent seasons. Nevei'theless Table
applications. The symbol p,' was used to repre- 1, giving the percentage of failures occurring in
sent the serviceability index of any section as each season for each type of pavement, sug-
determined l)v measurements made on the t' gests that the damage rate was relatively low
index day, and the plot of p,' versus time was in winter for both types of pavement and
termed the "serviceability history" of a section. relatively high in spring for flexible pavements.
(Usually the last three days of all index period Changes in the effect of load with seasons
16 THE AASHO ROAD TEST, REPORT 5
SMOOTHED SERVICEABILITY -
4
0
81-WEEKLY SERVICEABILITY
- INDEXES
\_p-1.5
0
- Flexible Pavement Sections .
I I I I I I I I I I I I I I I I
_T -
BI-WEEKLYSE RVICEABILITY-- SMOOTHED SERVICEABILITY -
HISTORY
-pI.5 -
— Rigid Pavement Sections
C I I I I I I I I I I I I I I I I
NOV. JAN. MAR. MAY JUtY SEPT. NOV JAN. MAR. MAY JUtY SEPT. NOV.
1958 1959 1960
Figure 18. Typical serviceability histories.
A. gives
1.3.5 Procedures for Analysis G=f3 (log W— log p) (5)
The analyses of performance resulted in Plotting G against log W for Eq. 5 gives a
empirical formulas wherein performance was straight line whose slope is /9 and whose inter-
associated with load and pavement design vari- cept on the log W axis is log p. For each Road
ables. To use mathematical procedures for the Test section the performance data given in
analyses it was necessary to assume some Appendix A were converted into values for G
analytical form or model for these associations. and log W and a straight line was fitted to the
In addition to the experimental variables the G. log W points. From these straight lines,
models include constants whose values were estimates of /3 and log p were obtained for each
either to be specified or to be estimated from test section. For the cases where the service-
the data. Thus the analytical procedures were ability loss was very small over the traffic test-
for the estimation of constants whose values ing period /9 may be nearly zero and log p
were unspecified in the model—constants that extremely large. Special rules were applied for
indicate the effects of design and load variables these cases in order to obtain logical values of
upon performance. The procedures also in- p and log p (see Appendix G).
cluded methods for estimating the precision The assumed relationship between /9 and the
with which the data fit the assumed model. The design and load variables was
procedures used in the Road Test analyses are - B0 (L1 + L2)B2
set forth in detail in Appendix G. +
(D + a2D2 + a3D3 + a4 )B1 L2B8
There are many different mathematical
forms that could be used as models for service- (6)
ability trends, and several of these may fit the in which
data with more or less the same precision.
Different models were tested for goodness of /30a minimum value assigned
=
fit to the Road Test performance data. Pref- to /9;
erence for one model over another was gov- = the nominal load axle
erned mainly by relative goodness of fit, but weight in kips (e.g., for
consideration was also given to relative agree- 18,000-lb single axle load,
ment with highway design practice and experi- L1 = 18; for 32,000-lb tan-
ence for traffic conditions beyond the Road dem axle load, L1 = 32);
Test.
The mathematical model ultimately chosen L2 = 1 for single axle vehicles,
for both the flexible and rigid pavement anal- 2 for tandem axle vehicles;
yses is of the form D1, D2 and D3 = the three pavement design
factors surfacing, base
(4) and subbase thickness for
(W
P )11 flexible pavement and re-
in which inforcement, slab thick-
ness andsubbase thickness
c1 ~ p ~ c0; for rigid pavement.
p = the serviceability trend value; The remaining symbols of Eq. 6 are positive
c0 = the initial serviceability trend value constants whose values were either to be as-
(for the Road Test c0 = 4.5 for rigid signed as was done for P. or to be estimated by
pavements, and 4.2 for flexible pave- means of the analysis.
ments—these values were the means Equations in this same form were deter-
of the initial serviceability of test mined from analysis of the rigid pavement data
sections); and the flexible pavement data, respectively.
THE AASHO ROAD TEST, REPORT 5
The analysis rationale assumes that estimates For example, it might be assumed that /3 is a
for /3 from the equation are better than esti- constant,
mates based only on the individual section per-
formance data. Consequently, the values of /3 f3=b0 (8)
estimated from the equation were used in con- or that /3 is a simple function of p, for example,
junction with the data to obtain new estimates
of log p for every test section. (9)
The algebraic form assumed for the associa- p b2
tion of p with the design and load variables is
The concept of a thickness index for flexible
AO (D + a4)1L2A3 pavements might be generalized after further
P = (7) research to a "structural index," S, where S
( L1 + L2)A2
would account for all pavement layers (their
where D (=a,D1 + a2D2 + aD,) represents a thicknesses and strengths) as well as the em-
"thickness index" of the pavement, L1 and L2 bankment soil. A single index for vehicle load,
are as defined for Eq. 6, and the remaining L, might be introduced so that L could account
symbols are constants whose values are either for all axle loads (including steering axles) and
to be assumed or to be estimated from the their spacing. Then it might be assumed that
analysis.
Evaluation of the constants in Eqs. 6 and 7
is reported in Section 2.2.2 for flexible and 3.2.2
for rigid pavements. so that the structural index is squared relative
Eqs. 6 and 7 when evaluated and used in to the load index. It may be noted that the ratio
conjunction with Eq. 5 thus represent the first of A2 to A2 in Eqs. 18 and 21 (see Section 2.2)
goal of the Road Test—to associate perform- is already of the order two to one, so that Eq.
ance with design and load variables. 10 appears to be a reasonable assumption at
At various stages in the development of the least for flexible pavements.
equations, tests were made for the significance As is explained in Appendix G, performanceS
of pavement design factors, and statistics were equations developed for the present report re-
computed to express the degree of correlation suit from a step-by-step fitting procedure where
between observations and corresponding pre- the results of one step are used as input for the
dictions from the equations. Finally, average next step. Modification of the fitting procedures
residuals were used to indicate the extent to will likely take the form of an over-all pro-
which observations were scattered from the cedure that determines all unassigned constants
corresponding calculated values of p and log W. simultaneously as a particular residual cri-
Average residuals, correlation indexes, and in- terion is minimized. Once the over-all fitting
ferences from the significance tests are sum- procedure is developed, the residual criterion
marized after presentation of derived equations can include both residuals from log W estimates
in Sections 2.2.2 and 3.2.2. and residuals from p estimates. Moreover, per-
Many different models and fitting procedures formance data from experiments that have
were studied and one selected from which the been analyzed separately in this report may be
performance equations fit the Road Test data combined in an effort to obtain a more general
with satisfactory precision. In time, other analysis.
models may be found that also fit the data satis- Although it was not possible to investigate
factorily and which may prove equally or more modifications of the type just described in time
useful. for inclusion in this report, the Highway Re-
search Board will undertake these studies. It
is hoped that further effort will produce modi-
1.4 NEEDED RESEARCH—GENERAL fied equations that can represent all the Road
Test performance data with at least the same
14.1 Modification of Performance Relation- precision as given in this report and that
ships simplifications can be introduced with little
Any further effort by the Highway Research sacrifice in precision over the equations re-
Board to fit a mathematical model to the Road ported herein.
Test performance data will likely involve modi-
fications either in the basic models for p, /3, and 1.4.2 Generalization and Extension of Rela-
p, or in the fitting procedures, or in both. It is tionships
the purpose of this section to mention several Discussion in the preceding subsection re-
possibilities for both types of modification that lates to the need for additional study of the
are contemplated in further work with the per- data obtained in the Road Test. A larger area
formance data. for future research, involves the extension of
Even if no changes are made in Eq. 4, it is the performance equations to include para-
possible to modify the formulas for /3 and p. meters that were not varied in the project. It
GENERAL INFORMATION 19
is important to know, for example, the effects serviceability trends observed under the normal
on pavement performance of variations in the traffic using the facility. A careful record of
characteristics of the soil and the materials the number and magnitudes of axle loads over
used in the pavement structure. The effects the test sections would be required.
of environment need study. Not only the dif- These experiments would provide for verifica-
ferences in performance associated with the tion of the coefficients. in the Road Test per-
existence of heavy rainfall, desert conditions, formance equations and for the inclusion of
frost, etc., must be considered, but also the terms in the equations relating to variables that
differences that may be associated with dif- were not under study in the AASHO Road Test.
ferent rates of traffic application and distribu- More specific areas for study in the satellite ex-
tion of axle loads in the traffic stream. (For periments are discussed at the ends of Chapters
example, at the Road Test a million axle loads 2 and 3.
of one weight were applied in two years to Field tests would be simple pavement per-
each section. What would have been the situa- formance experiments, with 2 or 3 test sections
tion had these loads, accompanied by several each, constructed as part of normal highway
million lighter loads, been applied in 20 years?) construction in a large number of locations
Studies designed to fill these gaps may fall where only one or two variations from normal
in four categories: (1) theoretical studies, (2) pavement design would be observed along with
major satellite studies, (3) field tests, and (4) the normal design. These studies would prove
laboratory tests. very useful to engineers who must use judg-
There should be continuing encouragement ment in the application of Road Test findings
of research into the mechanical and physical and in their attempts to evaluate new designs
laws involved in pavement performance. Only and new materials. However, the field tests
through such theoretical work will there be would not be designed in such a way as to per-
developed rational mathematical models by mit analyses that would result in important
which performance can be related to the funda- modification of the Road Test equations them-
mental properties of materials and to the selves. Many states have constructed test pave-
dynamic characteristics of the loading. ments in the field test category in the past. If
Since the completion of such theoretical work traffic records are available, further study of
appears to be years away, immediate attention these pavements would be extremely useful.
should also be given to means for extending the Laboratory tests are those needed in the
empirical models developed at the Road Test study of materials characteristics as they might
to include additional important parameters. A affect pavement performance. Here again more
most effective device for this purpose is the detailed recommendations are given at the ends
so-called satellite study. These studies have of Chapters 2 and 3.
been described* as relatively small road tests
in different parts of the country (and other 1.4.3 Serviceability of Pavements
countries) involving consideration of variables
most of which were not included in the AASHO It is believed that the serviceability-perform-
Road Test. A very important finding of the ance concept developed at the Road Test has
Road Test was that, within the range of pre- added a new technique of value in the design
cision of measurements systems and estimation and maintenance of highway pavement. It is
techniques available, no significant interactions emphasized, however, that the specific service-
were found among the design variables. There- ability indexes developed for the Road Test,
fore, in the design of satellite experiments were based on very small samples of the Ameri-
where the variables are like those in the Road can highway network by a very small group of
Test (structure thickness, base type, etc.) highway engineers. There is no reason to think
balance in the experiment can be attained that more extensive sampling will result in
through the use of partial rather than full major modification of these indexes, but if the
factorials.** This means that to test a given system is to receive widespread use, it is im-
number of variables any satellite experiment perative that other groups, working under the
will require only a small fraction of the test same rules as the Road Test Rating Panel,
sections that would have been required had make subjective ratings of many sections of
the AASHO Road Test shown that significant pavement over the entire country containing
interactions existed. many types of distress leading to loss of serv-
Such satellite experiments are also different iceability. Accompanying these rating sessions
from the Road Test in that traffic is not a vari- should be objective measurements of those ele-
able. The test sections would be constructed as ments that may be involved in serviceability
part of the regular highway system and their such as, slope variance (roughness), rut depth,
cracking, faulting, patching, and slipperiness.
* "Extending the Findings of the AASHO Road Test" Regression analyses of the ratings in terms of
before the Design Committee, AASHO, at the AASHO the objective measurement data will produce
meeting in Denver, Cob., October 1961.
** See Ham, R. C., and Irick, P. E., "Fractional Fac- new more generally applicable serviceability
torial Analysis," HRB Road Test Conference, May 1962. indexes.
Chapter 2
o
863. 864 6 4 iT Ti
mrr
6 j_ 624 9 8 471 ii
__
:: jp_
8
869 870 - - __ a
829 830 i iT Th I LL JL J. J! 320
6 837
825
838
826
o T ir 0 L L AtI
12 2 603 604 -
3 8 1 481
12 2 443
-
444
3 JL 1
16 1 2 315 316
0 109 110 -
851 852 -
6 A 2 589 590 A 2 455 456 2 307 308
8 --- ---
3 .1
- -
-r 8
2 -r 6 8
2 ., 6 12
2 305 -. 3097
819 820 -
-
...
12 3 575 576 12 3 425 426 16 I 327 328
16 821 822 _L i _J_ _i2.
0 823 824
- 4 2 595 596 84 2 437 438 8 2 313 314
o _L 6 9
16
-
877 878
6 ..L J.L. JL
._.. 12 I 625 626 12 I 477 478 16 3 265 266
- - 1. _!_ J_ ._1I. 4 2 605 606 4 2 439 440 8 2 297 298
5 _Q. IL ZL 0 8 3 587 588 3F422 39TW 336
8 849 850 0 4 2 J.L. JL 12 I 621 622 12 I 479 480 16 3 255 256
879 880 .J1 .J±.
6 - - - 4 3 579 580 4 3 423 424 8 I 325 326
02145146 8
3 i JL JL
3 I
Subsurface Studies 4 - - '-' 12 2 593 594 '-' 2 2 445 446 '
16 2 301 302
Design 5 - 1. J.?.L J.L
.
- - T - - T 475 76 T - r 264
162
o—-L-
49 I To 8 2 591 592
IiI272I
Test Section No. 8 2 447 448 12 2 311 312
OC 41 = 0 4 3 123 124 12 3 581 582 12 3 427 428 16 1 I 1 333 1 334
Lane Lane = = =
8 2 1 =
139 1 140
I
= Shoulder Paving Study Shoulder Paving Study Shoulder Paving Study
2 Shoulder ravng Study
0 861 862 Design 2 -
Design
- D es gn
.JesIgn
i . esIgn
831 832
o 8
CL
4,TestSectionNo. 0 =
ction No.
CC
a'
C CC
Test Section No.
- CC
a'
c OC
Test Section No.
_
16 817 818 75 753
3 [6 0 855 856 - rn • Lane I Lane .CQ.
Lane .co. . or
Lane Lane
no. °.r .
Lone Lone
8 845 846 JL 2 i- __
177 I 178 L
63 638 435 436 291 292
~,6 . ——
23O- 3O4 - 438 75 .
4.
161 16?J I.r
2 3 8 1751176 3 6 20 AL.L 2931294
713 - - - _____
t8, 84 - - JL- - - . ..aii I 2Z.
639 640 295 296
4 30 5 0 4 4 2 0
__ -._J_ ±Q1.LQ LI..gI
. Base Type Study. Base Type Study Base Type Study Base Type Study
N t e. Sh aded sec tons are replicates rtn;,,,, A rte;,,nA flni,,., A
TABLE 4
CHARACTERISTICS OF MATERIALS, FLEXIBLE PAVEMENTS'
Asphaltic
Crushed Cement Asphalt Concrete
Grave 1
Item Subbase Stone Treated Treated
Base Base Base Base Surface Binder
Mix Mix
'Identification: Subbase, uncrushed natural sand-gravel; Crushed stone base, crushed dolomitic limestone;
Gravel base, uncrushed natural gravel; Treated bases, asphalt cement or portland cement and subbase material.
'AASHO T99-57.
'Laboratory density using Marshall procedure.
'Before subgrading.
'85-100 penetration grade asphalt.
'Type I portland cement.
FLEXIBLE PAVEMENT RESEARCH 23
placed on the roadway and rolled in 3-in, lifts p = 5.03 - 1.91 log (1 + SV)
to the required density. Construction operations - 0.01\/C + P - 1.38RD2
were limited to short blocks as in the case of the
subbase and embankment soil. (11)
Data concerning the materials used in the in which
construction of the special bases for Design 4
are also given in Table 4. Certain of these sec- p = the present serviceability index;
tions had gravel bases, some cement-treated and SV = the mean of the slope variance in
some bituminous-treated. The gravel used for the two wheelpaths;
cement and bituminous treatment was essen-
tially the subbase material as used in the flex- C + P = a measure of cracking and patch-
ible pavement experiments. The gravel base ing in the pavement surface; and
material was slightly coarser than the subbase = a measure of rutting in the wheel-
and contained some plastic soil fines. It was paths.
obtained from a local source.
(In this equation and throughout this report,'
The asphaltic concrete binder course was a logarithms are to the base 10.)
mixture of dense-graded crushed dolomitic lime-
stone aggregate 1-in, maximum size and natural Slope variance was discussed in Section 1.3.
sand with about 4.5 percent of 85-100 penetra- Cracking, C, in Eq. 11 is defined as the area, in
tion grade paving asphalt. The asphaltic con- square feet per 1,000 sq ft of pavement surface,
crete surface course (the uppermost layer) was exhibiting class 2 or class 3 cracking. Class 2
a mixture of dense-graded crushed dolomitic cracking is defined as that which has progressed
limestone aggregate, 3/4 -in. maximum size, and to the stage where cracks have connected to-
natural sand with about 5.4 percent of 85-100 gether to form a grid-type pattern. Class 3
penetration grade paving asphalt (Table 2). cracking is that in which the bituminous sur-
Details relating to the materials, the construc- facing segments have become loose. Patching,
tion testing and construction operation are P, is the repair of the pavement surface by skin
given in Report 2. patching or deep patching expressed in square
feet per 1,000 sq ft of pavement surfacing. Rut
2.2 PAVEMENT PERFORMANCE depth, RD, is defined as the mean depth of rut
in both wheelpaths of the pavement where the
The serviceability concept and the derivation rut is the depression under the center of a 4-ft
of the serviceability indexes are described straightedge. The mean rut depth was esti-
briefly in Section 1.3 of this report and in detail mated by sampling in each wheelpath at 25-ft
in Appendix F. intervals. The relative significance of these
This section describes the flexible pavement terms is discussed in Appendix F.
present serviceability index. As required by the During critical periods it was often apparent
first objective of the Road Test, it gives the that a section had reached a present service-
principal relationships showing flexible pave- ability level of 1.5 or that it would reach this
ment performance as a function of design and level prior to the next regularly scheduled in-
load variables. Italso presents the results ob- dex day. Where this was the case it was some-
tained from the paved shoulder and the special times necessary to determine the serviceability
base experiments (Designs 2 and 4). A sum- index by expedient methods. Cracking and
mary of the material contained in each major patching were measured in the usual way, but
subsection precedes the text of the subsection. since it was not feasible to obtain slope vari-
ance with the project profilometer, a special
2.2.1 Serviceability Index for Flexible present serviceability index equation was devel-
Pavement oped that included a rut depth variance term
This subsection contains the equation used to instead of slope variance. The variance in rut
determine the present serviceability index of depth was determined by measuring the depth
each flexible pavement test section (Eq. 11). of rutting at 5-ft intervals in both wheelpaths.
It also includes tables giving for each design It is emphasized that this expedient was used
the number of axle load applications sustained only in cases where sections were nearing fail-
ure and it appeared that major maintenance
before the section's serviceability fell to 1.5 for would be required before the next regular 2-
unweighted applications (Table 5), to 1.5 for week index day..
weighted applications (Table 6), to 2.5 for un- DS 7322 gives the complete serviceability
weighted applications (Table 7), and to 2.5 for history of each section as well as cracking,
weighted applications (Table 8). patching, rut depth and slope variance data by
Eq. 11 was used to determine the level of wheelpaths. Section history charts showing the
serviceability of the surviving flexible pavement trends of cracking, patching, serviceability,
sections every two weeks during the period of roughness index, deflection and other items of
traffic operation. information are also available for every test
-- Inner
i'jW.1Ij:..
--------------------------.
• __
•• __
__
-• • ___
__
__
ni..a•uuuuiu•ua•u•uu•..._
nu••••uuuu•••uu••au•uu•uu w
'-3
MEM IMM
-
---------------------------__
----------------------------__
- _
-
I
--------------------------_
--------------------------_
--------------------------_
-------------------------
_
74 --------------------------
-
-
---
- ••.
------------------------_
•r
- -------,
_________ ------------------
_________
---- __
Miifl.1LJ I
• -------
_____•\•• '----------------___
IL\
J
--------------------------
--------------------------
----------------------------
-- -------------------------
- .•
--------------------------
-------------------------
section. Figures 19, 20 and 21 are examples of 2.2.2.1 Main Factorial Experiments (Design
these charts as they may be found for each 1).—This subsection contains the results of the
section in DS 4199. major Road Test flexible pavement analysis, the
Basic data relative to the performance of the pavement performance analysis, and develops
factorial sections for both weighted and un- the relationships for flexible pavement sought
weighted application are given in Appendix A.
Data for a present serviceability level of 1.5 in the first objective. These relationshis have
and 2.5, are also given in Tables 5, 6, 7 and 8. been reduced to four equations containing terms
Load applications for each design of pavement for-the variables included in the test. Eqs. 13,
are given for those sections that were removed 17, 18, and 19 are for the case where load appli-
from the test and p values for those sections cations have been adjusted by the seasonal
that survived the test. weighting function; similar equations are given
for unweighted applications.
2.2.2 Performance as a Function of Design and
Load Graphs and tables were constructed from the
equations for use in the study of performance
This subsection gives relationships between over the wide range of designs and loads in-
flexible pavement performance and variables cluded in the Road Test.
that describe load and pavement design. Per-
formance data, models, and analytical proce- A convenient presentation of the relation-
dures described in Section 1.3 are used to obtain ships for the axle loadings of the Road Test is
specific performance-design-load equations for shown in Figure 22. For example, to deter-
the factorial experiments. This section also in- mine what pavement structure would have sur-
cludes associations of performance with design vived a million 22.4-kip single axle loads at the
and load variables for the paved shoulder Road Test before its serviceability level dropped
studies and for the special base type studies. to 2.5, the chart is entered at 1,000,000 applica-
1
I
THICKNESS INDEX'O.44D1+ 0.1402+0.1103
SURFACING THICKNESS, INCHES (2 IN. MINIMUM)
01
, ,/-
- i--
_- -iJ_L
1
0000
ME
t-j --
r
EXTRAPOLATION
ni HHHI
10 100 1,000
WEIGHTED AXLE LOAD APPLICATIONS IN THOUSANDS
Figure 22. Main factorial experiment, relationship between design and axle application
at p = 2.5 (from Road Test equations).
TABLE 5 00
PERFORMANCE DATA, EXPERIMENT DESIGN 1, UNWEIGHTED AXLE APPLICATIONS TO
p = 1.5 OR p AT END OF TRAFFIC TEST'
Axle Subbase
Load Thick. Unweighted Axle Applications (x 10)
(kips) (in.)
(a) Loon 2
(b) LooP3
12S 0 64 65 77 77 72 87 78 80 178
0 109
4 4. 73 90 87 87 589 78 109 (2.3)
4 87
8 73 88 722 100 561 (1.6) 109 611 (3.6)
8 77
24T 0 64 70 72 72 75 80 72 88 175
0 95
4 3 . 76 81 80 86 102 80 100 627
4 82
8 74 80 555 91 111 614 88 558 (3.3)
74 . .
(c) LooP 4
(d) Loop 5
(e)Loop6
Axle Subbase
Load Thick. Weighted Axle Applications (x 10)
(kips) (in.)
(a) Loon 2
(b) Loop 3
(c) LoOP 4
(d) Loop 5
Axle Subbase
Load Thick. Unweighted Axle Applications (x 101)
(kips) (in.)
(a) Loor 2
4 628 (2.6)
(b) Loon 3
(c) LooP 4
(d) LooP 5
22.4S 4 64 71 80 67 80 83 77 98 608
4 485
8 70 73 79 80 97 169 101 561 (2.7)
8 91
12 86 78 (2.9) 90 506 1099 611 701 (3.5)
12 74 Id
40T 4 2 72 80 62 81 99 87 102 242
4 462 tj
8 72 76 85 80 93 123 107 373 800
8 151
12 85 459 153 98 182 (3.0) 611 764 (3.2)
12 80
w
(e). Loop 6
0
4-In. Surface 5-In. Surface 6-In. Surface
1 CIO
Values in parentheses are values of p.
TABLE 8
PERFORMANCE DATA, EXPERIMENT DESIGN 1, WEIGHTED AXLE APPLICATIONS TO
p 2.5 OR p AT END OF TRAFFIC TEST'
Axle Subbase
Load Thick. Weighted Axle Applications (x 10')
(kips) (in.)
(a) Loop 2
1-In. Surface 2-In. Surface . 3-In. Surface
(d) LooP 5
(e) Loon 6
CO
1
Values in parentheses are values of p. - • Cn
36 THE AASHO ROAD TEST, REPORT 5
tions on the abscissa and the thickness index D (base thickness, in.), and D3 (subbase thick-
(4.5) is read on the ordinate scale. Asphaltic ness, in.), and of the load variables, L1 (nomi-
concrete surfacing, base and subbase may be nal axle load, kips*) and L2 (1 for single axles
combined in any combination for an index of or 2 for tamdem axles).
4.5, provided it meets the conditions for use of The function /3 determines the general shape
the thickness index equation stated on the of the serviceability trend with increasing axle
chart. Many combinations of structural layers load applications, W. If /3 = 1, the trend is
will meet these conditions. One, for example, a straight line; if /3 > 1, the serviceability
is 4 in. of surfacing, 10 in. of base and 12 in. of loss rate increases with applications; and if
subbase. /3 < 1, the loss rate decreases with axle load
Since these equations represent serviceability repetitions. Graphs of the performance data
trend data observed in the test, some Road Test for flexible pavements in Appendix A indicated
sections failed sooner and some later than indi- that designs failing early in the Road Test
cated by the smooth curves. Thus, some allow- tended to have an increasing rate of service-
ance should be made for the scatter of the data ability loss (/3 > 1), while more adequate de-
as shown, for example, in Figure 25. Through signs as a rule had a decreasing loss rate
a residual analysis it was found that the scatter (/3 < 1). Estimates of /3 were obtained from
corresponds to approximately ±14 percent of the performance data of a number of sections
the thickness index values given by the curves. that experienced relatively little serviceability
If comparisons are made with observed per- loss in the Road Test. The average of these
formance of actual highways in service, addi- values was approximately 0.4, and this value
tional allowance should be made to account for was assigned to /30, the assumed minimum value
differences between the Road Test and the ac- for /3 in Eq. 6.
tual highway in materials, environment, and The function p is equal to the number of load
loading history. applications at which p = 1.5, and is assumed
These relationships are not intended to be to increase as design increases and to decrease
design equations. However, they can serve as a as load increases. The over-all aim of the per-
basis for design procedures in which variables formance analysis is to arrive at formulas for
not included in the Road Test, such as soil type, /3 and, p in terms of D,, D4, D:i, L1 and L4 so that
are considered. Eq. 12 may be used to predict the value of p
Tables and discussion are included to show after a specified number of applications, W. Or
the basis for determining the significance or if Eq. 12 is solved for log W,
nonsignificance of the various effects. Correla-
tion indexes show the degree of correlation 42
found in the relationships; mean residuals, the
degree of scatter of the observed performance log W = log p + log ( 2.7 (13)
data from the predictions of the performance FE
equations.
The thickness index found to apply to Road then Eq. 13 may be used to predict the number
Test flexible pavements is of interest in itself. of applications required to reduce p to a speci-
For the weighted applications case the thick- fied value.
ness index equation (Eq. 19) indicates that an For the flexible pavements, /3 and p are given
inch of surfacing was about three times as by particular cases of Eqs. 6 and 7 of Section
effective as an inch of base and four times as 1.3.5, as follows:
effective as an inch of subbase in improving
pavement performance within the range of de- BO (L1 + L2)B,
sign studied. /3=0.4+ (14)
The use of the seasonal weighting function (D + 1)B,L2133
on axle load applications was found to increase A0 (D ..+ 1) A, L2A,
the correlation index from 0.48 to 0.70 and to (15)
reduce the mean residuals by 15 percent. = (L1 + L2)A2
The general model used to represent pave- in which D is a thickness index given by
ment performance was Eq. 4. For flexible pave-
ment test sections in the factorial experiments D = a1 D1 + a2D2 + aD3 (16)
the average initial serviceability trend value
was c0 = 4.2, and since c1 = 1.5, c0 - = 2.7,
and the trend curves are represented by If the coefficients a, a2 and a3 in Eq. 16 are
each assigned a value of one, D is the total
structure thickness. In the Road Test analyses,
p = 4.2 - 2.7 (-!!- (12)
1
* For example, for single axle loads of 18 or 22.4
Both /3 and p are positive functions of the de- kips, L1 = 18 or 22.4; for tandem axle loads of 32 or
sign variables, D1 (surfacing thickness, in.), 40 kips, L1 = 32 or 40.
FLEXIBLE PAVEMENT RESEARCH 37
however, these coefficients were permitted to squares that can be used to determine the rela-
vary so that the three elements of the pave- tive significance of various effects. Because
ment structure might each enter into the thick- there were complete factorial experiments with
ness index with a different weight per unit replication in each ioop, analysis of variance
thickness. could be used to separate out and determine
Several analyses of variance were made in mean squares for the separate linear effects
order to infer how D1, D 2 and D3 might be of D1, D 2 and D3 ; mean squares for the sepa-
brought into the expressions for /3 and p. Par- rate and combined non-linear effects of .D1, D 2
tial results from one such analysis are given and D3 ; mean squares for the separate and
in Tables 9 and 10, which show first the num- combined interaction effects of D1, D 2 and D3 ;
ber of test sections that entered the analysis and mean squares for unexplained effects rep-
for each ioop. The second part gives mean resented by replicate differences. Por each of
TABLE 9
ANALYSIS OF VARIANCE FOR LOG p ESTIMATES' WITHIN Loops,
UNWEIGHTED APPLICATIONS
Effects2:
Lane mean difference 13.75 0.28 0.14 0.02 1.81
D. (surface) linear:
Lanes combined 14.47 4.10
Lane interaction 0.02 0.00 0.00 0.01 0.08
D2 (base) linear:
Lanes combined 11.18 6.75 .i
Lane interaction 0.27 0.07 0.01 0.01 0.01
D2 (subbase) linear:
Lanes combined 0.71 4.60 7.76 6.38 7.97
Lane interaction 0.12 0.07 0.02 0.02 0.04
D1, D2, D3 non-linear:
Lanes combined 0.87 0.21 0.01 0.11 0.07
Lane interaction 0.42 0.02 0.00 0.01 0.15
D1, D2,D3 interactions:
Lanes combined . 0.07 0.20 0.12 0.07 0.13
Lane interaction 0.07 0.01 0.01 0.02 ' 0.02
Replicate differences:,
Lanes combined 0.28 0.02 0.01 0.01 0.06
Lane interaction 0.14 0.01 0.01 0.01 0.06
Within lane 1 2 1 2 1 2 1 2 1 2
Coefficient
for log (D+1) ' 8.12 9.07 6.23 5.38 8.83 9.07 9.14 9.50 9.87 11.57
Percent of variation
explained by regression 68 88 66 69 83 89 84 86 82 75
Mean square for
unexplained variation 0.32 0.12 0.16 0.11 0.08 0.05 0.07 0.06 0.06 0.12
1Data from which this table arose are the estimates log as described in Appendix G.
'Mean squares for effects; underlined values considered to be significant relative to replicate differences pooled
with interaction effects.
38 THE AASHO R0 D TEST, REPORT 5
these effects, Tables 9 and 10 show mean significant at the 5 percent level. Although a
squares for the two lanes combined in any loop few scattered interactions did reach this level
and also mean squares for lane interactions of significance, the general finding was that
that bring about dissimilar effects in the two D,, D2 and D, interactions were similar to
lanes of any loop. In any column of the tables, replicate effects and could be pooled with the
significance is attained for a stated effect if replicate effects. Moreover, mean squares for
the mean square for the effect is a sufficient non-linear effects of D,, D 2 and D3 on log p
multiple of the mean square for an unexplained estimates were generally of the same size as
or residual effect. For example, D1 , D 2 and D3 the interaction effects, and since the linear
interaction effects must have mean squares effects of these variables were highly significant
about eight 'times the mean square for repli- in almost every case, the linear expression
cate effects if the interaction effects are to be a,D1 + a 2D2 + a,D3 accounts for practically all
TABLE 10
ANALYSIS OF VARIANCE FOR LOG p ESTIMATES' WITHIN Loops,
WEIGHTED APPLICATIONS
Effects2 :
Lane mean difference 13.25 0.32 0.14 0.04 1.55
D, (surface) linear:
Lanes combined 16.58 689 6.94 7.87 3.83
Lane interaction 0.00 . 0.00 0.00 0.00 0.01
D, (base) linear:
Lanes 'combined 11.04 7.78 6.16 6.11 4.04
Lane interaction 0.14 , 0.07 0.03 0.01 0.00
Da (subbase) linear:
Lanes combined 0.62 6.94 . 7.51 7.20 7.07
Lane interaction 0.08 0.08 0.01 0.01 0.00
D,, D2, D3 non-linear:
Lanes combined 0.90 0.13 0.05 0.09 0.04
Lane interaction 0.45 0.04 0.01 0.01 0.09
D,,-D,, D3 interactions: .
Lanes combined 0.10 0.07 . 0.09 0.03 0.11
Lane interaction 0.03 0.01 0.01 0.02 '0.01
Replicate differences:
Lanes combined 0.27 0.01 0.01 0.02 0.05
Lane interaction 0.13 0.01 0.01 0.01 0.05
Within lane 1 2 1 2 1 2 1 2 1 2
Coefficient
for log (D+1) 8.39 9.07 7.47 6.52 9.27 9.10 10.30 10.14 10.09 10.41
Percent of variation
explained by" regression 71 88 81 84 87 93 91 93 85 77
Mean square for
unexplained variation 0.32 0.13 0.11 0.07 0.06 0.03 0.04 0.03 0.05 0.09
Data from which this table arose are the estimates log P^ as described in Appendix G.
'Mean squares for effects; underlined values considered to be significant relative to replicate differences pooled
with interaction effects.
FLEXIBLE PAVEMENT RESEARCH 39
___ •7
EXTRAPOLATION
10 100 1,000
WEIGHTED AXLE LOAD APPLICATIONS IN THOUSANDS
Figure 23. Main factorial experiment, relationship between design and axle load
applications at p = 1.5 (from Road Test equations).
If W represents weighted applications ob- Figure 24. Main factorial experiment, relationship
tained throuh the use of seasonal weighting between design and load at p = 2.5.
40 THE AASHO ROAD TEST, REPORT 5
function described in Section 2.2.2.1.1, then the 10' (D + 1) 8 .94 L24 '7
analysis gives the following .equations: P= (21)
(L1 + L 2 ) 45
0.081(L1 +L2 ) 323
p = 0.4 + (17) D = 0.37D1 + 0.14D2 + 0.10D3 (22)
(D + 1) 519 L2323
Thus for a particular pavement design and
10593 (D + 1)936L 4.33
= (18) axle load, either Eqs. 17, 18 and 19 or Eqs. 20,
(L1 + L,) 21 and 22 give values for /3 and p that may be
D = 0.44D1 + 0.14D2 + 0.11D3 (19) substituted in Eq. 12 if p is to be estimated
from W, or in Eq. 13 if W is to be estimated
If the applications are unweighted, then the whenp is given. Figures 22 and 23 show how
performance equations are as follows: W varies with D in Eq. 13 when p is fixed at
2.5 and 1.5, respectively. Each figure has ten
0.083(L1 + L2)487 curves, one curve for each test load used in
p = 0.4 + ( 20)
(D + 1) 873 L2487 the Road Test.
I I 111111
I 2 Kip Axle Load
U liii
C1 liii ______________________
____
to 100 1POO 0
I I I I
l8Kip Axle
I - —
I 10 100 100
IM1111111lIllIjIlI'
111111110iiiiiu.ii1iirn
I1111111110 1IIIIIIII
II1I!1I IW
L-4-
IIlIIIIllhIIIIIIIIIIIuIIIIII
WEIGHTED AXLE LOAD APPLICATIONS AT p. 2.5, 1000.
IJJIJU
IuIIIIIIIIIIIIIIIIIIIuIIiOI
I to 100
WEIGHTED AXLE LOAD APPLICATIONS Al p. 2.5, 1000.
000
Figure 25. Main factorial experiment, relationship between design and single axle load
applications at p = 2.5.
FLEXIBLE PAVEMENT RESEARCH 41
I 11111111
24KiPIAxte Load ipAX1B Load
-
100 IPOO
0.0
.IIi'** HillIIIIIIIIIHI WWI
Figure 26. Main factorial experiment, relationship between design and tandem axle
load applications at p = 2.5.
Figure 24 shows design requirements when tions do not appear in the graphs, their per-
the final serviceability value is p = 2.5 for a formance data were used in the development
range of single and tandem axle loads at three of the performance equations.
levels of load applications. In this and the re- The performance data in Appendix A, De-
maining graphs for flexible pavement perform- sign 1, give a minimum of 5 and a maximum
ance (Figs. 24, 25 and 26), the final service- of 10 (p, log W) pairs for each test section.
ability level is p = 2.5. The choice of 2.5 for When p is fixed at 3.5, 3.0, 2.5, 2.0 and 1.5
final serviceability was arbitrary. The level of there can be as many as 5 log W observations,
serviceability at which states actually perform and when log W is fixed at t = 11, 22, 33, 44 and
major maintenance will be established by a 55 index days there can be as many as 5 ob-
survey of pavements scheduled for overlay or served values for p. Corresponding to each
reconstruction. observation, log W or p, is a calculated value,
Figures 25 and 26 show the correspondence log"W or P , obtained from the performance
between the individual curves of Figure 22 and equations. Differences between calculated and
performance data from Appendix A for each observed values are the residuals A log W =
of the ten traffic lanes. Each point represents log"W - log W and A p = - p. Absolute
the observed number of weighted applications values of these residuals are summarized in
at which the serviceability of a test section was the first part of Table 11 which shows for each
2.5. Horizontal deviations of the points from lane the number of residuals of each type as
the curves represent prediction errors or resid- well as mean absolute residuals. Mean absolute
uals when Eqs. 13, 14, 15, and 16 are used to values for A log W in Loop2, lane 1 were found
predict the life of a section (to p = 2.5) whose to be extreme relative to the other lanes and
design and load values are specified. were omitted from the grand means. Table 11
Points shown (Figs. 25 and 26) represent thus shows that mean values for A p and A log
only those sections whose serviceability fell to W were 0.53 and 0.27 for unweighted applica-
2.5 by the end of the test. All remaining sec- tions, and 0.46 and 0.23 for weighted applica-
tions would be represented by points whose tions.
abscissas are to the right of 1,114,000 applica- Log W residuals are horizontal deviations
tions. The number of such sections for any from the performance equation curves and are
lane can be found by subtracting the number thus of special interest in the use of these
of points shown from 22 in Loop 2 and from curves. The second part of Table 11 shows a
30 in all remaining ioops. Although these sec- further summary of log W residuals. The cor-
TABLE 11
SUMMARY OF PERFORMANCE EQUATION RESIDUALS AND REPLICATE DIFFERENCES
Load
p Residuals . Log W Residuals
Number of Mean Absolute Mean Absolute
Loop Lane Sections Number Number of
L, L2 of Residuals Wtd. App. Residuals Unwtd. App. Wtd. App.
Unwtd. App.
2 . 1' 2 1 22 97
55
0.39
0.44
0.41
0.51
42
83
0.85
0.40
0.92
0.36
2 6 1 22
3 1 12 1 30 37 0.53 0.46 138 0.34 0.28
2 24 2 30 31 0.57 0.44 146 0.37 0.29
4 1 18 1 30 62 0.58 0.53 136 0.23 0.19
2 32 2 30 66 0.44 0.41 138 0.18 0.16 CD
relation indexes are given as 0.48 and 0.70 it has been supposed that a satisfactory model
for unweighted and weighted applications, and fit is indicated whenever mean absolute
respectively, and corresponding root mean residuals are about equal to replicate mean
square residuals are 0.36 and 0.31. differences. Table 11 gives this comparison
The general nature of the over-all A log W for unweighted applications to be 0.53 vs 0.46
distribution (except for Loop 2, lane 1) is in- for p and 0.27 vs 0.15 for log W. For weighted
dicated by the fact that about 60 percent of all applications the comparison is 0.46 vs 0.46 for
A log W is contained within one mean absolute p and 0.23 vs 0.17 for log W. It is quite pos-
residual and about 90 percent within two mean sible that other models and fitting procedures
absolute residuals. The distributions support may do equally well, and that some will repre-
the statement that in about nine out of ten sent better the long-time performance of high-
cases, observations agree with corresponding ways in actual service.
performance equation estimates to within plus 2.2.2.1.1 Seasonal Weighting Function.—
or minus two mean residuals. In other words The concept of a seasonal weighting function
there is approximately 90 percent confidence to allow for changing load effects in a changing
(Table 11 includes root mean square resid- environment was discussed in Section 1.3.4.
uals, twice whose value can be used to set limits The weighting function, q,, used in flexible
with approximately 95 percent confidence.) pavement analyses is given by
that log W will be observed between logAW ±
0.5* for unweighted applications and between = [2c1( - dti]2 (23)
log W ± 0.46 for weighted applications. In
terms of the thickness index D, thesbands
correspond approximately to D ± 0.14 ID + 1) in which d is an estimate of the average deflec-
forunweighted applications and to D ± 0.11 tion under a 6-kip wheel load of eight sections
(D + 1) for weighted applications, where D is in Loop 1 (the non-traffic loop) during index
obtained by entering the appropriate perform- period t. Deflections were generally taken twice
ance equation (or curve) with fixed W and cal- during each index period and averaged, then a
culating D. For relatively heavy designs, the 3-point moving average was used to smooth
uncertainty represented by two mean residuals the deflection history of the eight sections. The
in log W is approximately 0.181) using the un- deflection d1 is the smoothed deflection for
weighted applications formulas and approxi- index period 1— 1.
mately 0.1412 using the weighted application Division by d, the 2-yr average of d, makes
formulas. All confidence limits such as these q t, a unitless factor and also makes the weight-
are relative to the Road Test conditions and ing function relative to the Road Test condi-
range of variables. tions. Whenever d = d1 = d, then q t = 1, so
The last part of Table 11 summarizes log W that the weighting function is unity if deflec-
and p differences observed between replicate tions in Loop 1 are unchanging and are at the
test sections. In all there were 32 pairs of 2-yr average value.
replicate sections in Design 1, and the mean The exponent 2 in Eq. 23 has been assumed
replicate difference is 0.46 for p, 0.15 for un- as an appropriate factor for increasing the
weighted log W differences and 0.17 for amplitude of q in periods of high increasing
weighted log W differences. In those pairs deflection relative to periods of low constant
where one section was out of test before the deflection. Data and values for q are given in
second, replicate differences were provided at Appendix B.
the missing points by assuming that the re-
maining differences would be as large as when In Table 11, the use of the seasonal weight-
the first section went out of test. ing function increased the correlation index
For whatever reasons two replicate sections from 0.48 to 0.70 and reduced the mean resid-
do not show the same performance, it can be uals in log W from 0.27 to 0.23.
expected that the performance data will deviate 2.2.2.2 Paved Shoulder Studies.—A study of
from any fitted equation. For a particular lane the effectiveness of paved shoulders (Design 2)
a satisfactory model and fitting procedure was included in the Road Test. A total of 48
should result in residuals that average about test sections was provided in this study.
the same as deviations of replicate observations Unfortunately, the pavements selected for
from their own mean. For two replicates, then, the tests were underdesigned to the extent
estimation errors should average to be about that 42 of the sections failed during the first
one-half the replicate differences if the fit is to spring of traffic operation and little information
be judged adequate. Since the performance of value was disclosed by the experiment. An
equations were developed across lanes and loops attempt was made to obtain additional informa-
it is expected that the average residual will be tion by studying the differences in performance
more than one-half the average replicate dif- of the outer and inner wheelpaths of the test
ference, but how much greatercannot be deter- sections of the main experiment.
mined in the absence of replicate lanes and The results of these studies pointed to the
loops. In the Road Test performance analyses fact that the pavement needed to maintain a
44 THE AASHO ROAD TEST, REPORT 5
certain serviceability at a given number of axle tion's serviceability had dropped to 2.5. Also
load applications would be considerably thinner shown are similar values for the 100-ft fac-
in the inner than in the outer wheelpath. torial sections with equivalent structural de-
2.2.2.2.1 Design 2.—The paved shoulder signs.
study (Design 2) involved six structural sec- Of the 48 160-ft paved shoulder sections, 42
tions (12 test sections) in each flexible pave- were removed from the test during the first
ment tangent of Loops 3, 4, 5 and 6. The sec- spring of traffic operation when the service-
tions were 160 ft. long with shoulders paved ability of their subsections had been reduced
with 3 in. of asphaltic concrete varying in to 1.5. The exceptions were the 2-3-8 design
width from 8 ft to 0 in the direction of traffic. (thickness, in inches, of surfacing, base and
The six structural sections in each tangent in- subbase, respectively) in lane 1 of Loop 3 and
cluded three different pavement designs and the 4-3-16 design in both lanes of Loop 6 These
three replicate sections. were the thickest structures included in the
The objective of the study was to determine paved shoulder test in these loops. Only three
the effect of width of paved shoulder on the subsections remained in test for the entire
performance of the pavement. traffic period. They were the first three sub-
Since the procedures used on the Road Test sections in the 4-3-16 structure in lane 2 of
did not produce objective measures of the serv- Loop 6.
iceability of pavements on a foot-by-foot basis, As shown by Figures 27 and 28, there was
each paved shoulder section was divided into no clear evidence of any effect of width of
four 40-ft subsections. In the analysis of per- paved shoulder on the performance of the sec-
formance each subsection was considered to tion that failed during the first spring. In the
have a paved shoulder whose width was repre- other three sets of replicate sections, there was
sented by its average width (that is, 7, 5, 3 some indication of an effect, but it was neither
and 1 ft in the direction of traffic). orderly nor consistent.
Data concerning the performance of the Of the 24 sets of replicate sections, 17 in-
paved shoulder subsections are shown graph- dicated no effect of paved shoulders on per-
ically in Figures 27 and 28. The values are formance. The average performance of these
averages of the total.number of the axle loads sections agreed closely with the performance
applied to corresponding subsections in each of the factorial sections having the same struc-
pair of replicate sections at the time the subsec- tural design. For the remaining 7 sets, the
400 400
U,
Loop. 3 12 Kip Loop4 - SKip
0
9 (AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS) (AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS)
300 300
U,
'U
0, jOOEOIGj
2
2 0
ISO ISO
,
-J
0
a
C
0
A.
400 400
00
Loop 5 -..- 22.4 Kip 0
Loop 6 - 30 Kip
2
(AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS) (AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS)
2 300 300
'0
CU
200 200
(4-3-16
'0
DESIGN)
OJ (6-3-8 DESIGN)
(5-3-4 DESIGN) 2 —
2 too (3-9-4 DESIGN) (3-3-4 DESIGN) - [ jjjDESIGNfl
R 100
:
C
•
-
llJllh
-
0.
A.
C -
Figure 27, Paved shoulder experiment, performance data for single axle loads,
unweighted applications.
FLEXIBLE PAVEMENT RESEARCH 45
400 400
00
S Loop 3 — 24 Kip Loop 4 32 Kip
(AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS) - (AVERAGES FOR EACH PAIR OF REPLICATE SECTIONS)
300 - 300
- 200 - 200
100
(4-3-0 DESIGN)
(2-3-8 DESIGN)
O •
11111 A.
600 600.. 5
0
0
80 (4-3-16 DESIGN)
400 400
300 300
200 2 200
)6-3-8
(8-3-4 DESIGN)
(3-9-4 DESIGN) (4-3-8 DESIGN)
100 liii R
11111
(3-3-4 DESIGN)
performance of the paved shoulder subsections, designs included in the test. The pavements
on an average, was appreciably better in five were underdesigned to the extent that all four
cases than it was for the counterpart factorial subsections in 42 of the 160-ft test sections
sections, and not. as good in two cases. How- went out-of-test in the first spring of traffic
ever, there was no clear indication of the effect operation, with pavement distress and failure
of varying width. occurring quite rapidly. However, the six test
Additional study of the condition of the pave- sections that survived the test beyond the first
ment in each paved shoulder section just prior spring were benefited to some extent by the
to maintenance work on the section indicated presence of the paved shoulder although they
that subsection serviceability decreased with provided little indication of the effect of its
increase in paved shoulder width in about as width on their performance. These sections
many cases as it increased. This also applied represented the thickest pavement design in-
to the depth of rutting. There were as many cluded in the paved shoulder test on Loops 3
sections showing an increase in depth of rut- and 6. If thicker designs had been used, the
ting in the outer wheelpath with an increase effect of the width of paved shoulders on pave-
in width of paved shoulder as there were show- ment performance might have been more
ing a decrease. evident.
Records of deflections in the outer wheelpath 2.2.2.2.2 Performance, by Wheelpaths.—
of the subsections also failed to reveal any Because all biweekly measurements that en-
potential benefit from the presence of the paved tered into the serviceability index for each
shoulders. flexible pavement section were made by wheel-
Failure of the experiment to provide con- paths, it was possible to make a complete
clusive evidence of the effect of width of paved analysis of the perfornance of each wheelpath.
shoulders on pavement performance can be at- The analyses were identical to those described
tributed at least in part to the selection of the in Section 2.2.2.1 except that the slope van-
46 THE AASHO ROAD TEST, REPORT 5
ance, cracking and patching, and rut depth thickness index differential could be made up of
observed in the inner wheelpath were used in asphaltic concrete, crushed stone base, or sub-
computing the inner wheelpath serviceability base, or by any combination of the three mate-
index for each 2-week period and those ob- rials. This is of interest because it indicates
served in the outer wheelpath were used in that for a given quantity of material in the
computing indexes for the outer wheelpath. pavement structure, the expected life of the
To simplify this analysis it was assumed that pavement would be greater if the inner wheel-
the thickness index for each wheelpath would path were constructed thinner than the outer
be the same as that determined for the entire wheelpath than it would be if both wheelpaths
section (for weighted applications) as in Eq. had identical structural thicknesses. Or, a
19. The other coefficients and constants in the structural design for a specified life would re-
performance equations for each wheelpath were quire less material if the outer wheelpath were
determined for weighted applications by the thicker and the inner wheelpath thinner than
same techniques as described in Section 2.2.2.1 it would be if a conventional uniform thickness
and in Appendix G. The resulting equations design were used. These comments relate to
for p and log p for the inner wheelpaths are quantities of materials only; the cost of build-
ing unconventional sections may outweigh the
0.028(L1 + L2) 26° savings in materials.
p = 0.4 + (24)
(D + 1)362 L226° 2.2.2.3 Special Base Type Experiments.—An
important investigation within the flexible
log p = 6.63 + 9.20 log (D + 1) - 5.02 log pavement experiment involved the study of the
(L1 + L2 ) + 4.47 log L2 (25) relative effectiveness of certain treated and
untreated bases. Four base types were studied:
The resulting equations for the outer wheel- crushed stone, gravel, cement-treated and
paths are bituminous-treated gravel. There were 48 test
0.087(L1 + L 2) 272 sections in this study.
p = 0.4 + (D + (26) The base experiment was designed so that
L2272
no mathematical analysis of the performance
logp = 5.82 + 8.72 log (D + 1) —4.47 log of the sections was attempted. The analysis
(L1 + L2) + 4.01 log L2 (27) was essentially graphical. However, it is antici-
pated that the Highway Research Board and
The equations may be compared with those others will later incorporate the special base
(Eqs. 17 and 18) obtained for the entire sec- data into the data from the main factorial ex-
tion. All terms in the equations are defined as periment in an effort to produce performance
they were for the previous equations. There is equations containing terms for the special base
perhaps more reliability associated with the materials.
outer wheelpath performance equations than The results of the analysis are presented in
with those for the inner wheelpath since the graphs (See Figs. 35 and 36), which permit
serviceability loss in the inner wheelpath was comparison of the performance of the stone,
generally less than for the outer wheelpath. cement-treated and bituminous-treated bases;
As a result the serviceability trends of inner that is, comparison of the thickness of the
wheelpaths for many sections had not been well materials that was necessary to maintain a
established at the time the section was removed level of serviceability of 2.5 at a specified num-
from test. ber of load applications. For example, for the
Figure 29 shows plots of W against D for 18-kip single axle load at 1,000,000 applications,
each axle load of the Road Test. Each graph the required thickness of base (where the sur-
includes a plot from the outer wheelpath equa- facing thickness was 3 in. and the subbase 4
tion, one from the inner wheelpath equation, in.) is shown to be approximately 13, 8 and 6
and one from the over-all section equation (see in. of stone, cement-treated and bituminous-
Section 2.2.2.1). The over-all section equation treated base, respectively. These values suggest
gives very nearly the same relationship as does that there was considerable difference in the
the analysis of outer wheelpath alone. This in- performance of the treated bases and the
dicates that since most distress in flexible pave- crushed stone bases. In fact, in all loops and at
ment occurs first in the outer wheelpath, all levels of serviceability the performance of
pavement structure designs based on the re- the treated gravel bases was definitely superior
quirements for the outer wheelpath alone may to that of the untreated crushed stone.
be nearly the same as the structure design Most of the sections containing the untreated
needed for an entire pavement. gravel base failed early in the test (Figs. 30
These plots show that the requirements for and 31); their performance was definitely in-
pavement structure in terms of thickness index ferior to that of the sections with crushed stone
necessary to maintain a serviceability of 2.5 base.
after one million applications of axle load The performance of four different types of
averaged considerably greater for the outer base course was studied in Design 4: crushed
wheelpath than for the inner wheelpath. This stone, gravel, cement-treated and bituminous-
FLEXIBLE PAVEMENT RESEARCH 47
II
z
'C
U
I
I-
6 6
I 11111111 I 11111111 I IIIIIIII I IIIl I
24 Kip Tandem Axle Load 18 Kip Single Axle Load1
iIii1 32 Kip Tandem Axle Load
j%JiI IA41
I 1111 11 5 I/flfl 5
I,.JilIII
4 4
11/ , lI
111111
— 11111111
II II
Ik 3 3
0
tO 100 1000 14000 I tOO 1000
H
14000 I 0 100 000 IQOoO
13 6 0
1 11111111 I 1111111' I IIIIIIII II III,iII
I I ,,ItII
22.4 Kip Single Axle Load 40 Kip Tandem Axle Load II I 111111
5 ,_l_uJl I 5
I II VII
I /.IIII
I I''
III
/ II11111
I 11111
4 4
hi
a
z
3 3
z
'C
U
I
2 2
IIIIIIIO
1 to too 1000 Iq000 I 10 too t000 10000 I tO tOO 1000 ILPJUU
WEIGHTED AXLF LOAD APPLICATIONS AT ,.2.5. 10005 WEIGHTED AXLE LOAD APPLICATIONS AT p2.5. I0OS
6
5 ---OUTER WHEELPATH
---INNER WHEELPATH
OVERALL SECTION
a AREA OF EXTRAPOLATION
4
a
z
z
'C
U
I
I- S
1111111111
to too 000 4000
WEIGHTED AXLE LOAD APPLICATIONS AT p.2.5. 1000$
Figure 29. Main factorial experiment, relationship between design and axle load
applications at p = 2.5.
THE AASHO ROAD TEST, REPORT 5
treated material. The crushed stone base was relate thickness of the four types of base course
the same as that used in the factorial sections; material to pavement performance and pave-
the gravel base consisted of a well-graded Un- ment capability (see Section 2.3.2) for various
crushed material; the cement-treated base con- levels of single and tandem axle loading and
sisted of the sand-gravel subbase material number of applications.
combined in a paving mixer with 4.0 percent by Basic information on performance of each
weight of Type I portland cement; and the subsection is given in Table 13 and Appendix
bituminous-treated base was a hot mix consist- A. The values listed for each subsection are
ing of the subbase material and 5.2 percent of either the number of unweighted axle load ap-
85-100 penetration grade paving asphalt (see plications to a serviceability of 1.5 or the final
Table 4). Three of the four types were included serviceability if the subsection was still in test
in six structural sections in each flexible tan- at the termination of traffic. Similar informa-
gent of the four major loops - (two replicate tion for a serviceability level of 2.5 is given in
sections for each base type). The sections were Table 14, and average values for pairs of repli-
160 ft long, and the base was constructed as a cate sections are shown in Figures 30 and 31.
wedge, that is, the thickness of the base mate- The effect of the thickness of the base on per-
rial decreased at a uniform rate from one end formance can be clearly seen except in the case
of the section to the other in the direction of of the gravel base sections in Loop 5.
traffic. The gravel base wedge sections failed early
As in the case of the paved shoulder experi- in the test and are omitted from the analysis of
ment, the 160-ft wedge sections were divided performance. The data in Tables 13 and 14
into 40-ft subsections, and each subsection was show that on Loops 3 and 4 there was a fairly
considered in the major analysis of perform- orderly effect of thickness of the gravel mate-
ance to be represented by its mean thickness of rial. However, on Loop 5, the serviceability of
base. Table 12 gives the pavement designs, all gravel subsections dropped to 1.5 at prac-
base types and average base thickness of the tically the same number of axle load applica-
subsections included in each of the four major tions (less than 100,000), indicating that
loops. .The main objectives of the study were to increasing the base thickness within the range
312-Ill-U
1200 BITUMINOUS
---------------------------------------L
000 0 AVB,OQB Final SonUjOsobilily OK SllbBBEtiOnB
LOOP 3
LOOP 4
12 KIP 18 KIP
§ 800 3-12-141-0
STONE
'I
312-IOI-4
600 CEMENT
GRAVEL
3-12-161-4
SHOVEL
400
3-12-161-4
STONE
20O I
59 76 34 3.1 12.5 9.5 6.5 3.5 2.5 9.5 65 35 NO 70 50 SO 142 10.8 72 38 142 lOB 22 38
AVERAGE BASE THICKNESS OF SUB-SECTION. INCHES AVERAGE BASE THICKNESS OF SUN-SECTION, INCHES
3-13-161-4
1200 BITUMINOUS
40' 3.8' 3.0 5.7'
1 14
1000 LOOP 5
22.4 KIP
b
0 600
.6
3-13-121-4
600 CEMENT
ti 400
•
200
3-13-18 1-4
4.4 11.1 TO 4.0 10.9 9.9 6.4 4.1 16.1 12.4 BA 4.9 10.1 128 8U 4.9 11.6 9.2 6.6 42 170 1S0 9.0 5.0
AVERAGE BASE THICKNESS OF SUB-SECTION. INCHES AVERAGE BASE THICKNESS OF SUB-SECTION. INCHES
Figure 30. Special base type experiment, performance data for single axle loads
(averages for each pair of replicate sections).
FLEXIBLE PAVEMENT RESEARCH 49
TABLE 12
DESIGN OF SPECIAL BASE TYPE WEDGE SECTIONS
Thickness (in.)
Base
Loop Base Type
Surfacing Subbase Average for Subsection
Range
1 .2 3 4
3-12-111-0
zoo BITUMINOUS
s' 40' 32' 50'
1114
0 0961096 FinU SorS'606bilily of SobBoBIlOISS
1000
LOOP3 LOOP4
24 KIP 32 KIP
5 600
US
5 600
3-12-10)-U
CEMENT
400
3-12-0)-U
3-(2-I4)-0 3-12-141-0 3-12-161-4
200 GRAVEL
ORAVEL STONE STONE
O
93 06 54 31 25 95 6.9 35 2.5 95 65 33 90 .70 50 00 IOU 04 7.2 3.0 14.2 10.6 7.2 38
4009UGE BOSE THICKNESS OF 000-SECTION. 16440$ AVEMOISE BOSE ThICKNESS EN S1JS-SECTION, INCHES
3-13-112-4
CEMENT
STONE
CEMENT
GRAVEL
14.4 11.1 7.9 4,6 10,9 6.6 64 4.1 16J IZ4 96 49 61 124 84 49 11,0 92 60 43 ITO 30 90 50
0006400 MUSE T1200NESS OF SUB-SECTON. INCHES OVERUSE MUSE ThICKNESS OF SUB-SECTION, INCHES
Figure 31. Special base type experiment, performance data for tandem axle loads
(averages for each pair of replicate sections).
TABLE 13 Cl'
0
PERFORMANCE DATA TO p = 1.5, BASE TYPE STUDY
Applications (x 10)
'Estimated applications.
TABLE 14
PERFORMANCE DATA TO p = 2.5, BASE TYPE STUDY
Applications ( x lOa)
Subsection ane 1 T
ane
Pavement Base Thickness (in.)
Loop Design and ___________________________
Base Type Avg. Rep. 1 Rep. 2 Avg.
Range Avg. Rep. 1 Rep. 2
'Average assumes 1,114 thousand applications for the replicate that did not fall to p = 2.5 by end of test.
52 THE AASHO ROAD TEST, REPORT 5
provided in Loop 5 did not improve its per- In making the assumptions necessary to specify
formance. It appears that the gravel material the following relationships consideration was
possessed internal stability that was adequate given to the strength characteristics of the
for the loads operating on Loop 3 and nearly so various materials. For example, the Marshall
for Loop 4 but certainly not for the loads on stability of the bituminous-treated base was
Loop 5. nearly as high as that for the surface and
In a graphical analysis of performance of the binder courses (1600 vs 2000 and 1800) and
stone, cement and bituminous base wedge sec- the compressive strength at seven days for the
tions, the relationships between changes in cement-treated base was 840 psi.
serviceability and changes in thickness of the The adjustments were made on the following
base material were first developed for each pair basis:
of replicate sections at several levels of axle From Eq. 19: 1-in, bituminous surface = 3-
load applications (Figs. 32, 33 and 34). The in, stone base, approximately; 1-in, bituminous
plotted points represent mean serviceability surface = 4-in, subbase, approximately; and
vs mean base thickness for each pair of repli- 3-in, stone base = 4-in, subbase, approxi-
cate subsections. The family of curves was mately. It is assumed that 1-in, bituminous
formed by connecting the plotted points for surface = 1-in, bituminous base; and 1-in, bi-
each of the six selected levels of axle load appli- tuminous base = 4-in, subbase. From average
cations. ratios of thickness of bituminous base and
Considerable judgment had to be exercised cement base given in Table 15 for Loops 5 and
in developing these performance relationships. 6: 1-in, bituminous base = 1.30-in, cement
In many cases, especially at the lower levels of base; 1-in, bituminous surface = 1.30-in.
axle load applications, the relationships were cement base.
well established by the plotted points. In other The relationships shown in Figure 35 were
cases, it was necessary to examine the field developed from Table 15 (adjusted thicknesses
condition survey records in order to establish for Loops 3 and 6). Because there was no con-
the trends of the curves. sistent indication of curvilinearity, straight
For example, in lane 1 of Loop 4 two sub- lines were fitted to the plotted points. These re-
sections having a 3-in, mean thickness of lationships compare the three types of base
cement-treated base were removed from the test with respect to their ability to give the same
at fewer than 100,000 applications. Thus, no level of performance under a range of single
points could be plotted for this thickness at and tandem axle loads and at six levels of axle
100,000 applications. The two subsections with load applications. The thicknesses of base are
5-in, mean base thickness had a mean service- those needed when the surfacing thickness is
ability at 100,000 applications of 1.65, but were 3 in. and the subbase thickness is 4 in.
removed from test before 300,000 applications. Figure 36 shows a similar comparison of the
The condition survey records showed that prior three types of base under a range of axle load
to 300,000 applications the surface of these sub- applications for each single and tandem axle
sections had been maintained back into the sub- load included in the test. The values used to plot
sections to a point where the base thickness the curves were taken directly from the rela-
was 5.6 in. Therefore, the trend of the service- tionships in Figure 35. Extrapolation or inter-
ability-base thickness relationship at 300,000 polation were used when necessary. For ex-
axle load applications was established by ex- ample, a curve is shown for each of four loads
tending the curve through a point plotted at in Figure 35 even though only three loads were
p = 1.5 and 5.6 in. of base thickness. The end actually applied to a particular base type.
points for the relationships at 700,000 and Additional information on the performance
1,114,000 applications were obtained in the of the special base type sections is given in
same manner. The curves at 500,000 and 900,- Tables 16 and 17. The thicknesses of base
000 applications were drawn to fit the other represent the amount necessary to prevent
relationships. Class 2 cracking of the surfacing at two levels
From curves constructed in this manner it of axle load applications. The values are appli-
was possible to estimate values of base course cable for the basic designs of the sections
thickness for various levels of axle load appli- (Table 12).
cations at any desired serviceability level. Table Because no preconceived mathematical form
15 gives these values for 100, 300, 500, 700, 900 was used in the graphical analysis of the base
and 1,114 thousand axle load applications at a study data, the curves shown in Figures 32
serviceability level of 2.5. The actual thick- through 36 do not necessarily follow the shapes
nesses listed are not directly comparable across that would be given were the models of Section
loops, or within Loop 6, since the thickness of 2.2.2.1 fitted to the performance data of Tables
surfacing and subbase was not the same for all 13 and 14. It is interesting, however, to super-
sections included in the study. Therefore, the impose the crushed stone base data from Table
base thicknesses for Loops 3 and 6 have been 14 upon the data and curves obtained from the
adjusted to correspond with the thicknesses of performance analysis of the factorial sections
surfacing and subbase used in Loops 4 and 5. (Design 1), all of which had crushed stone base.
FLEXIBLE PAVEMENT RESEARCH 53
Figure 37 shows this superposition, and it may apparent bias. A systematic procedure for
be inferred that the crushed stone base data fitting the augmented data to the models of
from Design 4 generally agree with the corre- Section 2.2.2.1 was under study when this re-
sponding data from Design 1. For the subsec- port was written. It was expected that applica-
tions having the thickest base, however, it tion of this procedure would result in a some-
would appear that the curves from the per- what modified performance equation that would
formance equation may tend to overestimate fit both the factorial and wedge section data for
the thickness index requirements when the crushed stone base, and would provide for the
thickness index involves a considerably greater differential effects on performance of the three
proportion of base thickness than was present base types that were discussed in the graphical
in the factorial experiment designs. Thus if analysis of this section.
the factorial performance data are augmented
by data from the crushed stone base study of 2.2.3 Structural Deterioration
this section, it is likely that certain coefficients In Section 2.2.2 performance was considered
in the performance equations will be modified. to be the trend of pavement serviceability with
However, in Figure 37 it was not possible to applications of load. The three principal ele-
show data points for those subsections whose ments that detract from serviceability of the
serviceability had not dropped to 2.5 by the end flexible pavements were slope variance (a meas-
of the test. Many of these points would fall to ure of longitudinal roughness), rut depth, and
the right of the curves and tend to reduce the cracking and patching.
TABLE 15
Thickness (in.)
1 Adjusted to correspond with thicknesses of surfacing and subbase used in Loops 4 and 5.
54 THE AASHO ROAD TEST, REPORT 5
4 4
3-(2-14)-0 Desigh 3-(2-14-0 NO
12Kip Single 24Kip Tandem
Clo
3
+00
10 12 K 8 10 2
,
4
3-(2-16)-4 Desig 3-(2-16)'-4 Desig
x Single 32
Ui
0
z
- 3
13
0 8
80 12 14 16 10 12 14 16
a
4-(3-19)-8 Desgn
30 Kip Single
I
4
I-
C'
810 12 14 16 18 8 10 12 14 16 18
BASE THICKNESS, INCHES BASE THICKNESS, INCHES
Figure 32. Special base type experiment, crushed stone base performance data
(averages of two replicate sections).
FLEXIBLE PAVEMENT RESEARCH 55
*
W
0
4
3-(2-I0)-4 Deslgn
18_KipSingle 7' +
4
13
C' - ------ 0
2 4 6 8 10 2 4 6 8 10
C' I,
4 6 8 10 12 4 6 8 10 12
5 5
4
0
z
I-.
3
C' C'
4 6 8 10 12 4 6 8 10 12
Figure 33. Special base type experiment, cement-treated base performance data
(averages of two replicate sections).
56 THE AASHO ROAD TEST, REPORT 5
.(td —±--
4
w 7
2 I)_OLesign
12 Kip Single - APPL.
:
CURVE (1,000) 3-(2-10-0Design
0-0 100
b--f 300 24 Kip Tandem
c—V 500
d— 700 2
e—S 900
f—x 1,114
Ix [3 From Coed.
Surveys
rf2
C,
2 4 6 8 10 2 4 6 8 10
—o
Lx
3- 3sin.
3
4 Kip Single
0 0
4 6 8 10 12 14 4 6 8 10 12 14
x4 , 4
H8)-4 Design - 3
/ -/141--
T
- 4 6 8 10 12 14 16 - 4 6 8 10 12 14 16
BASE ThICKNESS, INCHES BASE THICKNESS, INCHES
STONE
-
I'
IS
I
CEMENT TREAL ED
5
BITUMINOU
(300,000
L 70NE
SIONE _--- 0 __
15 'S
_I
0 10
Figure 35. Special base type experiment, relationship between base thickness and
axle load at p = 2.5 (surfacing, 3 in.; subbase, 4 in.).
TABLE 16 TABLE 17
SPECIAL BASE SECTION PERFORMANCE DATA, SPECIAL BASE SECTION PERFORMANCE DATA,
BASE THICKNESS NOT ASSOCIATED WITH CLASS 2 BASE THICKNESS NOT ASSOCIATED WITH CLASS 2
SURFACE CRACKING, FEBRUARY 20, 1960, SURFACE CRACKING, NOVEMBER 30, 1960,
500,000 AXLE APPLICATIONS 1,114,300 AXLE APPLICATIONS
Gravel 1 9.3 OT OT -
- Gravel 1 OT OT OT -
-
2
9.5
10.4
OT
OT
OT
OT -
- 2
OT
OT
OT
OT
OT
OT -
Stone 1
11.3
7.7
OT
9.2
OT
-
- 11.0 Stone 1
OT
11.4
OT
11.0 -
OT
- 11.0
2
7.7
9.4
9.2
9.1 -
-
11.0
7.6 2
13.0
13.4
13.2
11.6 -
-
11.2
8.0
Cement 1 -
10.4
-
9.2
6.2 6.9
7.6
9.1 Cement 1 -
13.5
-
12.6
7.8 8.0
9.2
9.5
2 -
-
6.3
6.2
7.0
6.9
9.1
7.7 2 -
-
7.5
7.5
8.8
8.2
10.2
8.0
Bituminous 1 2.4
6.4
-
-
6.9
5.3
7.7
6.7 Bituminous 1 2.6 -
7.5
-
8.2
5.3
8.1
7.6
2
2.6
3.1
3.1
-
-
5.3
5.3
5.3
6.7
5.7
5.7
2
2.8
3.2
3.8
-
-
6.0
5.6
5.6
8.0
6.8
6.8
OT = out of test. Thickness is average of two sub- I OT = out of test. Thickness is average of two sub-
sections; for thickness of surfacing and subbase see sections; for thickness of surfacing and subbase see
Table 12. Table 12.
58 THE AASHO ROAD TEST, REPORT 5
2.2.3.1 Transverse Profile Changes.—Studies Rutting of the pavement was due principally
were made of the seasonal changes in elevation to decreases in thickness of the component
of the pavements and of the rutting in the layers. Based on average data from 51 sections
wheelpaths. The studies of rutting included that were trenched in 1960 (see Tables 19, 20
such factors as the extent to which changes in and 21), a rut could be attributed to changes in
thickness of the structural components affected thickness of 32 percent, 14 percent and 45 per-
the depth of rut, and how much of the thickness cent, respectively, in surfacing, base and sub-
change
duewas due todisplacement.
densification and howwere
much base, and to a rut in the embankment soil equal
was to lateral Studies to 9 percent of the total rut.
made also of the seasonal changes in physical
condition and strength of the pavement corn- Only 20 percent of the change in thickness of
ponents. the surfacing and 4 percent of the change in
On an average, the pavement in the various subbase thickness could be accounted for by
loops heaved approximately 0.4 in. during the increases in density of the materials. In the
winter with the edges rising about 0.6 in. and case of the base only 30 percent of the change
the interior portion about 0.3 in. (see Fig. 44). in thickness determined in the summer of 1960
Most of this heaving was attributed to expan- could be accounted for by increases in density.
sion of the embankment soil. However, the increase in the density determined
20 ac
— 40 KIP_I
32 KIP
IS KIP
C,
0 200 400 600 800 1,000 100 0 200 400 600 800 1,000 1,00
15
30 KIL
10
22 I
32
I,
Bituminous—Treated Base
C,
0 200 400 600 800 1.000 1.200 - 0 200 400 600 800 1,000 1,200
SINGLE AXLE LOAD APPLICATIONS AT p 2.5, IOOs TANDEM AXLE LOAD APPLICATIONS AT p2.5, 10001$
Figure 36. Special base type experiment, relationship between base thickness and
axle load applications at p = 2.5 (surfacing, 3 in.; subbase, 4 in.).
FLEXIBLE PAVEMENT RESEARCH 59
I' LJ I L.
10 100 000 10 100 1000
32 Kip Tandem
=1-I
I'
0 100 1000 10 100 1000
I!li iI iIflhI I
5
III
48 Kip Tandem xj JJfl
4
3
10 lC) 1000
WEIGHTED AXLE LOAD APPLICATIONS AT p2.5. 1000's WEIGHTED AXLE LOAD APPLICATIONS AT p,2.5, 1000's
in the spring of 1960 accounted for all of the Since corrective measures appropriate to the
decrease in thickness of the material. prevention of rutting must depend to a large
In sections that survived the test, the rate of extent on the reasons for the rutting, consider-
development of rutting, during the first year of able effort was expended in the study of these
traffic generally exceeded the rate observed dur- matters.
ing the second year. Information for use in these studies was ob-
In the special base studies there was a level tained from a number of different types of
of base thickness above which the surface rut measurement, as follows: (1) Six sets of sur-
depth remained constant with increase in base face elevation data in which conventional pre-
thickness and below which it increased rapidly cise level surveying 'techniques were used to
with decrease in thickness. establish the transverse profile from measure-
The bituminous-treated base and surfacing ments at three locations per test, section. (2)
material offered greater resistance to consolida- 'Transverse profile measurements with the auto-
tion and displacement at low temperature than matic recording electronic 'device, shown in
at high temperature. Figures 38 and 39. The continuous profile
In nearly all of the flexible pavement test traces were used with the precise level measure-
sections ruts were formed in the wheelpaths ments to establish the elevation at every foot
during the course of the traffic tests (see Fig. across the pavement. (3) Settlement rod
66). Since there appeared negligible loss of measurements. After construction, vertical rods
surface material due to abrasion or other were installed in the pavement and anchored
causes, the observed rutting must be attributed to plates previously placed at the various pave-
to one or both of two conditions: (1) addi- ment layer interfaces and in the embankment.
tional consolidation under traffic of one or more Layer thickness changes were' determined from
layers of the pavement structure and/or the measurements of the surface elevations with
embankment material; and (2) displacement reference to the tops of the rods as shown in
outward from the center of the wheelpath of Figure 40. (A full description of the settlement
material in one or more layers of the pavement rods, their installation and the measurement
structure and/or the embankment material. program is contained in Appendix 4.) (4) Pre-
60 THE AASHO ROAD TEST, REPORT 5
-.
- -.- - -,- - -- -- .. -
cise level measurements made on the surfaces in the entire pavement with seasons. For ex-
and the tops of the subsurface layers during ample, the 3-6-4 pavement under the 2- and
the trenching program. From these measure- 6-kip axle loads of Loop 2 rather consistently
ments made every foot across the section, cross- went up about 0.04 ft between October 1959 and
sections of the pavement and embankment sur- March 1960. By October 1960, the pavement
face were obtained. (5) Routine biweekly rut had gone down to about 0.01 ft below the Octo-
measurements using the device shown in Figure ber 1959 level. Although the magnitude of this
41. (6) Measurements of rut depth at 5-ft rise and fall varied from section to section, the
intervals in the special (wedge) base sections, same general trend was observed in most of the
made with the same device, five times during pavements. The rise of the pavement surface
the test period. noted in the early spring (March) of 1960 is
Transverse profiles of the pavement surface attributed to the presence of frost in the struc-
taken with precise levels at different times dur- ture and embankment.
ing the traffic test are summarized in Table 18 The changes in elevation at the edge, outer
and are shown graphically in Figures 42 and wheelpath, between wheel paths, inner wheel-
43. For any one section comparison of the pro- path and at the center of the pavement between
files taken in the three periods demonstrates the the fall of 1959 and spring of 1960 are shown
magnitude of the absolute change in elevation in Figure 44. There was almost twice as much
heaving at the extreme edges of the pavements
as in the interior portions. This situation was
much the same for all loops although the actual
amount of heaving varied appreciably from
loop to 1001). The greater heaving at the pave-
ment edge was attributed to the presence of
more moisture in the base, subbase and embank-
a b a d I g h
0
ment soil beneath the shoulders. Heaving de-
creased from Loop 1 through Loop 4, the
RUT DEPTH b-
rcrse trend occurred from Loop 4 through
a,b.c..?c ARE PNEUMATICALLY OPERATED DEPTH PRCBE'
Loop 6 (Fig. 44). To this extent the data are
Rut Depth Mode
inconsistent insofar as the effect of pavement
ELECTRONIC LEVEL-1 thickness on frost heaving is concerned.
Data concerning the condition of the embank-
1--FIXED LEG ment soil in the spring for Loops 3, 4, 5 and
ADJUSTABLE LEG
TQLVfJ.1RJ55.. 6 in 1960 and for Loops 2 and 4 in 1961 are
given in Table 19. (The data from the 1960
MOVEGALE DEPTH CURR AGE -
study were obtained from the trenching pro-
gram described in detail later in this discussion.
Continuous Transverse Profile Mode Data from the 1961 trenches were obtained dur-
Figure 39. Schematic of transverse profilometer. ing the post-traffic special study program de-
FLEXIBLE PAVEMENT RESEARCH 61
I2 Iz
3-6-0 Design
1 -
364 Design
---- --
4-6-8 Design
z
0
I-
>
-j
z
-Ic
z
12 KIP SINGLE 24 KIP TANDEM
U
4-6-4 Design
-Ic
IC
5-6-4 Design
4-9-12 Design
5-9-12 Design
z
-Ic
6-9-8 Design
6-9-16 Design
zI- -
5-3-12
5-9-4
wo
04
ci
ouJ
0.2
01• • - • - - - - - -
EOBI EOBIQ EOBIG EOBI EOB I Q
scribed in AASHO Road Test Report 6.) The moisture exhibited no heave. However, at 85
computed value of saturation of the soil for percent saturation, 15.5 percent moisture and
Loop 2 (OWP, spring 1961) was the highest at the same density the soil heaved about 1.7
(82.3 percent). Also its density was appre- percent. Since soil conditions in Loop 2 were
ciably lower and its moisture content higher more adverse than those in the other loops
than these values for the other loops. (nearly 85 percent saturation was noted) a
Information on the frost potential of the greater degree of heaving was expected in this
embankment soil reported by the Corps of loop.
Engineers (DS 2300) shows that when frozen Additional data regarding vertical move-
in the laboratory the material at 72 percent ments of the flexible pavement structure are
saturation, 114-pcf density and. 13 •percent shown in Figures 45 and 46. They were ob-
TABLE 19
CONDITION OF EMBANKMENT SOIL, SPRING MONTHS
1960 1961
Loop Number of Moisture
Content Density Saturation CBR Number of Moisture
Content Density Saturation CBR
Sections (pcf) (%) Sections (pcf) (%)
2
3
-3 15.1
- -
113.3
-
80.7
-
2.5 -26 17.1
- 109.2
- 82.3
- 1.8
-
4 3 14.1 115.4 79.6 4.0 14.1 116.8 81.3 5.0
5
6
3
3
16.1
13.1
110.2
115.1
79.5
73.4
2.7
6.6
-
- - -
- -- -
-
All 14.6 113.5 78.4 - - - - - -
(b) BETWEEN WHEELPATHS
2
3
-3 15.0
- -
112.5
-
78.5
-
2.7 -62 - 16.8
- 108.6
- - 79.6 2.8
-
4 3 14.3 113.8 78.9 3.7 13.8 116.0 79.6 3.8
5
6
3
3
16.2
13.9
109.4
115.1
78.3
77.9
3.0
3.9
-
- -
- . -- -
- -
-
All 14.9 112.7 78.4 - - - - - -
FLEXIBLE PAVEMENT RESEARCH 67
tamed from measurements taken in the fall and veloped in both wheelpaths of the pavement
spring periods of the traffic testing phase. The surface, very little of it was apparent in the
figures show the seasonal changes in elevation embankment soil. This was considered to be
in the wheelpaths of the pavement surface and evidence that reduction in thickness of the sur-
of the top of the embankment soil in ten test facing, base and subbase courses was to a very
sections. Although some heaving took place large degree, responsible for the rutting ob-
within the pavement structure during the win- served in the wheelpaths of the pavement sur-
ter, the greater part of the heaving observed at face. If sections that were failing at a rapid
the top of the pavement originated in the em- rate were not maintained, rutting or distortion
bankment soil and the major part of the settle- of the pavement in the wheelpaths would prob-
ment of the pavement surface occurred as a ably extend into the embankment soil as was
result of movements within the structure. This the case in the 5-3-4 section.
is in agreement with the findings of the trench Layer thickness changes were also measured
studies that are discussed later in this section. by means of settlement rods anchored to plates
The transverse profiles shown in Figures 42 located at various levels in the pavement struc-
and 43 not only present a clear picture of the ture and on the embankment. Examples of these
magnitude of the ruts in the surface of these data are shown in Figures 48 and 49 for the
particular pavements but show the location of crushed stone base and the bituminous-treated
the ruts with respect to the pavement edge. base. Generally, for all bituminous-treated base
These locations for the 18 test sections are thicknesses and for all stone base thicknesses
summarized in Table 20. For these particular greater than 11 in., very little thickness change
sections ruts average 2.9 and 8.9 ft from was found in the material underlying the base.
the edge in the single axle lane and 3.2 and Sections with these thicknesses of base also had
9.0 ft from the edge in the tandem lane. There small losses in serviceability over the test
is reasonable consistency among the means for period. This is considered evidence that the
the, group of sections in each loop. In these rutting in pavement designs of substantial
sections the tandem axle vehicles were appar- thickness can be attributed primarily to reduc-
ently driven closer to the centerline of the pave- tion in thicknesses of the pavement layers.
ment than the single axle vehicles (6 in. on the Further information was obtained from
average). trenches cut into 12 sections in the spring and
During 1959, a number of trenches were cut summer of 1960 and from 27 more trenches cut
into sections, whose conditions had deteriorated in the fall of 1960. Data from these tests are
to the point where their removal from the test given in Tables 21, 22 and 23, and a summary
was imminent. In fact, the serviceability of of the data is shown graphically in Figures 50
some of these sections had dropped below 1.5, and 51. Values of the actual change in thick-
the level at which they were normally removed ness and those computed due to densification
from test. The purposes of this work was to are listed for each section trenched, for the
develop some preliminary information concern- mean of the sections in each loop and for all
ing the amount of wheelpath rutting at the top loops combined. Also listed are initial values
of each of the component structure layers as of thickness and density of each component and
well as to obtain information on the existing similar values determined at the time of trench-
condition and strength of the materials. Also, ing. The means of the sections for each loop
since a comprehensive program of trenching (Fig. 50 and 51) are for measurements made in
work was being planned for 1960, another ob- the outer wheelpath and between the wheel-
jective of the work was to develop an accept- paths. In addition to the change. of thickness
able procedure for executing the program. data for the pavement layers, the depth of rut
While the trenches were being made, precise reflected in the subgrade is indicated for the
levels were taken at 1-ft intervals on the top outer wheelpath in Figure 50. A comparison
of each of the layers. The trenches were 3 ft in of the magnitude of the embankment rut to the
width and the levels were taken on both faces. total pavement thickness change serves as
In addition, cores for density determinations further evidence that most of the observed sur-
were taken of the surfacing course, and in- face rut can be attributed to changes of thick-
place density, CBR, and moisture content deter- ness of the pavement layers.
minations of the granular materials and em- Comparison of the as-constructed densities
bankment soil were made. In some cases plate and the densities measured in the trenches in
load tests were made on the embankment soil. the three programs provide a means for assign-
Transverse profiles of three of the 1959 ing portions of the change in thickness of each
trenches are shown in Figure 47, two from component to densification. The amount of
Loop 6 and one from Loop 4. Those in Loop 6 change of thickness that would be expected due
were cut when the serviceability of the pave- to the measured change in density is shown for
ments was about 1.5, and that in Loop 4 when each loop and each structural component in the
its serviceability was about 0.5 and failure was solid bars (except that densities of base and
in a more advanced stage. In the two Loop 6 subbase were not determined in case of the fall
sections, although pronounced rutting had de- group of trenches). The total observed thick-
CHANGE IN STRUCTURE THICK.. IN. CHANGE IN ELEV.OF TOP OF EM8..IN. CHANGE IN ELEV. OF TOP OF PVMT.. IN.
0 00 ' (
0 CA
C'
TTTEIITIIIT ___ ___ ORIG
PR59 0
--\------ 1-159R
PR6O
ALL6O
-DRIG
EEt E EE SPR 60
'ALL 60 CD
x
II
I E1uj
tiRIG
I__J'----- I_ 1__---------- I
-ALL59Q
PR 60
ALL6O
FALL 60
--- ------')RIG
---ALL58
4? - - - -- - PRS
----+=T-----+--------•<_L.. --------•-::T
I_I-.------ 3PR6O
'ALL60'
----RIG58
JPRSg
FALL O CD
)RIG
ALL58
5PR59
OFALL 59 0
ALL60
ORIG
SPR 60
FALL 60
- --RIG
----ALL580
SPRS9
-------------------....
--=--
-- ------
--
- __
-----------
__
ORIG
---__
IF
----- -- ---- _
--- -------- ----..
58.
---.
-- FALL .0
ORIG
-------------- -- 0
FALL .0
0•
-- ----- ---------...
----------- ------...
01
a
-----
--- -------
----------_ -._
------------------..,
-- -- U
uu um --_ 00
.. :.
-------------------...,
----- -------------....
.0
0 2 4 6 8 10 12
4,,
5.70"
4 3.24"
-F
3.42"
-J
0 2 4 6 8 10 2
o
ZZZ ZZZ z
o o o o o o o o
20
SUBBASE SUBBASE
U)
LU
C)
z
IIFJTCING PLUS
!jT11T SURFACING PLUS BASE
'I'I I SUBBASE
SUBBASE
ness change, for whatever cause, is shown for concrete accounted for all of the total thickness
each ioop and structural component in the change in the surfacing material. The base
hatched bars. The fall 1960 trenches (near the course in nearly all the trenches became thicker
end of the traffic test) indicated that only about rather than thinner between the wheelpaths
25 percent of the thickness change in the sur- without undergoing much change in density.
facing in the outer wheelpath of these sections Presumably the material was forced into this
could be assigned to densification of the ma- position from the wheelpath locations. Thus
terial. The summer trenches indicated that between the wheelpaths there was considerable
about 25 percent of the thickness change in the reduction in subbase thickness accompanied by
subbase under the outer wheelpath and less a reduction on the. average in subbase density.
than 50 percent of the thickness change in the The data from all the trenching studies lead
base material of these sections could be assigned to the conclusion that changes in thickness of
to densification. the components of the flexible pavements at the
Data taken from between the wheelpaths in AASHO Road Test were due primarily to
the trenches are shown summarized by loops in lateral movement of the materials.
Figure 51. (Values for each section are avail- In connection with the data given in Tables
able in DS 4180.) Densification of the asphaltic 21, 22 and 23, in the outer wheelpath there was
Spring 1960 Summer 1960 FoIl 1960
LOOP 3 LOOP 4 LOOP 5 LOOP 6 MEAN LOOP3 LOOP4 LOOP5 LOOP6 MEAN LOOP3 LOOP 4 LOOP5 LOOP6 MEAN
SURFACING SURFACING SURFALG
-- ——— ——
- -— BASE
U - — 1-
BASE BASE+
z
- -1.0 - •
TOTAL THICKNESS CHANGE
THICKNESS CHANGE DUE TO . - - - -4- --DENSITIES NOT DETERMINED
DEN SI FICAT ION
E RUT DEPTH
SUBBASE SUBBASE
tm 1
SUBBASE +
0 —ff]
-1.0
I-u
w
x
C)
z
TOP OF EMBANKMENT TOP OF EMBANKMENT TOP OF EMBANKMENT
0 co . . n
_
-1.0
Figure 50. Summary of layer thickness changes in outer wheelpath, 1960 trench study.
BASE
- BASE - BASE +
Figure 51. Summary of layer thickness changes between wheelpaths, 1960 trench study.
TABLE 20
LOCATION OF RUTS FROM TRANSVERSE PROFILES;
DISTANCE TO CENTROID OF RUT FROM PAVEMENT EDGE
2 3-6-4 8.6 None 8.1 8.4 2.6 3.5 4.0 3.4 8.0 8.1 8.0 8.1 4.0 3.0 4.0 3.7
3 4-6-4 8.1 8.0 - 8.1 2.0 2.5 - 2.3 9.0 9.0 - 9.0 4.0
4.0
3.0
3.0
--
3.0
3.5
3.4
4-6-8 8.5 8.5 8.5 8.5 3.0 3.0 3.0 3.0 9.0 9.0 9.0 9.0
Mean 8.3 2.7 9.0 3.5
4 5-6-12 8.5 8.1 8.0 8.2 3.0 3.0 3.0 3.0 9.0 9.0 9.0 9.0 3.5 3.0 4.0 3.5
3-6-12 8.7 8.3 8.0 8.3 3.0 3.0 3.0 3.0 9.0 9.0 9.0 9.0 3.2 3.7 3.5 3.5
4-6-12 8.7 8.7 - 8.7 3.0 3.0 - 3.0 8.7 8.8 - 8.8 3.3 3.0 - 3.2
3.3
5-0-12 8.2 8.3 8.0 8.2 3.0 3.0 3.0 3.0 8.7 9.0 9.0 8.9 3.0 3.8 3.0
5-6-4 9.0 9.0 - 9.0 3.0 3.0 - 3.0 9.0 9.0 - 9.0 4.0 3.0 - 3.5
3.4
Mean 8.5 3.0 8.9
5 4-9-8 9.0 8.7 - 8.9 3.2 3.0 - 3.1 8.2 8.3 - 8.3 3.3 3.3
3.5
-
3.0
3.3
3.2
5-9-12 8.5 8.5 8.3 8.4 3.0 3.0 2.5 2.8 9.0 9.0 9.0 9.0 3.0
3-9-12 9.0 8.7 8.7 8.8 3.0 3.0 3.0 3.0 8.5 8.7 9.0 8.7 3.0 3.0. 3.0 3.0
5-9-4 9.0 8.8 8.5 8.8 3.0 3.0 3.0 3.0 8.7 9.0 9.0 8.9 3.2 3.5 3.0 3.2
5-9-4 8.7 8.7 8.7 8.7 3.0 3.0 3.0 3.0 8.0 9.0 8.7 8.6 3.0 3.3 3.0 3.1'-
.1k
Mean 8.7 3.0 8.7 3.2
2.3 2.3 2.3 2.3 9.0 9.0 9.0 9.0 3.5 3.7 3.0 342
6 6-3-16 8.0 8.0 8.0 8.0
6-9-8 8.0 8.0 - 8.0 3.0 3.0 - 3.0 9.0 9.0 - 9.0 3.0 3.0
3:0
-
3.0
3.0
3.0
6-9-8 8.0 8.0 8.0 8.0 2.5 3.0 2.7 2.7 9.0 9.0 9.0 9.0 3.0
4-9-16 8.7 8.5 8.3 8.5 3.0 3.0 3.0 3.0 9.0 9.0 9.0 9.0 3.0 3.0 3.0 3.0
6-9-16 9.0 9.0 9.0 9.0 3.0 3.0 3.0 3.0 9.0 10.0 10.0 9.7 3.0 3.0 3.0 3.01
Mean 8.3 2.8 9.1 3.1
Total mean 8.4 3.0 8.8 3.4
TABLE 21
CHANGES IN THICKNESS AND DENSITY, OUTER WHEELPATH,
TRENCH PROGRAM, SPRING 1960
Thickness (in.) Density (pfc) Change in Thickness (in.)
Loop Design Total Due to
Initial1 Trench2 Initial3 Trench4 Observed Densification
(a) SURFACING
3 4-3-8 4.04 3.81 149.3 150.8 -0.23 -0.04
4-6-4 4.17 3.65 148.9 148.3 -0.52 + 0.02
4-6-8 3.93 4.07 149.2 149.9 + 0.14 -0.02
Mean 4.04 3.84 149.1 149.7 -0.20 -0.02
4 5-6-12 5.31 4.73 149.2 151.6 -0.58 -0.09
5-6-8 4.87 4.94 149.1 150.9 +0.07 -0.06
5-3-12 4.90 440 148.6 152.5 -0.50 -0.13
Mean 5.03 4.69 149.0 151.7 -0.33 -0.09
5 5-9-12 5.03 4.29 150.5 150.8 -0.74 -0.01
5-6-12 5.03. 4.83 149.0 150.8 -0.20 -0.06
5-9-8 5.06 4.53 149.3 149.6 -0.53 -0.01
Mean 5.04 4.55 149.6 150.4 -0.49 -0.03
6 6-6-16 5.66 5.90 149.5 153.0 + 0.34 -0.13
6-9-12 5.84 5.48 149.6 150.3 -0.36 -0.03
6-9-16 5.94 5.80 148.1 151.2 -0.14 -0.12
Mean 5.78 5.73 149.1 151.5 -0.05 -0.09
Over-all Mean 4.97 4.70 149.2 150.8 -0.27 -0.05
(b) BASE
3 4-3-8 3.32 3.13 145.0 143.4 -0.19 +0.04
4-6-4 5.60 5.66 142.6 145.8 +0.06 -0.13
4-6-8 5.96 5.96 143.5 149.1 0.00 -0.23
Mean 4.96 4.92 143.7 146.1 -0.04 -0.08
4 5-6-12 6.50 6.45 140.6 146.3 -0.05 -0.26
5-6-8 6.00 6.09 140.7 149.0 + 0.09 -0.35
5-3-12 3.10 2.87 139.9 137.2 -0.23 +0.06
Mean 5.20 5.14 140.4 144.2 -0.06 -0.14
5 5-9-12 9.24 9.53 137.1 147.8 +0.29 -0.72
5-6-12 6.14 6.09 143.1 140.5 -0.05 +0.11
5-9-8 9.02 8.74 138.6 146.9 -0.28 . -0.54
Mean 8.13 8.12 139.6 145.1 -0.01 -0.32
6 6-6-16 6.16 5.77 138.1 141.5 -0.39 -0.15
6-9-12 9.22 8.66 141.9 136.2 -0.56 +0.37
6-9--16 8.60 8.57 140.2 141.4 -0.03 -0.07
Mean 7.99 7.67 140.1 139.7 -0.33 +0.02
Over-all Mean 6.57 6.46 140.9 143.8 -0.11 -0.14
(c) SUBBASE
3 4-3-8 7.98 7.37 131.7 134.0 -0.61 -0.14
4-6-4 3.74 3.72 133.6 137.4 -0.02 -0.11
4-6-8 8.14 7.56 134.0 128.3 -0.58 + 0.35
Mean 6.62 6.22 133.1 133.2 -0.40 -0.01
4 5-6-12 11.38 11.04 130.3 143.4 -0.34 -1.14
5-6-8 7.98 7.19 136.8 135.2 -0.79 + 0.09
5-3-12 1232 11.02 137.3 135.2 -1.30 +0.19
Mean 10.56 9.75 134.8 137.9 -0.81 -0.24
5 5-9-12 12.12 11.54 136.7 131.2 -0.58 +0.49
5-6-12 11.96 10.84 135.9 134.7 -1.12 +0.11
5-9-8 7.88 7.46 129.3 135.3 -0.42 -0.37
Mean 10.65 9.95 134.0 133.7 -0.71 + 0.02
6 6-6-16 15.60 14.91 139.5 131.3 -0.69 + 0.92
6-9-12 11.88 11.48 136.9 134.8 -0.40 +0.18
6-9-16 16.54 16.27 136.6 141.3 -0.27 -0.57
Mean 14.67 14.22 137.6 135.8 -0.45 + 0.19
Over-all Mean 10.63 10.03 134.9 135.2 -0.59 -0.02
'Cores taken at 1, 6 and 11 ft from pavement centerline at third points in section; data are interpolations from
these measurements.
' Thickness deter-mined from transverse profile plot at maximum depth of rut; surface profiles prepared from
25 precise level measurements at 1-ft intervals.
Average of two tests made at randomly selected locations.
Average of two tests in outer wheelpath, one from each side of trench.
FLEXIBLE PAVEMENT RESEARCH 75
TABLE 22
CHANGES IN THICKNESS AND DENSITY, OUTER WHEELPATH,
TRENCH PROGRAM, SUMMER 1960
more consistency in the behavior of the surfac- wheelpaths data (Fig. 51). Here, as expected,
ing than of the base and subbase components. there was less reduction in the thickness of the
In fact, the thickness of the surfacing of all the layers than in the wheelpaths; and as far as
sections that. were trenched had decreased in the surfacing is concerned, it was all accounted
94 percent of the cases and the density had for by changes in the density of the material.
increased in 98 percent. For the base course Actually since a large part of the change (re-
these values were 75 and 76 percent and for the duction) in thickness of the surfacing in the
subbase course 42 and 94 percent, respectively. wheelpaths was apparently due to lateral move-
The summary of the data (Fig. 50) serves as ment, it is not clear why this material did not
further evidence that there was more uniform- exhibit an increase in thickness between the
ity in the changes in thickness of the surfacing wheelpaths rather than a decrease.
in the wheelpaths particularly for the summer In connection with the data in the tables and
and fall trenches than in case of the granular figures., the sum of the thickness changes in
courses. This is also shown for the between each layer plus the rut measured on the top of
00 U) 00 -C] 0
0_0. 000 .0- 0 - r4'i '1
000000 0000o000 6 6ddod o
00
000C)U)1.0 0,
- 0'U)t-,-40o 0 CQ)I..) 0) 'C']00C']C']CC CCC0(00,-4CCCC
r 1'. c C')CCc-4
C CQCQCQ'I C LL)- 00 COC')CCL))C']-* CC LO C)COCCCCC)
C'] C'] C'] C'] C'] C'] CC C'] CC C'] CC C'] CC C'] CC C'] C'] C'] C'] C4 C'] C']
00 , 00 CO 00 CO ( 00 '-400 '-4 ' 00 ' '-1 ' ' .-4 '-4 4 CO 0000-̀ O00,-,-400,-400
1144
COO)
I
C')
I II I I I
CCC)CCCCO)
11111
CO U) CO CO CO C') 4 ' U) U) ' ' ' ' IL) U) 4)) IL) CC CC CC
I I I I
0000L))
-I I I I I
00000000']'
0)
4)) CC 4))
FLEXIBLE PAVEMENT RESEARCH 77
the embankment soil does not always equal the of the asphaltic concrete with axle applications
depth of the existing rut, on the pavement sur- for the thickest pavement designs in the fac-
face. This is because any uplift of the surfac- torial experiment, Loops 5 and 6. As a refer-
ing on either side of the rut will add to the ence, the 50-blow Marshall density is shown as
depth of the rut. However, if this is ignored an the horizontal line on each plot. Initial data
approximation of the extent to which changes were taken from the as-constructed density
of thickness of each of the layers of the pave- measurements. Later, data were obtained from
ment contributed to the rut can be made. For special tests during the period of test traffic.
example, in the 51 sections trenched in 1960, Figure 53 shows the effect of type of base mate-
changes in thickness of the surfacing, base, sub- rial on the void content of the asphaltic con-
base and the embankment soil contributed 32, crete surface with axle applications. The great-
14, 45 and 9 percent, respectively, to the total est change in void content took place in case of
rut. the stone base sections and the least change in
Figure 52 shows the increase in the density the bituminous-treated base sections.
04)
U) C') U)
.2
—- 0
'-
0) 0) Cl)
U) C') 0)
0)
( 0)
Cd
') 0)
N'N00 C)C'1 00 'o 00 4O 00C')COL ,-1C)NC)0) 'l NCO UZI -41f)C)CO* 00
CO ' '-, 0000 CO Cl! CO,-4' 3 000 00000C' Lq OC') "-* *
- 0
000000 000 00 l0 000 000000000 o
++I I IIHH+IIIIl I I I I -•I-• I .1 I I I I. I I I I I I I -'I-' I I I -1-• I I I I I U)'
U) 0) 0)
Q
C') 4)
v '4-4
ft'
U)j S3qiSUp U)j
U)-
E 0)
•
'-
-4- - 042
CO 0) U)
0) 4)0) 4)
C') C') C") C") C') C') 00 C') CC C') CC C') CC (N CC C') C') '-4
00,-400.-400 00,-100'H00'H00,-400
00
IICC I C IICC& II J,I 0') -4-4
4-4 U)'H
- 0 0) 0
')
-4--
U)
Of) C') C')
(4'
C
CO' CC o
0) ', %-,
(.4 ,. (.4 C')
THE AASHO ROAD TEST, REPORT 5
phase. The lines fitted to the points by eye Single Design Thickness (in.)
-
Base (in.)
indicate that there was a level of base thickness
above which the surface rut depth remained
Axle'
(kips) Surf.
-
Base Subbase Bitum. Cement Stone
essentially constant and below which it. in-
creased rapidly. 12 3 Variable 0 4 5' --
Disregarding the variations in surfacing and 4 61 - 8 14
18 3 Variable
subbase thickness from ioop to loop and inter-
polating or extrapolating for the loops where 22.4 3 Variable 4 7 10 14'
each base type was not represented, estimated
values of minimum base thickness above which 30 4 Variable 4 9.5 12 -
little or no reduction in surface rutting was 30 4 Variable 8 - - 13
noted with increase in base thickness are as
follows: 1
Interpolated or extrapolated.
.61
We
.4
I
-2
C_ --- --- J_ J_ 0
.6
LL
0 200 400 600 600 1000 12C
0 .T_!
-; •— .4
.2
200 400 600 800 1000 12 0 200 400 600 600 1000 IN
4C 40
20 20
00 200 400 600 800 1000 1200 00 200 400 600 800 1000 121
UNWEIGIITED AXLE LOAD APPLICATIONS, 1000'S UNWEIGHTED AXLE LOAD APPLICATIONS, 1000'S
Figure 54. Effect of axle load application on depth of rut; typical sections from
main factorial experiment.
THE AASHO ROAD TEST, REPORT 5
Relationships between rut depth and axle could be established between cracking, pave-
load for the late summer of 1959 and for the ment design and load applications.
fall of 1960 are given in Figures 60 and 61. Eq. 29 was developed from which the number
The values of rut depths are for thicknesses of of axle loads sustained by the pavement before
the bases where rutting was no longer reduced Class 2 cracking of the surface occurred could
with further increase in thickness. Relation- be computed for any design and load. By in-
ships of this type for the crushed stone base cluding a deflections term in the equation it
are not shown since it was not possible to deter- was found that a somewhat better prediction
mine the thickness of this material beyond of load application could be obtained (Eq. 30
which the depth of rut for the 12-kip single and and 31).
the 24-kip tandem axle loads remained constant More surface cracking occurred during
(Fig. 59). periods when the pavement structure was in a
2.2.9.2 Cracking.—Cracking was also an ele- relatively cold state than during periods of
mentof structural deterioration that detracted warm weather.
from serviceability and performance of flexible Generally, cracking was more prevalent in
pavement. Records were maintained of the de- sections having deeper ruts than in sections
velopment of cracks in order that relationships with shallower ruts.
.40 - - -
.21
.81
Bituminous Treated Base
-- .8 I -- 200 400 500 800
•I ---
Bituminous Treated Bose
--
- 4-10.5-4 Design 4-10.5-4 Design
30 Kip Single Axle Load mAe Load
.60 .6
41 - .4
.21 -.- -- - .2
0 - - - - - -
.8
Slone Bose
--- -
200 400 600 800 1000 1201 - 200 400 600 800 1000 1200
Stone Base
4-14.2-8 Design 4-14.2-8 Design
30 ip 48 Kip Tandem Axle Load
.6 .60
z
I
U
i .4
I-
a
IA
0
Ti.9{.I .•.i.40
.2
200 400 600 800 - 1000 1200 0 200 400 600 800 1000 1200
UNWEIGHTED AXLE LOAD ApPLIcATIoNs, 1000'S UNWEIGHTED AXLE LOAD APPLICATIONS, 1000'S
Figure 55. Effect of axle application on depth of rut; typical sections from
special base experiment.
FLEXIBLE PAVEMENT RESEARCH
I
An important element of the serviceability
and the performance of flexible pavements was
D BASE
cracking of the surfacing material. Although
cracks do not in themselves have much effect on
CRUSHED STONE BASE
the ability of the pavement to serve traffic, they
serve as indications that something about the
pavement design is inadequate and that failure
of the pavement is likely to occur at an earlier CEMENT-TREATED BASE
date than would be the case if no cracking ap-
peared. MARCH JUNE SEPT. DEC. MARCH JUNE SEPT.
For purpose of classification, cracking was Figure 56. Relative effect of seasons on rutting.
'4
U)
Ui
I
C-,
z.
I-
a-
Ui
0
I-
A
2 4 6 8 10 12 14 Cr
0
\
4 6 8 10 12
(fl •6
0
Ui
0
.4
S • •SI
LATE SUMMER 1959
0 0 0 0 0
0-0-0-0- 0-0— 0— 0-0-O—.
0 00 0
6 8 10 12 14 16 18 6 8 10 12 14
.8 --
'0
30 Kip Single Axle Load
\ - FALL 1960_
\ I
(I.• . I
S
0
_______
6 8 10 12 14 16 18 6 8 10 12 14
Figure 57. Effect of thickness of bituminous-treated Figure 58. Effect of thickness of cement-treated base
base on depth of rut, on depth of rut.
82 THE AASHO ROAD TEST, REPORT 5
(0 0.6
.8 w FALL 1960
I
12 Kip 6ingle Axie Load I
zo
\\.L...___FALL 1960
0.4
Ui -i------
C
LATE SUMMER I959
I- I
02
959
1
2 20
L. 25 30 35 40 45 50
TANDEM AXLE LOAD, KIPS
2
LATE AMER 1959I —
T7 2D2 + a3D3 + a, )A, L24
w - A0 (a1D1 + a(L1+L2)A2 (28 )
in which
C.
10 12 14 16 18 20
W0 = number of weighted axle appli-
BASE THICKNESS, INCHES cations sustained by the pave-
Figure 59. Effect of thickness of crushed stone base ment before appearance of
on depth of rut. Class 2 cracking;
FLEXIBLE PAVEMENT RESEARCH
Eq. 28 resulted in
log W. = 5.484 + 7.275 log (0.33D1 + 0.10D2 0.6
and 5. Also plotted are points showing the Figure 61. Relationship between axle load and
number of applications at which time Class 2 minimum rut depth in cement-treated base
wedge sections.
cracking was actually observed to occur in the
test sections in these lanes. The fit of Eq. 29 to
the data in the other Road Test lanes was about 0.001 in.) were used and in Eq. 31 when spring
the same. deflections, d8, were used:
Eq. 29 may be used to predict W, for pave-
ments similar to those in the Road Test; how- log W = 7.847 + 5.919109 (0.33D1 + 0.10D,
ever, in case of pavements in service, a some- + 0.08D, + 1) + 2.850 log L2 - 3.204 log
what better prediction of W can be made by a (L1 + L 2) - 1.106 log clj
slightly different approach. At the Road Test
this involved the measurement in each test sec-
tion of creep speed deflection under a 12-kip The mean absolute residual was 0.16 and r2
single axle load in the fall of 1958, shortly after = 0.84.
pavement construction was completed and
again in the spring of 1959, during the period log W = 8.131 + 4.526 log (0.33D1 + 0.10D2
of adverse environmental conditions for flexible + 0.08D3. + 1) + 2.185 log L, - 2.434 log
pavements. It was assumed that the as-con- (L1 + L,) - 1.296 log d3
structed fall deflection, d,, or the spring deflect-
tion, d0, would help in the prediction of W
when used in conjunction with design and load The mean absolute residual was 0.16 and r2
information. The model chosen for analysis = 0.85.
was identical to Eq. 28 except that a term, The improvement in the squared correlation
dA4, was included in the denominator. coefficient, from 0.79 in Eq. 29 to 0.84 and 0.85
Evaluation of the constants in this model re- in Eqs. 30 and 31 means that it is possible to
sulted in Eq. 30 when fall deflections ci, (in explain about 5 percent more of the variation
84 TUE AASHO ROAD TEST, REPORT 5
l000
III, •.•'
--- WHi
__.. . 1
lie
IF
_,y
I • OBSERVED VALUE (ONE TEST SECTION)
I COMPUTED FROM EQUATION
- AREA OF EXTRAPOLATION
__
1111111
10 . .. . 00 1000
WEIGHTED AXLE LOAD APPLICATIONS TO CLASS 2 CRACKING (Wc), I000S
Figure 62. Relationship between pavement design, axle load applications
and Class 2 cracking.
in log W if the deflection (either d, or d) is attained by Eq. 30. The fit of Eq. 31 is very
used than can be explained in terms of load and similar and the other lanes were similar to
design alone. The mean residual of. ' 0.18 for those shown. - To show a smooth curve for the
log W predictions from Eq. 29 indicates that relationship it was necessary to adjust the pre-
observations whose residuals are less than the dicted number of applications by the deflection
mean residual will range from 64 percent to term from the equation. Thus the plots in Fig-
151 percent of the corresponding prediction. ure 63 are .for cracking thickness index, D, on
With the addition• of the deflection term, how- the ordinate and for log W, + 1.106 log df on
ever, the mean residual of 0.16 corresponds to the abscissa. The extrapolated portion of the
the narrower range of. 69 percent to 144 per- curve was computed assuming a 12-kip axle
cent. Although these gains are prObably not load deflection of 0.012 in. (approximately the
great enough to warrant the time and expense minimum deflection to be expected regardless
required for taking deflection measurements, of thickness).
it must be remembered that Road Test con- A bargraph showing the time of appearance
struction was exceptionally uniform. In normal of Class 2 cracking as it occurred in the test
highway construction a greater spread in de- sections of the main factorial experiment is
flection for a given design may be expected and shown' in Figure 64. Most cracking occurred
thus a prediction of W by Eq. 30. or Eq. 31 during periods when the pavement structure
may be considerably more precise than one was in a relatively cold state, slightly above
from Eq. 29 without the deflection term. freezing. The ..peak period for the appearance
Figure 63 shows the . fit to the observation. of Class 2 cracking was in April for both spring
FLEXIBLE PAVEMENT RESEARCH
seasons. Rutting on the other hand occurred flection was considerably greater at each season
more during warm weather. (see Fig. 56, Sec- (spring, summer and fall) in the sections with
tion 2.2.3). gravel and stone base (9 in. thick) than in sec-
Figure 65 shows the number of sections, tions with bituminous- and cement-treated base
cracked and uncracked, for different levels of of the same thickness.
rutting. Among sections having deeper ruts, a The deflection of the sections with 9-in, gravel
greater proportion of the sections also had and stone bases reached a maximUm during the
cracking than was the case among sections spring, decreased during the summer and con-
having shallower ruts. tinued to decrease until late in the fall.
The deflection of the sections with 94n.
2.3 DEFLECTION AS RELATED TO bituminous- and cement-treated bases increased
DESIGN, LOAD, SPEED AND only to a slight degree in the spring; reached a
TEMPERATURE maximum during the summer, and decreased
Relationships were developed between flexible during the fall.
pavement deflection and pavement design, load, The high level of deflection of the sections
vehicle speed and pavement temperature to pro- with stone and gravel bases in the spring was
vide a basis for the deflection vs pavement per- considered due toadverse subsurface conditions
formance studies reported in Section 2.4. at this time; the lower level in the fall, to an
Findings were as follows: improvement in these conditions.
In the main factorial experiment the as- The deflections of the sections with gravel
phaltic concrete surfacing was much more effec- base were somewhat lower than those for sec-
tive, inch for inch, in reducing pavement de- tions with stone base, although the performance
flection, particularly during the spring, than of the stone base was considerably better than
was the base or subbase. that of the gravel base sections (Section 2.3.2).
The subbase was somewhat more effective The deflection occurring within the pavement
than the base in restricting deflection in both structure (surface, base and subbase), as well
=-
the spring and fall (see Section 2.3.1). as that at the top of the embankment soil, was
In the special base experiment the level of de- greater in the spring than during the succeed-
I 11111111 1 I HIlill I
LOG WC+1.106 LOG d 7.847+ 5.919 LOG lD+l)+ 2.850 LOG L2-3.204 LOG IL.+ i2)
Dc .33o,+.IOD2+.O8D
U 11 II
_
I'
11111111 1111111
Milli
11111111
11111
U10
___IIIIIIlIN!IIIiIIUIIIIIHI
los
UIIIIIIIIIIIIIII 11111111 $06 $07 $08
Wcd 06 WEIGHTED APPLICATIONS TO CLASS 2 CRACKING X FALL CREEP SPEED
DEFLECTION (IN 0.001 INCHES) TO 1.106 POWER
Figure 63. Relationship of design to axle load applications at Class 2 cracking and
fall creep speed deflection.
Me THE AASHO ROAD TEST, REPORT 5
ing summer months. This was considered to be periodically throughout the traffic testing phase
due to the higher moisture contents of the base of the Road Test. The principal purposes of
and subbase that ekisted in the spring. this program were (1) to provide information
A high degree of correlation was found to that could be used in the fulfillment of objec-
exist between the deflection at the top of the tive 5 which called for, ". . . test procedures,
embankment and the total deflection (Section data, charts, graphs, and formulas which will
2.3.3). reflect the capabilities of the various test sec-
A pronounced reduction in deflection accom- tions. . . ." (these studies are reported in Sec-
panied an increase in vehicle speed. Increasing tiOn 2.4), and (2) to develop relationships
the speed from 2 to 35 mph reduced the total showing how deflection is influenced by pave-
deflection 38 percent, the embankment deflec- ment design, load, vehicle speed, etc. (these
tion 35 percent, and the partial deflection 67 studies are discussed in this section). Data
percent (see Fig. 78, Section 2.3.4 for definition upon which the deflection studies are based are
of partial deflection). given in Appendix C.
In the studies of the effect of the temperature Taken together, the studies showing how
of the asphaltic concrete surfacing upon deflec- deflection is influenced by design and load and
tion, it was found that between 80 and 120 F the studies showing how performance may be
the deflection was essentiall y constant. At about predicted by deflection, may add to the under-
80 F it began to decrease as the temperature standing of the mechanics of load support of
decreased. As shown by Figures 87 through 91, flexible pavements. This should prove helpful
the extent of the decrease varied, depending to engineers in pavement design work and in
upon such factors as the age and traffic history the evaluation of the capabilities of existing
of the pavement, the speed of the vehicles, the pavements.
design of the pavement, the type of base, and
the time of the year when the tests were made. 2.3.1 Deflection as a Function of Design and
An extensive program was conducted in Load, Factorial Experiments
which deflections of the flexible pavement sec- Deflections of the surface of the pavement
tions under moving loads were measured were measured with the Benkelman beam under
01 0 NO CRACKING
The equations resulting from these analyses CRACKING
were used to determine the relative effect of the
thickness of surfacing, base and subbase courses
on deflection. Although there was considerable
variation among the coefficients from one 1001)
to the next, there was no consistent trend that
indicated the need of including a load-design
interaction term in the across-loop equations.
Consequently, it was assumed that the best esti- 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
mates for a, and a were the means of the
a2, RUT DEPTH,INCHES
values from the individual 1oop equations. Figure 65. Main factorial experiment, relationship
Coefficients from the Loop 2, lane 2 equation between cracking and rut depth (Loops 3, 4, 5
were not included. and 6, all test sections in test).
13
Jill \\
4.
....
................... . . .. ....... S...
c
S --
LA
t
- -
7•
.
j
3V1I iii; I
• y _' -. .\,• S . •
• .,k,:t .
Moderate rutting, no displacement of longitudinal Class 3 cracking.
grid tine.
.200
Spring Defletions Fall Def'lections
0d°125 D1 +.02O 02+028 03
Log do 1.07+1.46 Log L,- .049 D1+.014 02+023 03
.160 ______ 4.42 Log (I+Dd .160 Log d .74+1.13 Log L1 -
3.61 Log (1+04 )
(/)
x
.120 .120
z
0
I.-
.080 .080
04C .040
0 0
0 0.2 0.4 0.6 0.8 1.0 12 14 0 0.2 04 0.6 0.8
Figure 67. Relationship between pavement design and creep speed deflection
(from Road Test equations).
_1 I
.OG d .74+1.13 LOG L1-3.61 LOG (14 Od)
6 Kip Single Axle Load 12 Kip Single Axle Load
1d .O49DI+.OI4D2.O2303
02 0.3 0.4 0.5 0.6 0.7 0.2 0.3 0.4 0.5 0.6 0.7 J.6 0.7 0.8
DEFLECTION THICKNESS INDEX (Dd) DEFLECTION THICKNESS INDEX IDdI DEFLECTION THICKNESS INDEX ID4)
Figure 68. Main factorial experiment, relationship between design and fall creep speed deflection.
2.3.2 Deflection as a Function of Design and The procedures used were the same as those
Load, Base Type Experiment outlined in Section 2.3.1.
Creep speed deflection tests were made every The results of tests made in the fall of 1958
two weeks during the period of traffic operation (before start of test traffic), and in the spring,
on the special base sections (Design 4). Single summer and fall of 1959 are shown in Figures
axle. loads were employed as follows: 70 through 73. The values given are the means
of the deflection in both wheelpaths for two to
four series of tests. The test dates for the
Loop Single Axle Tandem Axle selected periods were as follows:
Lane (kips) Lane (kips)
Avg. Temp.
3 . 6 and 12 6 . Season Test Dates of Surfacing
4 12 and 18 12 (°F)
5 12 and 22.4 12 Fall 1958 Oct. 5; Nov. 8 72
6 . 12 and 30 12 Spring 1959 March 20, 31;
April 16, 27 68
Summer 1959 Aug. 3, 17, 31 96
Measurements were taken in both wheel- Fall 1959 - Oct. 12, 26; Nov. 8 55
paths at the midpoint of each 40-ft subsection.
FLEXIBLE PAVEMENT RESEARCH 91
After examining a number of deflection the lower level in the fall months was due to
histories, these dates were selected to represent improved subsurface conditions.
the associated seasons. However, they do not Asphaltic concrete is more able to support
necessarily ,correspond to the maximum or and distribute load at low temperature than at
minimum values of deflection for each base high (see Section 2.3.1). The relatively low
type. Generally, deflections were greater in the deflections of the bituminous base sections in
spring in the stone and gravel base sections; the spring and the higher values in the summer
they were greater in the summer for the bi- may be attributed to this fact. The relatively
tuminous and cement base sections. low deflections in the fall months may be par-
Curves were constructed by eye through the tially due to the temperature of the bituminous
data of Figures 70 and 73, and the values of de- base and surfacing, and to better subsurface
flection for 9-in, thicknesses of each type of conditions.
base were obtained from the curves. These Although the deflections of the gravel base
values are plotted in Figure 74, which shows sections were somewhat less than those of the
the differences in level of deflection of each base stone base sections, the performance of the
type for the four periods of testing. For each stone base was considerably better than the
period, the level of deflection of the gravel and gravel. Perhaps the gravel possessed less in-
stone base sections was considerably greater ternal stability than the stone; yet, it may have
than that for the bituminous and cement base been somewhat less resilient.
sections.
The deflection of the gravel and stone bases 2.3.3 Deflection at Embankment Level
reached a maximum during the spring, de- In Sections 2.3.1 and 2.3.2, all references have
creased during the summer, and continued to been to deflections of the pavement surface
decrease during the fall; the pattern being measured at creep speed with the Benkelman
much the same as for the factorial sections. beam. Additional studies were conducted in
The deflection of the bituminous-.and 'cement- which deflections were measured' with elec-
treated bases increased only to a slight degree tronic devices at speeds ranging from creep
in the spring; and for some unknown reasons speed to 50 mph. Small linear variable differ-
reached a maximum during the summer, and ential transformers (hereafter designated
decreased during the fall. LVDT) were used as transducers that trans-
The high level of deflection of the stone and lated movements into changes in electric cur-
gravel' sections in the spring was due to adverse rent which were recorded by conventional
subsurface conditions that exist at this time; means. These devices were used to measure
.I80 .160
\l8 Klp Single Axle Load \ 22.4 Kip Single Axle Loo
Load
.120 .121
Dec .080
.040
rI
.4 0.6 02 1.0 4 0.6 02 1.0 1.2 0.8 0.8 1.0 1.2 1.4
DEFLECTION THICKNESS INDEX (Dd ) DEFLECTION THICKNESS INDEX tDd) DEFLECTION THICKNESS INDEX IDd I
Figure 69. Main factorial experiment, relationship between pavement design and
spring creep speed deflection.
Spring 1959
---CRUSHED STONE BASE
-GRAVEL BASE
BITUMINOUS TREATED BASE .080
S
'%% ...•_S_•N._
.061 - .060
FaIl 1958 -
-.--- __.4
z
U
.040 ____________ •
—
______ ______ ______ .040 ———:--
z2 — .-
.020 • .020 —4--
r — --
.4 6 B 10 2 14 2 4 6 B 10 12 14
z
0
.02C
'U
-J
'A-
La
0
C
2 4 6 B 10 12 14 4 6 8 10 12 14
BASE THICKNESS. INCHES BASE THICKNESS. INCHES
Figure 70. Special base type experiment, relationship between base thickness and deflection, 12-kip
single axle load, Loop 3.
Fall 1958
.106
Uj .08C
I
z
U
z
2.060
I-
U
La
-J
LA-
.040
.026
'2 4 6 8 10 12 14 16
bC 100 — I
_ \Summer59 FaIl 1959
080
In
'U
I
U
--CRUSHED STONE BASE
z 'A----- GRAVEL BASE
z
0
.06C 160
CEMENT TREATED BASE
46 -
020 )2C
2 4 6 8 10 12 14 N 6 8 10-1214
BASE THICKNESS. INCHES BASE THICKNESS, INCHES
Figure 71. Special base type experiment, relationship between base thickness and deflection, 18-kip single
- axle load, Loop 4.
.101 .1011
Summer 1959 Fall 959
.080 .081
Ui
I
U
\ A--GRAVEL BASE
-----CEMENT TREATED BASE
1
.060 -\- .060 + BITUMINOUS TREATED BASE
-
0
I
U 'S
'S
.040
.020 .020
( B 8 10 12 14 16 4 6 8 10 12 14 16
BASE THICKNESS. INCHES BASE THICKNESS. INCHES
Figure 72. Special base type experiment, relationship between base thickness and deflection, 22.4.kip
single axle load, Loop 5.
.121J
----CRUSHED STONE BASE Spring 1959
CEMENT TREATED BASE
+ BITUMINOUS TREATED BASE .IOIJ
.080
I
2.06C
Fall 1958
TT 1Ec11
z
U
z
O.04(,
_ --
6 8 10 2 14 lb 18 4
.08 -
Summer 1959 FaIl 1959
Il) .060
Ui
- 160-
I
U
z
.040 40
I-.
U
Ui
-J
L.J.020 120_618
0
6 8 10 12 14 6 lB - - - 14 -
BASE THICKNESS. INCHES . BASE THICKNESS. INCHES
Figure 73. Special base type experiment, relationship between base thickness and deflection, 30-kip single axle
load, Loop 6.
94 THE AASHO ROAD TEST, REPORT 5
deflections at various levels beneath the pave- face and on the top of the subbase. The differ-
ment surface as well as on the surface (see Ap- ence between the deflections referenced to the
pendix D for a description of the test equip- deep plates (total deflection) and those refer-
ment and procedures). enced to the plates on the top of the embank-
Because dynamic deflections could be meas- ment (deflection within the pavement structure
ured in only one test section at a time, nearly only) was the deflection of the surface of the
two weeks were required to obtain a complete embankment. Figure 75 shows relationships
set of readings in all of the instrumented sec- between deflection and total structure thickness
tions. These tests were normally made during for five of the traffic lanes at four periods dur-
scheduled traffic operations under the regular ing the traffic testing phase. The data shown
traffic test vehicles. Except during the winter include manual Beñkelman beam deflections
months, these measurements were taken con- taken at creep speed on the surface, and sur-
tinuously throughout the traffic testing phase. face deflections and embankment deflections
Basic data from these tests may be obtained taken at the regular test traffic speeds of 35 mph
from the Highway Research Board in DS 5141, (no beam deflections were taken under the 32-
which defines deflection as the mean of the five kip tandem axle load). The four periods cor-
highest readings taken during any individual respond to four seasons and the data for each
observation period. Readings were not re- season represents the averages of the biweekly
corded unless the center of the dual wheels deflection measuring series. The selected series
passed within 51/ 2 in. of the center of the LVDT for each season are as follows: spring 1959—
installation. May. 1 and May 19; summer 1959—June 16,
Deflections of the pavement surface were July 2, July 15 and August 12; spring 1960—
referenced to plates installed 6 ft below the top April 8, April 19, and May 4; summer 1960—
of the embankment, on the embankment sur- May 31, June 23, and July 13. The average ab-
solute deviation from the plotted means (dy-
namic tests) for each of these periods was
spring 1959, 0.0035 in.; summer 1959, 0.0029
.080 .080 in.; spring 1960, 0.0034 in.; summer 1960,
0.0033 in. It may be noted that the creep speed
deflections are considerably greater than the
.060 .060 high speed deflections. Data concerning the
lB Kip Single
U, effect of speed on deflection are presented in
UJ
I - Axle Load
U Section 2.3.6.
z
.040 The lines in Figure 75 were constructed
through the data by eye. Furthermore, no at-
tempt was made to separate the effects of the
.020 .020 three individual layers of structure—deflection
Uj
0 is plotted against total thickness (that is,
D = D, + D + D, in the notation of Section
0 2.3.1).
Table 25 gives the total deflection, structure
deflection and embankment deflection of five
STONE .100 axle loads for four periods of time. The values
=• GRAVEL were taken directly from Figure 75 and indi-
= BITUMINOUS cate that deflections (total, structure, and em-
CEMENT
.080 bankment) in general were greater during the
spring than during the succeeding summer
22.4 Kip Single ,gle periods.
Axle Load ood
.060 .060 Of interest in this connection are the data in
U)
Ui Section 2.5.6 (see Table 34) on the condition
I
U
z
of the embankment soil and base and subbase
.040 .040 courses in the spring and summer of 1960.
0 These data, obtained from the trenching pro-
gram, show that for base and subbase mate-
.020 rials the moisture contents were higher in the
spring and for all three materials the CBR val-
ues were higher in summer. In fact, the sum-
mer CBR values of the subbase and base were
0 Cr 0 double those of the spring and those for the em-
mcI, - : Fz
W
U) -U) bankment soil were 40 percent higher. In view
In
of this, it was expected that the densities of the
materials would be greater in the summer.
Figure 74. Special base experiment, seasonal
deflection for 9-in, base thickness. However, this was the case only for the sub-
FLEXIBLE PAVEMENT RESEARCH 95
---CREEP SPEED
.040 DEFLECTIONS
TOTAL DEFLECTION
AT 30 MPH
.020
--------EM8ANKMENT
DEFLECTION AT 30 MPH
C __ 1 __--
-
12 14 0 10 20 12 14 16 lB 20 12 14 lb Is 20
06'C
.040
U
z
.020
I LJTT EFHIT
23 IS Il 19 21 23 15 17 lb 21 23
.041
.02CIj_
C
5 I, 19 JI 23 15 - 17 I921 23I5 17 I 21 23I5 17 -
is 2
.020
'1J LJ
C
is 20
__
22 24 20 18 20 22 24 26 18 20 22 . 24 26 18 ZO 22 24 26
.060
.040
Figure 75. Main factorial experiment, relationship between design and deflection.
96 THE AASHO ROAD TEST, REPORT 5
base (Table 34). The increase in moisture con- Use of total deflection as a measure of the
tent was probably responsible for the higher capabilities of flexible pavements has been sub-
structure deflections that occurred in the spring ject to question, because total deflection alone
months. does not show the sharpness of bend or curva-
Figure 76 includes plots of embankment de- ture of the surfacing. •At the Road Test
flection against total deflection for the dynamic attempts to measure the curvature of the sur-
deflection data obtained in the spring and facing directly with curvature strips (plastic
summer of 1959. A high level of correlation strips with strain gages cemented to their
existed. Extrapolation of the relationships in- surfaces) placed between the layers of asphal-
dicates. that the embankment deflection would tic concrete produced no useful data. However,
approach zero when the total deflection is about a simple statistic ("partial deflection") related
0.008 in. for both the spring and summer to maximum curvature of the surfacing was
period. obtained from the analog traces of deflection.
The partial deflection was simply the depth of
2.3.4 Deflection Basin the deflection basin measured under a 2-ft
Several sidestudies produced the information chord at the bottom of the basin. Figure 78
concerning the configuration of the deflection shows the manner in which partial deflections
basin. The deflections used in these studies were obtained from the deflection traces.
were measured either with the electronic re- Figure 79 shows plots of partial deflection
cording devices (LVDT's) or with Benkelman against total deflection under single axle loads
beams especially equipped to record complete for all six test loops, indicating the degree of
analogs of the deflection traces. correlation between partial and total deflection.
Data on distribution of deflection were de- Within each loop, the ratio of partial deflection
veloped from the traces of deflection by con- to total deflection decreased as the load in-
ventional influence line techniques. These traces creased.
were obtained from placements of the vehicle
wheel to the left, right, and directly over the 2.3.5 Deflection-Load Relationships
deflection measuring point. This made it pos-
sible to develop contours of equal deflection, In Eqs. 38 and 39 in Section 2.3.1, it was
ranging from the maximum at the center of shown that, based on a wide range of structure
the basin to zero around the periphery. Figure designs and loads, the deflections measured at
77 shows typical deflection contours of the creep speed in fall 1958 and spring 1959
pavement surface and of the top of the embank- increased with load with upward curvature
ment under a single and a tandem axle load. (Fig. 80). Specifically, fall deflections, d1, in-
TABLE 25
ESTIMATES OF DYNAMIC DEFLECTION
(from Figure 75)
'At 30 mph
FLEXIBLE PAVEMENT RESEARCH 97
050
creased in proportion to L1113 and spring deflec-
tions, d8, varied with L1146 . Spring 1959
As a further check on the effect of wheel load
.040
on deflection, special studies were made where +
d,,,b —.0071 .855 d,
load was varied and other variables were kept r 2 .833
constant. This program included three series of z
0
tests. The first two were made on the single .030
axle lane of Loop 6, one during the spring and
one during the fall of 1959. The third series
was made on the single axle lane of Loop 4 .020
during the fall of 1959. There were six test sec- Ui
tions in each series. z4
m
The sections in Loop 6 for series 1 and 2 were .010
of the following designs (inches of surfacing,
base and subbase, respectively) : 5-9-12, 5-9-16,
6-6-12, 6-9-12, 6-6-16 and 6-9-16. The six sec-
0
tions in Loop 4 included in the third series of .010 .020 .030 .040 .050 .060
tests had the following designs: 4-6-8, 4-6-12,
5-3-8, 5-6-8, 5-3-12 and 5-6-12. 50
Both single and tandem axle vehicles were
Summer 1959
included in these studies. Axle loads for test 1
-
I
series 1 and 2 were 6, 12, 18, 22.4 and 30 kips (0
ul 40
single axle and 24, 32, 40 and 48 kips tandem U
z dem —.0070 +944 d 1
axle. For series 3 they were 6, 12 and 18 kips r.882
single axle and 24 and 32 kips tandem axle. z0
.0027
.030
Benkelman beam deflections were taken at Ui
eight locations (4 in each wheelpath) in each -J
U.
UI
section using an improvised form of the re- 0
N-TRANSERSE REFERENCE
0 _
25 Iii ___ ___ ___ ___
,.o
;oI
\ '\
1%
so Ip
1
LONGITUOINAL
I_
REFERENCE -60
6 5 4 3 2 I 0 I 2 3 4 5 6 7
1
z0 -TRANSVRSE REFERENCE I—
zs/
-AXLE 21
ss
ZLi 0
0Lt U DIN AL
)
LLONGI7
REFERENCE
\\
ci 40 135
30 /
U
0z
. .' _5 ' "\o3s'•-•••._
\ \
N 1' o
Ix
zu 10
20
4 3 2 I 0 1 2 3 4 5 6 7 8 9
z
0
TRANSVERSE REFERENCE j._-'
RLEI \
s-AXLE 2
10
Is
(
LLONG TUDINAL
REFERENCE
20 S
_
I
I I
4 3 2 0 2 3 4 5 6 7 8 9
DISTANCE FROM TRANSVERSE REFERENCE LINE. FEET
Figure 77. Deflection contours, March 1959.
DIRECTIONOFMOVINGWHEEL
OIO
Cr
- O .010 .020 .030 .040 .00 Le
J .040 -
Figure 79. Main factorial experiment, relationship between partial deflection and total deflection.
100 THE AASHO ROAD TEST, REPORT 5
For structure deflection: shows that all the sections included in the tests
log d1,,. = 0.336 + 0.611 log L1 - 0.534 log L2 were in good condition at the time of study:
(44) Fall Serviceability Value, p
= 0.83, rms error = 0.055, and L1S/L1T Loop Design Thickness
= 0.546. Index' Aug. Sept. Dec.
In only two sections (5-3-12 and 5-6-8 in
Loop 4), and then only in the fall test series, 4 3-6-12 0.51 3.5 3.2 3.5
4 5-6-12 0.61 3.5 3.4 3.5
were the deflection-load relationships curved 4 5-6-4 0.42 3.4 3.3 3.3
upward; i.e., the exponent of L1 was greater 4 5-6-4 0.42 4.0 3.9 3.9
than unity. The others curved downward. The 6 4-9-16 0.69 3.7 3.6 3.7
disagreement between this and the clear up- 6 6-9-8 0.60 2.8 2.8 2.7
6 6-9-8 10.60 3.6 3.4 3.4
ward curvature shown in the studies reported 6 6-9-16 0.79 3.6 3.3 3.2
in Section 2.3.1 was not explained.
Fall thickness indexes were computed from the fall
2.3.6 Deflection-Speed Relationships normal procedure deflection Eq. 33.
Because vehicle speed was not a variable in
the Road Test its effect upon pavement per- The deflections were measured using the elec-
formance was not determined. However, con- tronic devices and recording equipment de-
siderable information associated with this scribed in, Section 2.3.3. They were taken at
question was obtained from special studies of speeds ranging from creep speed to 50 mph
the effect of speed on pavement deflection. using two single axle wheel loads, 6 and 9 kips
The objectives of the tests were to determine on Loop 4 and under 6 and 15 kips on Loop 6.
the effect of load, pavement design and temper- The loads of the test trucks were randomized
ature of the surfacing upon the deflection-speed but all trips necessary for one speed were run
relationships. Three series of tests were made, in sequence before progressing to another speed
the first in August, the second in September, level.
and the third in December 1959. The following The deflection of the pavement and that of
the structure alone (surfacing, base and sub-
base) was recorded on paper tape and for these
records the total deflection and the partial de-
60 flection (see Section 2.3.4) was obtained.
U,
In the analysis of the data, equations were
Fall Creep SpLed DeflLtions
Ui
I LOG dF .74+ 1.13 LOG L1 -3.61 LOG (1+0)
developed for each section included in the tests,
U
for each load used, and for each date tested.
z
0
.120 _Dd.O49DI+.Ot 4D2+.O33D3
-D
I The model selected to fit the observed data was
I-
U
Ui
d = 10A*h, (45a)
.080
0 in which
d = deflection;
= 0.40 v = vehicle speed;
A0 = test section constant; and
A = speed coefficient.
10 IS 20 25 30
60
The regression was performed on the log
transformation
In
Ui Spring ' Creep Speed Deflections
I
1) LOG d, I.07+1.46L00 Li- 4.42 LOG (I+D) / log d = A, + A1v . (45b)
.120
Dd.l25D1+.O2ODZ+.O2$D3
Good agreement was found between observed
t
1
0
I-
0 070 values of deflection and those computed with
U I-D =0.80 the equations. Examples of the level of agree-
Ui I
.080
0 ment for the tests made in September 1959 on
0
Ui
the 5-6-12 design in Loop 4 and on the 6-9-16
Ui
0.
U, design in Loop 6 are shown in Figure 85.
.040
In a study of the speed-deflection data for
the sections under the same load no consistent
or orderly effect of design on the speed coeffi-
25
cient was disclosed. Consequently, equations
0 5 10 15 20 30
SINGLE AXLE LOAD, KIPS
were developed for the means of the four sec-
Figure 80. Main factorial experiment, relationship tions in each ioop. This was done by averaging
between spring and fallcreep speed deflection and the values of A0 and A 1 appearing in the in-
axle load (from Road Test equations). dividual equations. These values, for the total
FLEXIBLE PAVEMENT RESEARCH 101
C.)
z Lis Lis LIS I A./
-.526 .567 533
LIT LIT LIT
yO
g .040 .040 .040
C-)
Ui
71
-J
U-
LIT 71 V
.040
Lis
IT
.564
.040
LIS
LIT
,544
/ 6-6-16 Design
.020
/ 6-9-16 Design
-0
I.,
10 20 30
ON
0 10
I
20
I I
30
O N
0 10
I
20 30
SINGLE AXLE LOAD, KIPS SINGLE AXLE LOAD, KIPS SINGLE AXLE LOAD, KIPS
0 20 40 60 0 20 40 60 0 20 40 60
TANDEM AXLE LOAD, KIPS TANDEM AXLE LOAD, KIPS TANDEM AXLE LOAD, KIPS
Figure 81. Variable load study, Loop 6, effect of axle load on creep speed deflection, spring tests.
deflection, the embankment deflection and the from 0.49 to 0.67. (Loop 4 sections may be
partial deflection, are given in Table 26 for compared with Loop 6 sections only under the
each date of testing and for each of the loads 12-kip axle load). However, Table 26 indicates
used in the two loops. clearly that the speed coefficient (A,) is reduced
All analyses indicate a marked reduction in as load is increased.
deflection with increase in vehicle speed. Since If partial deflection is considered a measure
most Road Test deflections were measured at of curvature of the deflected surface, then the
creep speed (2 mph), and since normal Road data in Table 27 show that there was a much
Test traffic operated at 35 mph it is of interest greater reduction in curvature with speed than
to consider the influence of design, speed and in case of the total or embankment deflection.
surfacing temperature on the reduction in de- If curvature is'considered a better measure of
flections between these speeds. performance of flexible pavements than total
The percentage reduction in deflection as deflection the apparent benefits of increasing
speed is varied from 2 to 36 mph is equal to speed are greater than indicated by the reduc-
100 (1 - 101) where A, is the speed coeffi- tion in total deflection.
cient. Information from Table 26 is repeated in
Table 27 along with computed values of these 2.3.7. Deflection-Temperature Relationships
percentage reductions. These data indicate no 2.3.7.1 Non-Traffic Loop 1.-During traffic
consistent variation of percentage reduction operations creep speed deflections were mea-
with surface temperature between 87 and 40 F, ured weekly on the non-traffic ioop using a 6-kip
and none with pavement structure design over axle load except during the winter when the
the range of average deflection thickness index pavement was frozen. The results of these tests
102 THE AASHO ROAD TEST, REPORT 5
are shown in Figure 86. Each line represents fall of 1958 were relatively low, increased some-
the average deflection (outer and inner wheel- what subsequently, then decreased to a very
paths) of three groups of sections, whose sur- low level when freezing occurred. Spring tests
facing, base and subbase thickness were as showed a high level of deflection after frost left
follows: the pavement; tests during the summer revealed
a gradual decrease. The deflections for the
Group 1 Group 2 Group 3
second fall increased somewhat erratically. De-
flections taken during the second year followed
approximately the same trends as during the
1-0-8 3-0-8 5-0-8 first.
1-0-16 3-0-16 .5-0-16
During the summer months, when the tem-
1-64 3-6-9 5-6-0
5-6-8
perature of the surface was relatively high and
1-6-8 3-6-8
5-6-16
the subsurface conditions good, the difference
1-6-16 3-6-16
in deflection of the 1-, 3- and 5-in, surfacing
groups of sections were of a low order of mag-
Included are those eight sections whose deflec- nitude. These differences increased appreciably
tions were used to determine the seasonal when the surfacing temperature was low and
weighting function described in Section 2.2. the subsurface conditions adverse, as during
Deflections in these sections -are listed in the spring months.
Appendix B. The data plotted in Figure 87 represent an
Figure 86 shows clearly the effect of seasons attempt to smooth out the variations in the level
of the year. Initial deflections taken during the of the observed deflections with seasons by
5-9-16 Design
[J 0. 0,
.060 .060
L 5
=.627 = .545
CIT L
.040 .040 .040
.020 .020
Figure 82. Variable load study, Loop 6, effect of axle load on creep speed deflection, fall tests.
FLEXIBLE PAVEMENT RESEARCH 103
.080 .080
LOG d=.I44+I.118 LOG L1 LOG d=.069+I.096 LOG L, LOG d=.187-f.937 LOG L1
-.942 LOG L2 -.944 LOG L2 7.599 LOG L2
.060 .060
= .55' .642
LIT LIT LIT
I .040: .040
.0 0
.020 .020
plotting the ratios of deflection; that is, the 1- increase as the temperature decreases below 80
to 3-in., the 1- to 5-in, and the 3- to 5-in. sur- F and those for the 1- to 5-in, group increase
facing groups of sections against surfacing tern- at a more rapid rate than those for the 1- to
perature. This was done to obtain an indica- 3-in, or the 3- to 5-in. There is no change mdi-
tion of the effect of temperature alone. cated in the ratio values above 80 F. The ratios
In all three plots, the ratio values begin to never reach a value of one because of the addi-
104 THE AASHO ROAD TEST, REPORT 5
I I
.025
U) .02C
U
I
zo
.015
O , 0
0 I
U D
.010 TOTAL DEFLECTION
-- SINGLE AXLE
o
TANDEM AXLE
.005 -, ---- I STRUCTURE DEFLECTION
SINGLE AXLE
TANDEM AXLE
- I I I
O
0 4 8 12 16 20 24 28 32
SINGLE AXLE LOAD. KIPS
O 8 16 24 32 40 48 56 64
TANDEM AXLE LOAD.KIPS
Figure 84. Variable load study, effect of axle load on creep speed LVDT
deflection, fall tests (6-9-16 design, Loop 6).
.04
OBSERVED VALUE 6-9-16 Design
- COMPUTED FROM EQUATION •
.03 - I I .03 -.-- 30 KIP SINGLE AXLE LOAD
5-6-12 Design
z z
18 KIP SINGLE AXLE LOAD
z .02
iZZ
TABLE 26
CONSTANTS AND COEFFICIENTS OF SPEED, DEFLECTION EQUATIONS'
TABLE 27
REDUCTION IN DEFLECTION AS SPEED VARIES FROM 2 TO 35 MPH
,
Loau emp. Total Defi. Emb. Dell. Partial Defi.
Loo (kips) (°F) A, % Red. A % Red. A, % Red.
thicknesses of surfacing were plotted against facing average 14 in. of base and subbase, as do
temperature. Curves fitted to the plotted points the sections with 4-in, or 5-in, surfacing. The
by eye are shown in Figure 88. tests were made over 24-hr periods on three
From season to season the level of deflections dates—October 2, October 13, and October 31,
changed; however, the change in deflection per 1958—at either two or three levels of surfacing
degree change in temperature was much the temperature. Creep speed normal deflections
same. Above about 80 F the deflection was were taken under a 12-kip single axle load.
practically constant. Below 50 F to about 30 F The results of the tests are shown in Figure
the change, with one exception, was most pro- 89, in which curves a, b and c represent the
nounced. The lone exception was in the group average deflection for each of the three groups
of sections having 5-in, surfacing, tested on of sections having the same thickness of surfac-
December 3, 1959. Before this date, air tem- ing, and curves d are averages of the deflection
peratures as low as 15 F had been recorded of the three groups of sections in each of the
with several minor cycles of freezing and thaw- loops. The same axle load (12 kips) was used
ing. These conditions may have influenced the in all tests. All thicknesses increased from
deflections of the thicker surfaced sections in a Loop 3 to Loop 5 to Loop 6. The deflections
different way than they did those having the were greater for the sections of Loop 3 than
thinner surfaces. If the deflection of the 5-in. for those of Loop 5 and of Loop 5 than for those
sections in the December 3 tests had decreased of Loop 6.
from 50 to 30 F at the rate shown by the dashed The results of these tests indicate that deflec-
curve (Fig. 88), the ratios of deflection of the tion was decreasing at 85 F, whereas on the
group of sections having the different surfacing non-traffic loop deflection did not begin to de-
thicknesses would beof much the same order of crease until the temperature reached 70 to 80 F.
magnitude as those shown in Figure 87. (Data The rate and extent of reduction of deflection
for the non-traffic loop temperature-deflection with temperature to about 45 F appears more
study are in DS 5190 and DS 5191.) pronounced than in the non-traffic loop study,
possibly because of the lack of oxidation of the
2.3.7.2 Traffic Loops.—In addition to the two
asphaltic cement at the time these early tests
complete routine creep-speed deflection cover-
were made, at least to the point where the re-
ages in the fall of 1958 before traffic, a limited
sponse of the asphaltic concrete surface to tem-
program of deflection-temperature tests was
perature changes was more pronounced.
conducted on sections of Loops 3, 5 and 6, both
in the single and tandem axle load lanes, as One series of deflection-temperature tests was
follows: conducted on the special base sections of Loop
6 in October 1960. Measurements were made
Loop 3 Loop 5 Loop 6 extending over a 24-hr period at three levels
of thickness of the bituminous-treated base, 8.6,
Surf. Base Sub. Surf. Base Sub. Surf. Base Sub. 12.4 and 16.1 in., at two levels of the cement-
(in.) (in.) (in.) (in.) (in.) (in.) (in.) (in.) (in.)
treated base, 9.3 and 11.8 in., and at two levels
8 3 3 8 4 3 12 of thickness of the stone base, 13.0 and 17 in.
8* 3 6 4 4 6 8 The surfacing thickness over all three base
4 3 9. 12 4 9 16
0 ————— - types was 4 in., the subbase thickness beneath
the bituminous and cement base, 4 in., and be-
3 0 0 4 3 12 5 6 12
12* neath the stone base, 8 in.
3 3 8 4 6 8 5 6
3 6 4 4 6 8* 5 9 8 The results of the tests are shown in Figure
- - - 4 9 4 5 3 16 90. In all three base types the reduction in
4 0 4 5 3 4 6 3 8 deflection with temperature change was some-
4 3 0 5 6 12 6 6 16 what more pronounced for the thinner cross-
4 6 8 5 9 8 6 9 12 sections. This is in agreement with the test
described previously. Also, the percentage re-
* Replicate section. duction was the greatest in case of the 16.1-in.
In each loop, average base and subbase thick- bituminous base.
ness is the same for each surfacing thickness. In the previously discussed tests, the deflec-
For example, in Loop 5 sections with 3-in. sur- tions were measured with the Benkelman beam
FLEXIBLE PAVEMENT RESEARCH 107
.070 I I I I I I I .i
I
I INCH SURFACING (AVERAGES FOR 1-0-8. 1-0-16, 1-6-0. 1-6-8 AND
1-6-16 DESIGN SECTIONS)
3 INCH SURFACING (AVERAGES FOR 3-0-8.3-0-16, 3-6-0. 3-6-8 AND
.060 - 3-6-16 DESIGN SECTIONS)
Lil A 5 INCH SURFACING (AVERAGES FOR 5-0-8, 5-0-16. 5-6-0, 5-6-8 AND
I 5-6-16 DESIGN SECTIONS)
LI
.050 -
z
O
5.040 /
LU
IL. I
ILL
o ,/
\
.O30 - •-
ILL S.---
LU
0.
In
0
A
_-.A •-..
a.020
U - / / NJ
a AAA A AA A
A
.010 -
I
A11 I I I I I
0
OCT NOV DEC JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV
(958 1959
.070
.060 -
U)
UI
I
LI
.O50 -
O /I
II
.04O
- H
II 1/
II
ILL II
I
I
.030 I I
U I
U III \ I
0.
U) I a-41
11/
I II I $ _• , •-.- -.
°- .020 • . 4 I S. ,/ •- -•/ A_. '•• •_•
Cr
I' • / / A _A _AA__A _•_.
LI I I /
/ I
\ A AA—
_.__ /
A_AA_A._ •
0(0 A
a A_A_ A A A
O il I I I I I I
DEC JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC
959 . 960
Figure 86. Seasonal deflection on non-traffic loop, 6-kip single axle load.
IM THE AASHO ROAD TEST, REPORT 5
at creep speed. One series of tests was made future pavement performance. Deflections in
on a single section of Loop 4 (5-6-12 design) flexible pavements under moving loads proved
where the deflection was recorded electronically to be highly effective for this purpose.
under. an 18-kip single axle load at a vehicle The performance of the flexible pavements
speed of 35 mph (the speed at which test traffic was. predicted with essentially the same preci-
operated). The tests were carried out in one sion from load-deflection data as from load-de-
day in September 1960 when temperatures of sign information.
the surfacing ranged from about 60 to 90 F. Deflections taken during the spring when the
Figure 91 shows the results of a straight line subsurface conditions were adverse gave a bet-
regression analysis. The reduction in deflection ter prediction of pavement life than, those taken
with temperature was of a low order of magni- in the fall.
tude, about 0.0012 in. per 30 deg temperature There was a high degree of correlation be-
change. tween deflection and rutting.
Obviously more data are needed before it can
be said that the effect of temperature on deflec- 2.4.1 Performance-Deflection Relationships
tion is less at moderate than at slow speeds of The principal purpose for the measurement
vehicle travel. and study of deflections under load in flexible
pavement (Section 2.3) was to find relation-
2.4 PREDICTION OF PERFORMANCE ships between deflection measured at a given
FROM DEFLECTION time and the future performance of the pave-
The fifth Road Test objective asked for rela- ment. It was assumed that the deflection of a
tionships that would employ information from given pavement under a particular load would
dynamic measurements in the prediction of serve as a better measure of the pavement's
ability to survive many applications of the load
than knowledge of the pavement structure de-
sign alone. For example, the deflection may be
expected to reflect the strength of the embank-
ment soil and the strength of the surfacing,
- AVERAGE DEFLECTION OF I INCH SURFACING SECTIONS
Figure 87. Deflection-temperature data, Loop 1, 6-kip The deflection data used in the derivation of
single axle load. the performance equations were those obtained
FLEXIBLE PAVEMENT RESEARCH 109
.05C
.045
S 0 MAY2I,1959
JUNE 26,1959
o AUG. 14, 1959
.040 DEC.3,1959— £ OCT. 2, 1959 -
DEC. 3, 959
MAY7.1960'
I),
w X AUG. 19,1960
5.035 • MAY 21, 1959—
z
z
,/" __ _O
JUNJ 26,I959
0 00 6 A
.030 • - — 0---
Ui
-J
U-
Ui
MAY 7, 1960 AUG. 19,1960
.025
Ui XXX £ X X 0 0 00
a. XX ox
Cl)' - 0
a- £ X0X0
UI AIIM 01 a
.020 -0-
0 / 0
L__AUG. 14, 1959
OCT. 2, 1959
.015
.005
30 40 50 60 70 80 90 100 110 120 130
030
I 1
3— Inch Surfacing Thickness
(I)
MAY 21,1959
025 o JUNE 26, 1959
0
z A
A A A A A
z DEC.3,1959
0 A A 66 6
020
UI X DXXD
-j
U
UI D0 D X AUG4I959 _
0 a aa a L___AUG. 19 , 1960
015 MAY
,
UI
a-
C/) OCT. 2, 1959 -
a-
UI
.010
0
.005
30 40 50 60 70 80 90 100 110 120 130
025
j
.
. x xa xx x 0
~EC. 3, 1959 •,_- x
14,1959
AUG. 19,1960
MAY 7, 1960
OCT. 2,1959
(IllS 7 1
30 40 50 60 70 80 90 100 ItO 120 130
SURFACING TEMPERATURE, °F
Figure 88. Deflection-temperature data, Loop 1, 6-kip single axle load.
110 THE AASHO ROAD TEST, REPORT 5
in the fall of 1958 and spring of 1959 (see Sec- serviceability level was set at both p = 2.5 and
tion 2.3.1) under the loads normally assigned p = 1.5.
to the single axle lanes of Loops 3, 4, 5 and 6
and lane 2 of Loop 2. The life data for these. For the fall normal deflections:
relationships in most cases were observed values log W2.5 = 7.98 + 1.72 log L1 - 3.07 log d1
of W at the level of p specified. In several cases,
where the specified level of. p was not attained = 0.47, r = 0.33
62
during the course of the Road Test, estimates
for W were obtained from the performance log W 1.5 = 8.48 + 1.76 log L1 - 3.32 log d1
equations given in Section 2.2.2.
To obtain common coefficients for all loops = 0.39, r = 0.34
the coefficients for log d for the individual loops
were averaged. An adjusted lane mean was For the spring normal deflections:
then determined for each loop. These adjusted log W2 .5 = 9.40 + 1.32 log L1 - 3.25 log d8
means were then regressed on log L1 to obtain
a coefficient for log L, and a constant term A0 . cl = 0.78, r= 0.21
Subsequently an analysis was made across all
loops in which the coefficients A0, A1 and A2 log W1.5 = 10.18 + 1.36 log L1 - 3.64 log d0 ,,
were determined in one step. Very little differ-
ence in the coefficients was found indicating C2 = 0.66, r = 0.24
that significant design-load interactions were
not present. Therefore, in Eqs. 48-51 the coeffi- Correlation indexes and mean residuals for
cients are those obtained from the analysis these equations show that log W predictions are
made across loops. closer to the observations when spring deflec-
Equations were derived for predicting log W,, tions are used rather than fall deflections, and
from log L and log d; both when d = ( fall also that better predictions are made for ap-
normal deflection) and when d = d8 (spring plications to p = 2.5 than to p = 1.5.
normal deflection). For each case the terminal Curves computed from Eqs. 48 and 50 are
.050 .050
I I I
tO Curves "a"- Loop 6 Sections Curves 'c'- Loop 3 Sections
U
I 2 SURFACING a
0
.04C .040
.030 1 SURFACING
5" SURFACING
•
U 4 SURFACING
.020 .020 A7I
-6"SURFACING
j .010
40 50 60 70 80 90 100 40 50 60 70 80 90 100
.050 .050
I I I
Curves ud_ Loop 4verages
I
Curves "b'- Loop 5 Sections
.030
5" SURFACING
0
U LOOP 6
U
.020 .020
a. A
U
U
0
010
40 50 60 70 80 90 100
010
40
jA
50 60
IL
70
-A
80
A T.
90
_
100
SURFACING TEMPERATURE, 'F SURFACING TEMPERATURE. 'F
Figure 89. Deflection-temperature data, traffic loops, 12-kip single axle load.
FLEXIBLE PAVEMENT RESEARCH 111
Bituminous-Treated Base
In .020
w
I 4-(3- 18)-4 Design
.030 ------------
z
0
8.6" BASE
I-
U
- -
.020
U. 2.4" BASE
Ui
0
0
Ui 0
Ui
CL .010 .005
- 16.1" BASE 1
a
0 -
55 60 65 70 75 80 85 90 95
50 ,55 60 65 70 75 80 SURFACING TEMPERATURE. F
Cement-Treated Base
4- (3-I3)- 4 Design by the end of test traffic. These sections are
U, .050 represented by solid points if terminal service-
Ui
I
U
ability was 1.5 before the end of the test and
1 by squares if final serviceability was between
9 3 BASE 2.5 and 1.5. Sections represented by open
.040
circles had serviceability indexes greater than
2.5 at the end of the test, and are plotted at
points estimated from the performance equa-
tions (see Section 2.2) for log W, 5. Summary
plots of Eqs. 48 through 51 are shown in Fig-
---• r- L8"BASE
ure 94.
.020 Because normal deflections could not be ob-
U tained with the Benkelman beam under tandem
axle loads, Eqs. 48 through 51 must be used
.010 in conjunction with single-tandem relationships
(see Section 2.2.2) in order to predict life under
tandem axle loads.
0 Mean residuals for log W are about the same
50 55 60 65 70 75 80 whether log W is predicted from the perform-
ance equation with given pavement design and
.040. I
load or predicted from the equations involving
load and spring deflections. Therefore, under
Crushed Stone Base the conditions of the Road Test, flexible pave-
U,
4 - (3-19)-8 Design ment performance was predicted with essen-
.030 tially the same precision from load-spring de-
13 BASE
flection information as from load-design infor-
mation. In actual highway practice it may be
BASE
that the load-deflection equations are better
predictors since they can reflect non-uniformity
of construction as well as differences between
Road Test materials and those used in the
actual construction.
The relationships in Figure 94 might be used
to indicate the magnitude of deflection, meas-
50 55 60 65 70 75 80 ured in the fall or spring, that could be con-
SURFACING TEMPERATURE. .F sidered "safe" for any specified number of load
Figure 90. Deflection-temperature data, special base applications before p = 2.5 or p = 1.5. If, for
type wedge sections, 30-kip single axle load. example, it is assumed that there is little risk
I I
\OILOG _Wg_9.40. 1.32 LOG L1 -3.25 LOG dl l8 Kip' Single' Axle Load
\
.I • 5
6 Kip Single Axle Load
S ••
0
0
-I
In
N
70
.020 .040 .060 .080 .100 .120
S I 10
.020 .040 .060 .080 .100 .120 .140 224 Kip Single Axle Load
01
-1
0
w
I-
x 12 Kip Single Axle Load
0
w S
U. S
0
0
0
-j I 0 .020 .040 .060 . .080 .100 .120
2
Nr
UJ7
0
___
.020 .040 .060
•
.080 .100 .120 .140
\
1
0
30 Kip Single Axle L
CREEP SPEED DEFLECTION • INCHES
OBSERVED- FINAL p LESS THAN 1.5
-J 6 -o-'.--
1% #
O OBSERVED- FINAL p MORE THAN 15. BUT LESS THAN 2.5
o ESTIMATED- FINAL p MORE THAN 2.5
.
I. 1- S
41 .1 I
.020 .040 .060 .080 .100 .120 .140
CREEP SPEED DEFLECTION • INCHES
Figure 92. Main factorial experiment, relationship between axle load application at p = 2.5 and spring 1959
creep speed deflections.
that a pavement life will fall much more than regression lines and equations are shown (Fig.
two mean residuals below the curves, then safe 95) along with plotted values of the observed
curves would be parallel to those of Figure 94 at data. Also shown are the correlation coefficient
a distance of about 0.50 units below the curves and the standard error of estimate for each
as shown. For example, this rule would deter- case. The dotted lines are located at a distance
mine that a spring deflection of 0.025 is safe of one standard error of estimate from each
for a pavement that is expected to carry one regression line.
million 18-kip axle loads without dropping
below p = 2.5. 2.5 AUXILIARY STUDIES
2.5.1. Overlays
2.4.2 Rutting-Deflection Relationships A study of the effectiveness of asphaltic con-
Because serviceability, and therefore pave- crete overlays included 99 flexible pavement
ment performance, was affected by the degree test sections. It was clear that overlays were
of rutting and because a high level of correla- highly effective as a means for extending the
tion was found to exist between deflection and service life of these pavements.
performance, deflection was expected to cor- Attempts at mathematical analysis designed
relate with rutting. Figure 95 shows this to be to establish specific relationships between per-
the case. formance and overlay design were unsuccess-
The analyses involved the spring normal de- ful, because the outcome of each analysis
flections (see Section 2.4.1) and the mean of proved to be highly dependent on the assump-
the rut depths in both wheelpaths of the main tions made concerning the mathematical model
factorial sections in lane 1 of Loops 3, 4, 5 and for the analysis. Further work will be at-
6 and in lane 2 of Loop 2. Regression analyses tempted by the Highway Research Board.
were made of data obtained at 140,000 and In the flexible pavement factorial experi-
610,000 axle load applications and at the end of ment, 83 of the 288 sections failed early in the
traffic testing (1,114,000 applications). The test. Each was rebuilt to a thickness considered
FLEXIBLE PAVEMENT RESEARCH 113
adequate to carry the remaining test traffic. sign of each section; (b) thickness, date placed
Their subsequent performance was no longer and type of overlay and percent asphalt from
observed. In another group of 129 sections, extraction tests; (c) data from Marshall tests
structural deterioration occurred at a much on the overlay material; (d) weighted axle load
slower rate. Many sections in this category application sustained before and after overlay-
were overlaid, and observations and studies of ing; (f) rut depths before overlaying and on
their performance were continued in the same final observation date; (g) cracking and patch-
manner as for in-test sections. Thus, for each ing; (h) level of serviceability trend before
of these sections the loss in serviceability with and after overlaying and on the final observa-
load applications could be determined from the tion date; and (i) deflections.
time it was overlaid until further maintenance Not all of the sections had reached a service-
was necessary, or until the end of the traffic ability level of 1.5 when they were overlaid.
test. It was intended to use trends in loss of Some were overlaid when one wheelpath only
serviceability of these sections to develop in- (usually the outer) had attained this level.
formation regarding the effectiveness of the Generally, these were sections having such
overlays. severe distress in one wheelpath that they were
The study included 5 sections from Loop 2, considered unsafe for the test vehicles at their
12 from Loop 3, 23 from Loop 4, 30 from Loop normal speeds, despite over-all serviceabilities
5 and 29 from Loop 6. Basic data obtained from over 1.5. In some sections, the rate of deteriora-
the study (Table 28) include (a) original de- tion of serviceability was so rapid that they
1 I I I
-J
N
a
0 .020 .040 .060 .080 .100 .120 0 .020 .040 .060 .080 .100 .120
7
••
%_ _
L
4
o .020 040 .060 .080 .100 .12
were not actually overlaid until their condition plications of axle loads that the sections, aver-
was below the 1.5 level. In many cases, this was aged by lanes, withstood before and after over-
because of the established policy under which laying, together with average values of the
the placement of overlays was often delayed initial and final overlay serviceabilities. The
until two or more sections had fallen, below the overlaid sections sustained an average of 671,-
1.5-serviceability level. (It is important to 000 weighted load applications, with an average
keep in mind that the analysis of- performance loss in serviceability of 0.5 (3.4 initial, 2.9
of the sections in the main experiment was final). Before overlaying, the same sections had
based on the entire serviceability trend of each sustained a lower average number of weighted
section, and that after a downward trend was applications (348,000) with a higher loss in
established for any section, the actual level of serviceability, 2.7 (4.2 original, 1.5 final). For
serviceability at which the section was removed all the sections, the average depth of rut before
from test had little bearing on the results.) overlaying was 0.69 in. From the time of over-
Moreover, the data (Table 28) show that the laying until the end of test traffic the ruts de-
level of serviceability of the overlays after veloped averaged 0.38 in. Although there are
placement was not as high as the average (4.2) appreciable differences in the values for the
of the sections when originally constructed. individual lanes and those for all lanes com-
This was not surprising in view of the short bined, the over-all trends were much the same.
lengths (100 ft) of the individual sections, and Several attempts were made at the Road Test
the necessity for placing many of the overlays to analyze mathematically the performance
at low temperature and during inclement of the flexible pavement overlays. In these
weather. analyses the overlay was considered to be a
The data in Table 28 have been further sum- fourth layer in the pavement system. It was
marized in Table 29, which gives weighted ap- assumed that the following model, similar to
_\ -- _FaII_958_DifIectIojs 8
30 KIP SINGLE
4
- 6KIP SINGLE-'
6 KIP SINGLE
.020 .040 .060 .080 .100 .120 .140 .160 0 .020 .040 .060 .080 .100 .120 .140 .160
I ~
aIJI958JetIectiJns - - _Sprng_I959_DefIJctions_
p=25 2.5
72 ' Log L 17 307 L og Log 5 9.40+ 1.32 Log L 1 ' 325 Log d}
K
KIPSLE_
~-30 KIP SINGLE
OP SINGL E
12 KIP SINGLE I 4
12 KIP IN
6 KIP SINGLE 6 KIP SINGLE
.020 .040 .060 .080 .100 .120 .140 .160 0 .020 .040 .060 .080 .100 .120 .140 J60
CREEP SPEED DEFLECTiON, INCHES CREEP SPEED DEFLECTION, INCHES
Figure 94. Main factorial experiment, relationship between performance and creep speed deflection (from Road
Test equations).
FLEXIBLE PAVEMENT RESEARCH 115
that reported in Section 2.2 for performance, of the pavements in the non-traffic loop over
would apply: the 2-yr period of the test.
2.5.2.1 Strength and Condition Data, Non-
Traffic Loop.—No traffic was permitted to op-
= /3' (log W' - log p') (52)
in which .100
(140,000 Axle Lcad Appliations)
G' = a function of serviceability loss in I I
4.2—p I
I
the overlaid pavement, log ;
.oeo -d,0252 + .0638 RD I
2.7 r2 .80
U rmsr .0069 -
W' = weighted axle application on the
overlaid pavement; ZF
0
o
.060 40 o ___
and p' = functions of design and load O1()
'Overlay Rolling 'G-kip axle load, Loop 2; 12-kip axle load, Loops 3, 4, 5, 6.
'Initial overlay serviceability lower than final value; for analysis these sections were assigned initial
Breakdown Intermediate Final serviceability equal to final serviceability plus 0.1.
Thick- 'Observations terminated prior to end of traffic due to maintenance; serviceability assumed to be 2.0.
Type neaa Method wt. Wt. Wt. 'Scratch coat-a level course used to fill ruts and depressions before placing first lift. This course placed
(in.) Roller Roller Roller either by hand or asphalt paver without vibrating screed in operation. In general, this course was com-
( tona) (tona)
pacted with 5-ton tandem roller.
I 3 Scratch coat; 2 lifts 10 3-wheel steel 12-15 Pneumatic 8 Tandem
lB 3 Scratch coat; 2 lifts 10 Tandem 19 100-psi pneu. 8 Tandem
II 2 Scratch coat; 1 lift 5 Tandem - - S Tandem
111 3 Scratch coat; 2 lifts 10 Tandem - - 8 Tandem
118 THE AASHO ROAD TEST, REPORT 5
spring and summer periods. This finding was crease from spring to fall at a slower rate. This
identical to that for these materials in the is evidence that some traffic during the spring
traffic ioops, and the increases in CBR and elas- and early summer will hasten the restoration
tic modulus from spring to summer are at- of pavement strength following the conditions
tributed to moisture changes rather than to associated with spring thaw.
density changes (see Section 2.5.6)., However, Data obtained from the trench studies on
in case of the subbase and base materials, the Loop 1. were used to develop correlations be-
increases in indicated strength are not as pro- tween Benkelman beam deflection and the re-
nounced as those for the traffic ioops. suits of the plate load tests. A sample of these
Benkelman beam deflection data provided correlations is shown in Figure 99. Within the
further evidence that flexible pavements under range of the test data the relationships are
traffic possess higher strengths and recover practically linear for 12- and 18-in, diameter
strength more quickly than pavements without - plates and slightly, curvilinear for the 24-in.
traffic. For equivalent designs, beam deflec- plate. In these tests the beam deflection values
tions on no-traffic pavement are higher and de- were obtained under a 6-kip axle load, and
7 142 I I I I I I I I
I I I I I I I
z
I-
U,
z
I-
138
85 U,
U, 'S 0
I-.
• 0
0
S 136
I I I I I I I I 134 I I I I I I I I I I I I I I I I
3 APR JULY OCT IJAN APR JULY OCT
APR JULY OCT
"Dv
959 1960
9 140
Subbase Subbase
8 138
.136
>-
I-
z
U,.
U,
- S
\• 0
- \ >. 134
\ 0
S
\ . S •
5 - 132
-
I II I I I I I II II I I I I
4 130
APR JULY OCT IJAN APR JULY OCT
1960 - 1959 1960
18 I I I I I
I
Embankment Embankment
- _•_
-S
- •
\ ,
./ /
//
5
— S
/
.\
—
I I I I I I I I I I I I I
APR JULY OCT APR JULY OCT APR JULY OCT
1959 1960 1960
Figure 96. Seasonal subsurface conditions; Loop 1.
120 THE AASHO ROAD TEST, REPORT 5
50 700
z
BASE COURSE I-, . . .
40 vj 600
-J
BASE COURSE
I'J /
U)
-J
>30 ,,5O0
3, .,
U 0 18 INCH PLATE /
0
20
In I I I I I I I ---------------------
APR JULY OCT
Ibpoif IbU WbU
41
I I I I I I I I I I 350
I
- SUBBASE - SUBBASE
30 - -
S —
/
V
V -
I
/ / V
20—
. S
/ ,250
/
/
/
/ 9
/5
24
/ INCH PLATE
5,I'
-
/ / I
0a //
I. /
10 1200 I
,
C I I I I I I I I I I I 150 I I II I I I II I II II II I II
APR JULY OCT JAN APR JULY OCT APR JULY OCT JAN APR JULY OCT
1959 1960 959 960
4
2 130
I I I I I I
I I I I I I I I I I I
UD I I I I I I I I I I I I I I I I_ -
• EMBANKMENT U, ,V EMBANKMENT /
,• -J / /'
w3 -'-- —II - 110 /S,// •
, 5— I /
-J
SI /
ID /
/
II /,. 30 INCH PLATE
S —I
I I /
/
70 I I I I I I I I I I I 'I I 1 I I I I
APR JULY OCT JAN APR JULY OCT
IbU 1959 1960
Figure 97. Seasonal strength tests, Loop 1.
the plate values are for a 16-psi unit load. neers, those developed in connection with the
Other investigators have found essentially the Loop 1 studies are given in Appendix H.
same correlation between the results of beam 2.5.2.2 Vertical Volume Changes.—Instru-
and plate load tests on flexible pavement mentation installed in selected sections of the
surfaces.* non-traffic loop (Loop 1) produced information
A considerable amount of effort was ex- concerning the seasonal changes in elevation
pended in developing relationships between the at the pavement surface at the top of the em-
conditions of the embankment soil (moisture bankment soil, and at 4 ft below the top of the
content, density and percent saturation) and pavement. (Relevant information from the
the strength of the material (CBR and plate, traffic loops is reported at the end of Section
load tests). Since relationships of this type are 2.2.3.1.)
of interest to many pavement research engi- Table 30 summarizes the data obtained from
sections of the following designs: 5-0-0, 5-6-0,
* Sebastyan, C. Y., "The Benkelman Beam Deflections 5-0-16 and 5-6-16. Figure 100 shows the
as a Measure of Pavement Strength.'? Proceedings, changes in elevation that were found at the
Canadian Good Road Association, 1960. three levels in the winter and fall of 1959 and
FLEXIBLE PAVEMENT RESEARCH 121
1960. The values plotted for the five sections all the movements at the pavement surface. In
represent the changes in elevation that occurred the remaining three sections the movements at
from the fall of 1958 when the initial readings the pavement surface were accounted for by
were taken. those at the embankment level plus those within
In the 5-0-0 section the ye rtical movements the granular layers.
of the surface (up in both winters; down in Of special interest is the fact that after dis-
both falls) wen practically tI ie same in degree appearance of frost, the pavement sections
as those of the embankment surface. At the settled appreciably (in most cases to a level
4-ft level, howe ver, the moveinents were in the below their original elevation even though they
reverse order. In the o ,her three sections, the were not subjected to traffic).
movements of the paveniient surface were simi- 2.5.2.3 Temperctture.—In two sections of
lar, that is, up in the v inter and down in the Loop 1 a sufficient number of thermocouples
fall periods; however, 1the heave at the pave- were installed to make it possible to study the
ment surface was appi eciably greater in the distribution of temperature in the pavement
5-6-0 section than in the other two. At the em- structure and underlying embankment soil.
bankment level the mo1vements followed those Figure 101 shows the data for March 25 and
of the surface, except thLat in the second winter May 25, 1960. When the temperature of the
the movements were do wnward in case of the surfacing was at the minimum level on either
5-0-16 and the 5-6-16 sel tions. At the 4-ft level date, the temperature of the underlying granu-
below the pavement s urface the movements in lar material was at a higher level; when the
these three sections were of a low order of temperature of the surfacing was at its maxi-
magnitude, excep.t that in the second ;rinter mum level, the temperature of the material
they were downward in the 5-0-16 and the beneath was at a lower level. In all cases the
5-6-16 sections to about the same degree as the "penetration" of a given temperature beneath
top of the embankment. In all four sections no the surfacing was appreciably different from
heaving due to frOst wa found at the 4-ft level. that beneath the crushed stone shoulders. When
As expected, the mov ements at the embank- the temperature of the surfacing was at its
ment surface of the 5-0-.0 section accounted for maximum level on either day, heat penetrated
Base Course
SPRING
[] SUMMER
20
I-
z
U,
I-
z
8 10 00
20 800
U,
z
Ui 3z
I-.
z
8 400
U,
cr
VA.l
p0
1959 1960 MEAN 1959 1960 MEAN
Embankment
40 800
U,
D.
20 400
.050
N 0) CC IN IN
o o o 6 o o o PLATE
++++ I .040
(Si
I8" PLATE
I
00 /
' U) IN 'N CC ,1 0 -
0 0 0 0 0 0 CO IN z -, .030
2w
7/o I
24' PLATE
I I I I I
.020
UJ 0/
IN IN IN ' ,1 IN 2!2
oo0 . o o. .010
d o o o o o
+1111 II
0
0 .010 .020 .030 .040 .050 .060 .070
CO 'CC 00 N U) 0 CO ELASTIC DE0RMATI0N,INCHES
0 (16 p.1 LOAD)
+ I +++++++
Figure 99. Relationship between Benkelman beam and
plate load deflections.
CO CC N IN 0 0 0
0 0 0 0 -1 'i - C' IN
dd00000
I •••I-• I I I I
into the granular material -and embankment
soil to a greater depth than beneath the should-
CO 00 00 CC IN 0 0 ' ers. Conversely, when the temperature of the
d d o 0 o o o o surfacing was at the minimum level, the tem-
-•1•-• -I-- I I I I I I perature at a given depth was greater beneath
the shoulders than beneath the surfacing.
IN
CO 0.1 . CC CO 00 CC IN
' 2.5.2.4 Serviceability Changes, Non-Traffic
Loop.—Pavements in Loop 1 were not subjected
++++++++ I to the test traffic. Over the 2-yr test period no
significant loss in serviceability was noted in
these sections.
0 0) 0) 0) CC N N C)) 0
00 • IN IN
0 0 0 0 0 0
IN• One of the purposes of Loop 1 was to deter-
++ I I I ++ I mine the performance of pavements subjected
to the effects of weather and time in the absence
of traffic. Consequently, lane 1 of this loop was
IN CO CC CC '-C -4 N N C)) kept free of traffic during the entire test period.
0 0 0 0 0 0 0 0 Because the flexible pavement sections inthis
++ LI I ++ I I lane were extremely short (25 ft) it was dif-
ficult to ascertain their serviceabilities without
relatively large measurement error. Small lo-
in CO ' CO
-CO0O4
CO N N
Cl! CO Cl! CO CO calized irregularities played inordinately large
0 .•
I -•l-• I I I -I-- -l--• I I parts in the serviceability determination. In
longer sections such irregularities were "aver-
aged out."
a) IN CC a) cq 0 ' Co 00 Another limitation of this experiment was
N 0.1 0
ddddddod its short duration. Obviously., the probable ef-
+++++++++ fects of environment and time on serviceability
over the normal life span of a pavement could
0 a) a) a) CC N N C)) 0
not be estimated satisfactorily from a 2-yr
0oo
dd00000
Cl! C' study. Nonetheless, a direct comparison was
++ I I I ++ I I available between pavements subjected to traffic
and those in Loop 1 that were not. Even though
only two years had elapsed, several cycles of
freezing and thawing occurred, and the normal
central Illinois precipitation and temperature
ranges were encountered.
00 a) a) Lo a) 0 0
C)) CC C)) 1)) CC CC CC
a) C) a) a) a) a) a) a) The serviceability trend values for the sec-
tions in Loop 1 are included along with the
a) a) CO •CO C. v other performance data in Appendix A. Ex-
cluding Section 857, wrhich was inadvertently
FLEXIBLE PAVEMENT RESEARCH 123
damaged, the serviceability trend values for the given in Appendix D. The basic data obtained
remaining 23 sections in the study averaged on in these tests are given in DS 5162 and 5163.
index days 11, 22, 33, 44 and 55 were 3.47, One special study was made to determine the
3.46, 3.37, 3.45 and 3.44. The differences among effect of vehicle speed on transmitted pressure
these values are well below the magnitude of to the surface of the embankment. The tests
experimental error associated with the service- were conducted simultaneously with those made
ability determining sytem. Thus, it is con- on December 2, 1959, to determine the effect of
cluded that no significant serviceability lOss was speed on deflection (see Section 2.3). Included
found in the flexible pavements of Loop 1 that were four sections of Loop 4 (3-6-12, 5-6-12,
were not subjected to traffic over the 2-yr and the 5-6-4 section and its replicate) tested
period of the Road Test. under four single axle loads (2, 6, 12 and 18
kips).
2.5.3 Embankment Pressure Variations in embankment pressure with
In connection with the problem of the me- seasons of the year are shown in Figure 102.
chanics of load support of flexible pavements, The pressure reached a maximum value during
pressures that are transmitted through the the spring and early summer and decreased
structure to the supporting soil were subjected subsequently. The pattern of pressure varia-
to a limited study. In LOop 4, 20 pressure cells tion with seasons was similar to that observed
were installed in certain sections. for creep speed deflections.
The variations in transmitted pressure with The effect of design (total thickness) of the
seasons and speed were of the same character pavement on transmitted pressure is shown in
as those shown for creep speed deflections. Figure 103. A fairly orderly effect is evident
The transmitted pressures were found to for the spring and summer periods. For the
vary approximately as a linear function of the fall, the 3-6-12 (21-in, pavement) appears to
applied load in the same manner as deflections. be out-of-line. No explanation was found.
In the routine program of pressure distribu- The pressure contours at the embankment
tion tests, observations of the pressure trans- level (Fig. 104) also show the effect of design
mitted to the surface of the embankment were of the pavement. The contours for one wheel
taken weekly except during the winter. Details of the tandem axle load show a slight effect of
of the test procedures and equipment used are the adjacent second wheel.
U I—I U
ji 1•1 t,-j
lop of Pavement
lop of Embankment
Figure 100. Vertical movement data, Loop 1 (elevation changes from initial elevation).
Maximum Surfacing Temperature
EMBANKMENT
66
N \ rL 64
Figure 101. Isothermal lines, 3-0-16 design, Loop 1, March 25, 1960 (left) and May 25, 1960 (right).
FLEXIBLE PAVEMENT RESEARCH 125
in which
P = transmitted pressure in psi; and
fI.
v = vehicle speed, in mph.
!
The coefficients determined for the four sec- I
tions are given in Table 31.
Examples of the analytical results are given 5-6-12 DESIN'
in Figure 105. Embankment pressures are plot-
ted as a function of vehicle speed. The points
represent observed values, and the dashed lines
represent computed values. The marked effect
of load and speed is clearly shown by these
data. 959 IQRO
Table 32 gives computed values of the co-
efficient A, and the percent reduction in de-
flection and pressure for two loads, 12 and 18 Inner Wheelpath-Loop 4
kips, and two designs of pavement. The per- 3-6-12 Dasian
centage values are for a speed range from 2 to
35 mph.
There was close agreement between the per- IV 'I \f
cent reduction of deflection and pressure with a8 I ••\'NV \ A i' \1
UI
8KIP SINGLE AXLE LOA
speed for the 5-6-12 design, but the reduction UI
Ui S
in transmitted pressure was much less than that CL IS
for the deflection for the 3-6-12 design. lI,\I
I I\Ihi 1 \...
In Figure 106, embankment pressure is 2 ! :
z -'/ ._ \IkI
plotted against wheel load. It appears that the <4
I
'4 14 14
10 10
0 0
8 8
5-0-12 DESIGN
1
3-0-12 DESIGN I a
WITH 3" OVERLAY
3-6-12 DESIGN 0
6 6
5-6-12 DESIGN
A A
16 lB 20 22 24 16 18 20 22 24 16 18 20 22 24
STRUCTURE THICKNESS.INCHES STRUCTURE THICKNESS, INCHES STRUCTURE THICKNESS,INCHES
Figure 103. Relationship between design and embankment pressure under 18-kip single axle load.
5-6-12 Design - 18 Kip Single Axle Load 3-6-12 Design- 18 Kip Single Axle Load 3-6-12 Design - 32 Kip Tandem Axle Load
Inner Wheelpath - Inner Wheelpath Inner Wheelpath
it of Pavement of Pavement of Pavement
-
10.4
Pgo
TRANSVERSE SCALE,FEET
0 I 2
I —
TABLE 32
PERCENT REDUCTION IN DEFLECTION WITH SPEED
5-6-12 3-6-12
I I
-
3-6-12 Design Inner Wheelpath
18 KIP SINGLE AXLE LOAD
December 2, 1959
2 KIP SINGLE AXLE LOAD 0
0 6 KIP SINGLE AXLE LOAD
2KIP SINGLE AXLE LOAD 0 00
.r-0•• 0
0 0
0 0
o 00
0
0• 0
00
0.
—o
0 0
01 0
—.0
000 0•..000 CIU-
0
0
0
••_•_.._..• •
:
0
0
•-
0
- T__..._•_ - - 116—
00 10 20 30 40 50 60 00 10 20 30 40 50 60
VEHICLE SPEED mph VEHICLE SPEED. mph
elastic subgrade moduli from 30-in, diameter the main traffic loops, relationships, similar to
plate load tests on the embankment. These data those foi Loop 1, were developed (Appendix H).
are presented in detail in Table 34 and in sum-
mary form in Figures 108, 109 and 110. The 2.5.7 Bituminous Surface Treatment
strength (plate load and CBR tests) of the A limited program of tests involving surface
granular layers and of the embankment soil was treatments was conducted on Loop 2 as a part
greater in the summer than in the spring. of the over-all study of pavement design. The
Furthermore, the elastic subgrade modulus of performance of the surface-treated sections was
the embankment soil in both of the periods was inferior to that of sections with equal thick-
somewhat higher beneath the thick than the nesses of base and subbase but with asphaltic
thin pavement sections. concrete surfaces. Their performance was im-
This increase in strength or stability of the proved appreciably as the thickness of the un-
embankment material is not explained by in- derlying base and subbase course was in-
creases in density or by decreases in moisture creased.
content as these changes were neither orderly The bituminous surface treatment experi-
nor consistent. However, the increase in ment was included in Loop 2 as a part of the
strength may have resulted from a drying study of pavement design. The experiment
cycle during the summer and a wetting cycle (Design 6) was a 3 by 2 factorial experiment
during the spring.* The marked increase in the with base and subbase thickness as the prin-
indicated strength of the base and subbase ma- cipal variables. The sections were replicated
terials from spring to summer was apparently so that there were 12 sections per lane. Sur-
caused by decreases in moisture content because facing for all sections consisted of two applica-
changes in density were small and inconsistent. tions of about 1/4 gal per sq yd of MC-5 bitumi-
In the study of the physical test data of the nous material, each covered by crushed stone
embankment soil obtained from the trenches in aggregate (graded from /8 in. down), and a
similar application of MC-5 with a smaller (/8
* "Pore Pressure and Suction in Soils." Conference, in. down) stone seal coat (see Road Test Report
British National Society of the International Society of 2).
Soil Mechanics and Foundation Engineering, William
Clowes and Sons Publishers, London, England. Basic information on the performance of the
FLEXIBLE PAVEMENT RESEARCH 129
sections in this experiment is given in Table 35 structures. Also, no paving machine was used
and in Appendix A. The values for each sec- to iron out minor irregularities.
tion are weighted and unweighted load applica- If the performances of sections in this experi-
tions to a serviceability of 1.0 or the final serv- ment and those in the factorial test are to be
iceability if the section was still in test at the compared, the comparison should be on a basis
end of test traffic. The applications are for a of an equal drop in serviceability rather than
lower level of serviceability (1.0) than was a drop to the same level of p. In the highway
used in the factorial experiment (1.5 and 2.5). system, surface-treated pavements are usually
This terminal serviceability value was chosen used for relatively low-traffic highways; there-
because the initial serviceabilities of the sur- fore, these pavements may have lower service-
face treatment sections were considerably lower abilities before major maintenance than high-
than those for sections with hot-mix asphalt traffic volume highway pavements.
concrete surface (2.2 compared to 4.2, on the The performance of the surface-treated sec-
average). The low initial serviceabilities of tions was appreciably improved as the base
these sections may be attributed to the impossi- thickness was increased in sections with and
bility of maintaining smooth grades during without subbase for both the 2- and 6-kip axle
construction of these short sections with thin load lanes (Table 35).
3-6-12 Design
December 2,1959
5
U)
-
U)
w
50 mph
0
0 2 4 6 8 10 12 14 16 18
4.
5-6-12 Design
a
December 2,1959
5m
50 mph
0<;
2 4 6 8 10 12 14 16 is
SINGLE AXLE LOAD ,KIPS
Figure 106. Relationship between embankment pressure and axle load (from Fig. 105).
130 THE AASHO ROAD TEST, REPORT 5
2.6 SUMMARY OF FINDINGS AND ment research may be found in Section 3.6.
NEEDED RESEARCH The principal objectives of the research at
the AASHO Road Test, excluding those dealing
2.6.1 Summary of Findings with bridge studies and with pavement mainte-
The major findings of the flexible pavement nance, were:
research at the AASHO Road Test have been 1. To determine the significant relationships
stated in the text and illustrated by means of between the number of repetitions of specified
charts and graphs. Summarization of the axle loads of different magnitude and arrange-
material contained in each major subsection ment and the performance of different thicknesses
of uniformly designed and constructed asphaltic
precedes the text of the subsection. Here the concrete, plain portland cement concrete, and
more important material contained in the sum- reinforced portland cement concrete surfaces on
marizations is repeated for the convenience of different thicknesses of bases and subbases when
those who do not wish to study the details of on a basement soil of knowii characteristics.
* * *
the development of the findings. A similar sum- 5. To develop instrumention, test procedures,
marization of the findings of the rigid pave- data, charts, graphs, and formulas, which will re-
5,000
10.000
0
9
0
9
15.000
0.000
U)
z
5.000
a.
>-
I-
-J
2.000
U)
(D
S
0.
9
TABLE 33
RESULTS OF MARSHALL STABILITY TESTS ON ASPHALTIC CONCRETE SPECIMENS AT VARIOUS TEST TEMPERATURES
flect the capabilities of the various test sections; equations for use in the study of performance
and which will be helpful in future highway de- over the wide range of designs and loads in-
sign, in the evaluation of the load-carrying capa- cluded in the Road Test. A convenient presenta-
bilities of existing highways and in determining tion of the relationships for the axle loadings
the most promising areas for further highway
research. of the Road Test is shown in Figure 22. For
any axle loading studied itis possible to deter-
Another important objective of the pavement mine a pavement structure that would carry a
research involved the study of the effectiveness specified number of applications before its
of paved shoulders and various treated base serviceability dropped to 2.5. Such determina-
materials. This objective, excluding that por- tions, of course, are subject to the limitations
tion dealing with material reported in Road discussed in Section 1.1.5.
Test Report 6, was stated: These equations represent empirical, service-
3. To make special studies dealing with such ability trend data observed in the test; some
subjects as paved shoulders, base types, pavement Road Test sections failed sooner and some later
fatigue..... and to correlate the findings of than indicated by the smooth curves. There-
these special studies with the results of the basic
research. fore, some allowance should be made for the
scatter of the data. An example of the scatter
Major staff efforts were directed towards the is shown in Figure 25. Most of the observed
fulfillment of the first objective-to find rela- points fall within approximately ± 14 percent
tionships between pavement performance, on of the thickness index given by the curves. If
the one hand, and pavement design, loading comparisons are made with observed perform-
and number of load applications, on the other. ance of actual highways in service, additional
Pavement Performance (Section 2.2.2).- allowance should be made to account for dif-
Onecontribution of the AASHO Road Test was ferences in materials, environment, and loading
the development of a definition of pavement history between the Road Test and the actual
performance in terms of the trend in pavement highway.
serviceability (ability to serve traffic) with in- These relationships are not intended to be
creasing number of load applications. design equations. However, they can serve as
Perhaps the major finding of the flexible a basis for design procedures in which vari-
pavement research was the set of relationships ables, such as soil type, not included in the Road
sought in the first objective. These relation- Test, are considered.
ships were reduced to a set of four equations Section 2.2.2.1 includes tables and discussion
containing terms for the variables that were showing the basis for determining the • signifi-
included in the test. These equations appear cance or nonsignificance of the various effects,
in the text as Eqs. 13, 17, 18 and 19 (for the correlation indexes to show the degree of cor-
case where load applications were adjusted by relation found in the relationships, and mean
a seasonal weighting function). Similar equa- residuals to show the degree of scatter of the
tions are given for unweighted applications. observed performance data from the predic-
Graphs and tables were constructed from the tions of the performance equations.
TABLE 34
CONDITION DATA OF PAVEMENT COMPONENTS, (FROM TRENCHING PROGRAM, 1960)
(b) SUBBASE
3 4-3-8. 4.9 4.4 134.0 136.2 23.7 ' 47.3 5.1 4.2 132.8 132.0 22.6 24.2
4-6-4 5.5 4.6 137.4 137.9 9.2 30.0 5.6 5.0 134.6 131.0 8.8 22.6
4-6-8 5.2 4.7 128.3 132.7 21.4 34.2 , 5.3 4.6 134.5 130.4 24.6 56.0
Mean ' 5.2 4.6 ' 133.2 135.6 18.1 37.2 5.3 4.6 134.0 131.1 18.7 34.3
4 5-6-12 4.8 5.2 143.4 129.0 23.2 49.8 4.8 4.6 1.35.9 126.0 14.9 65.8
5-6-8 5.3 4.8 - 130.4 21.6 54A 5.6 .4.6 - 131.4 22.7 49.5
5-3-12 5.4 4.6 - 132.9 9.4 45.6 5.2 4.8 - 132.5 16.8 26.3
Mean 5.2 4.9 143.4 130.8 , 18.1 49.9 5.2 4.7 135.9 130.0 18.1 47.2
5 5-9-12 , . 5.8 4.8 131.2 139.9 26.8 62.2 5.9 5.1 '139.3 139.7 26.7 28.8
5-6-12 5.9 4.5 ' 134.7 134.8 18.2 42.6 6.2 5.0 134.0 137.3 14.9 29.7
5-9-8 6.3 5.4 135.3 136.9 26.2 74.7 7.2 5.8 132.6 134.8 18.8 52.9
Mean
• 6.0 4.9 133.7 137.2 23.7 59.8 6.4 5.3 135.3 137.3 20.1 '37.1
6 . 6-6-16 5.7 4.8 131.3 138.6 39.1 10.3 6.0 4.8 134.2 131.0 23.2 74.2
• 6-9-12 . 5.6 4.5 134.8 130.3 35.7 85.4 5.8 5.0 135.2 136.0 28.7 . 39.0
6-9-16 5.8 5.4 141.3 142.5 36.2 60.3 6.2 5.4 136.8 132.6 28.8 44.2
Mean 5.7 4.9 135.8 137.2 37.0 52.0 6.0 5.1 135.4 133.2 26.9 52.5
Over-all Mean 5.5 4.8 136.5 135.2 24.2 49.7 5.7 4.9 135.2 132.9 21.0 42.8
FLEXIBLE PAVEMENT RESEARCH 133
C ..
The thickness index found to apply to Road
b.
C) N 0, ..
Test flexible pavements is of interest in itself.
00 N C' ,-C 0') '-I C) N in 0
0) ,.
For the weighted applications case the thick-
ness index equation (Eq. 19) indicates that an
inch of surfacing was about three times as
CO C'1 LtD ,. C) N 01 Gq CO 01 0) CO CO C) b. C C)
effective as an inch of base and four times as
N 00 b. N CO '' o N U) M Co N to 14 oo
'-C
b, effective as an inch of subbase in improving
pavement performance within the range of
design studied.
o
The use of the seasonal weighting function
C'1CO 00 o COC) C'1 00 '
CC) C') 0) ' C') 0) ' C CC) .4 on axle load applications was found to increase
' the correlation index from 0.48 to 0.70 and to
01
reduce the mean residuals by 15 percent.
N C') C) c' Special Base Type Experiment (Section
0.1 C) o C) C') 00 01 C'] C, ,.4 0
' ,- ' '
- , 2.2.2.3) .—An important investigation within
the flexible pavement experiment involved the
ClC
study of four types of base: crushed stone,
gravel, cement-treated and bituminous-treated
C') ' C') C') VO CO o gravel.
.5 The design of the base experiment was such
Cd
• - that no mathematical analysis of the perform-
Lo LCO e
4 ,
00 CUD
, .
CC)
.
ance of the sections was attempted. The
• analysis was essentially graphical. However,
Cd
0)0
it is anticipated that the Highway Research
Board and others will incorporate the special
base data into the data from the main factorial
experiment in an effort to produce performance
equations containing terms for the special base
Ca
IV materials.
TABLE 35
SURFACE TREATMENT EXPERIMENT PAVEMENT
ig 0
C'.] I!) CC) C)4 c, PERFORMANCE DATA AT p = 1.0 FOR PAIRS OF
0) C'] -1. C!) U) U) CC) 00 .'-1 0)
REPLICATE SECTIONS'
2S 0 48 716 (1.7)
° ' 315 797 (1.4)
4 685 (1.2) (1.9)
'00o
C')U)C)to,
C')C,]0'C
COC)NNCOCC
COkbC)'-i C!D01b.
,-ICO,-4O
00
Ct')
r.0 C) 25 (1.1) (1.9)
C') ' ' 'Cl' -'C C') ' Ct
.-1 i - "C 'C 6S 0 30 30 722
• ")C'C
C0 20 95 749
V
V 'ba
raG
4 20 722 (1.3)
CU00 ,-CO00CC c 4 34 755
C') C C') 0 ' C') VD CC) C') CCC') CC) C') C') C') CC) 0) .P,
Cl2
V -
Cd
w (b) WEIGHTED
'C'.]CC) '0)')C')
C').
V~ C! 01
'"]'-'r
2S 0 50 •372 (1.7)
'l'"]'*. *
''•'"]'-* CC].,:I 124 751 (1.4)
4 268 (1.2) (1.9)
0)4-C 27 (1.1), (1.9)
C'] C-]C'] C-] CO C-I CO
0O' 00 400,-'C v-1,-I00 '-1 6S 0 32 ' 32 402
C)CCLC.0 0bCrDC'L, DCC0C C0 22 92 522
U)LC)LLD U)UD 4,1
'V
4 - E-4 '
4
•
22 402 (1.3)
. 0CC/I 5 36 552
) 0
co . CL) CO 'Values in parentheses give serviceability at end of
test traffic. V
134 THE AASHO ROAD TEST, REPORT 5
150
SPRING 1960
SUMMER 1960
U
50
0 0
LOOP 3 LOOP 4 LOOP 5 LOOP 6 LOOP 3 LOOP 4 LOOP 5 LOOP 6
6
150
z
U
Ui
a
Co 4
- 100
U,
-J.
0 w2
50
-J
L)
I-
C,,
4: Co
-j
Ui U
Lel
LOOP 3 LOOP 4 LOOP 5 LOOP 6 - LOOP 3 LOOP 4 LOOP 5 LOOP 6
The results of the analysis are presented in considerable difference in the performance of
graphs (Figs. 35 and 36) which can be used to the treated bases and the crushed stone bases.
compare the performance of the stone, cement- In fact, in all loops and at all levels of service-
treated and bitmuinous-treated bases, that is, to ability the performance of the treated gravel
compare the thickness of the materials that was bases was definitely superior to that of the un-
necessary to maintain a level of serviceability treated crushed stone.
of 2.5 at a specified number of load applications. Most of the sections containing the i.intreated
For example, for the 18-kip single axle load at gravel base failed very early in the test. Data
1,000,000 applications the required thickness from these sections are shown in Figures 30
of base (where the surfacing thickness was 3 and 31 which show that their performance was
in. and the subbase 4 in.) is shown to be ap- definitely inferior to that of the sections with
proximately 13, 8 and 6 in. of stone, cement- crushed stone base.
treated and bituminous-treated base, respec- Paved Shoulder Studies (Section 2.2.2.2).—
tively. These values indicate that there was A study of the effectiveness of paved shoulders
6 150 60
5 125 50
00 40
(U
30
Ir
U
° 50 20
25 10
0 o 0
LOOP3 LOOP4 LOOP 5 LOOP 6 LOOP3 LOOP 4 LOOPS LOOP 6 LOOP3 LOOP4 LOOPS LOOP 6
SPRING 1960
SUMMER 1960
5
0
LOOP 3 LOOP 4 LOOPS LOOP 6 LOOP 3 LOOP 4 LOOP 5 LOOP 6 LOOP 3 LOOP 4 LOOP S LOOP 6
Figure 110. Base course condition data, spring and summer 1960.
136 THE AASHO ROAD TEST, REPORT 5
was included in the Road Test. A total of 48 The bituminous-treated base and surfacing
test sections was provided in this study. Un- material offered greater resistance to consolida-
fortunately, the pavements selected for the tion and displacement at low than at high
tests were underdesigned to the extent that 42 temperature.
of the sections failed during the first spring of Cracking was also an element of structural
traffic operation and, thus, little information deterioration that detracted from serviceability
of value was disclosed by the experiment. and performance of flexible pavement. Records
An attempt was made to obtain additional were maintained of the development of cracks
information by studying the differences in per- in order that relationships could be established
formance of the outer and inner wheelpaths of between cracking, pavement design and load
the test sections of the main experiment. applications.
The results of these studies indicated that the Eq. 29 was developed, from which the number
pavement needed to maintain a certain service- of axle loads sustained by the pavement before
ability at a given number of axle load applica- Class 2 cracking of the surface occurred could
tions would be considerably thinner in the inner be computcd for any design and load. By in-
than in the outer wheelpath. cluding a deflection term, it was found that a
Structural Deterioration (Section 2.2.3).— somewhat better prediction of load application
Studies were made of the seasonal changes in could be obtained (Eqs. 30 and 31).
elevation of the pavements and of the rutting More surface cracking occurred during
in wheelpaths. The studies of rutting included periods when the pavement structure was in a
such factors as the extent to which changes in relatively cold state than during periods of
thickness of the structural components affected warm weather. Generally, cracking was more
the depth of rut, and how much of the thick- prevalent in sections having deeper ruts than
ness change was due to densification and how in sections with shallower ruts.
much was due to lateral displacement. Studies Deflection as Related to Design, Load, Speed
were made also of the seasonal changes in and Temperature (Section 2.3) .—Relationships
physical condition and strength of the pave- were developed between flexible pavement de-
ment components. flection and pavement design, load, vehicle
On an average, the pavement in the various speed and pavement temperature to provide a
loops heaved approximately 0.4 in. during the basis for the deflection vs pavement perform-
winter with the edges rising about 0.6 in. and ance studies reported in Section 2.4.
the interior portion about 0.3 in. (Fig. 44). In the main factorial experiment the asphal-
Most of this heaving was attributed to expan- tic concrete surfacing was much more effective,
sion of the embankment soil. inch for inch, in reducing pavement deflection
Rutting of the pavement was due principally (particularly. during the spring months) than
to decreases in thickness of the component was the base or subbase.
layers. Based on average data from 51 sections The subbase was somewhat more effective
that were trenched in 1960, 32 percent of the than the base in restricting deflection in both
depth of rut could be attributed to a reduction the spring and fall (see Section 2.3.1).
in surfacing thickness, 14 percent to a reduc- In the special base experiment the level of
tion in base thickness and 45 percent to a re- deflection was considerably greater at each
duction in subbase thickness—a total of 91 season (spring, summer and fall) in the sec-
percent. Thus, only 9 percent of a surface rut tions with gravel and stone base (9 in. thick)
could be accounted for by rutting of the em- than in sections with bituminous- and cement-
bankment. treated base of the same thickness.
Only 20 percent of the change in thickness of The deflections of the sections with gravel
the surfacing and 4 percent of the change in base were somewhat lower than those for sec-
subbase thickness could be accounted for by in- tions with stone base although the performa1ce
creases in density of the materials. In the case of the stone base was considerably better than
of the base only 30 percent of the change in that of the gravel base sections (Section 2.3.2).
thickness determined in the summer of 1960 The deflection occurring within the pavement
could be accounted for by increases in density. structure (surface, base and subbase), as well
However, the increase in the density deter- as that at the top of the embankment soil, was
mined in the spring of 1960, accounted for all greater in the spring than during the succeed-
of the decrease in thickness of the material. ing summer months. This was considered to be
In sections that survived the test the rate of due to the higher moisture contents of the base
development of rutting during the first year of and subbase that existed in the spring.
traffic generally exceeded the rate observed A high degree of correlation was found to
during the second year. exist between the deflection at the top of the
In the special base studies there was a level embankment and the total deflection (Section
of base thickness above which the surface rut 2.3.3).
depth remained constant with increase in base A pronounced reduction in deflection accom-
thickness and below which it increased rapidly panied an increase in vehicle speed. Increasing
with decrease in thickness. the speed from 2 to 35 mph reduced the total
FLEXIBLE PAVEMENT RESEARCH 137
deflection 38 percent, the embankment deflec- of load support of flexible pavements, pressures
tion 35 percent, and the partial deflection 67 that are transmitted through the structure to
percent (see Fig. 78 for definition of partial the supporting soil were subjected to limited
deflection). study in Loop 4, in which 20 pressure cells were
In studies of the effect of the temperature of installed in certain sections. The variations in
the asphaltic concrete surfacing upon deflec- transmitted pressure with seasons and speed
tion, it was found that between 80 and 120 F were of the same character as those shown for
the deflection was essentially constant. At creep speed deflections. The transmitted pres-
about 80 F, it began to decrease as the tempera- sures were found to vary approximately as a
ture decreased. Figures 87 through 91 show linear function of the applied load in the same
that the extent of the decrease varied, depend- manner as deflections.
ing upon such factors as the age and traffic Marshall Stability vs Temperature (Section
history of the pavement, the speed of the ve- 2.5.4) .—A limited series of tests was conducted
hicles, the design of the pavement, the type of in the project laboratory utilizing Marshall
base and the time of the year when the tests stability equipment in which molded specimens
were made. of the asphaltic concrete material were tested
Prediction of Performance from Deflection over a temperature range from 40 to 160 F. A
(Section 2.4) .—The fifth Road Test objective well-defined effect of temperature was found
asked for relationships that would employ in- both for the surface and binder course mixture.
formation from dynamic measurements in the In each case the log log of Marshall stability
prediction of future pavement performance. was shown to vary as a linear function of the
Deflections in flexible pavements under loads temperature of the material at the time of
moving at creep speed proved to be highly testing (Fig. 107).
effective for this purpose. The performance of Physical Test Data (Section 2.5.6).—A great
the flexible pavements was predicted with es- deal more structural deterioriation of the flexi-
sentially the same precision from load-deflection ble pavement sections took place during the
data as from load-design information. Deflec- spring months of the year than in the summer,
tions taken during the spring when the sub- fall and winter months (Table 1). In view of
surface conditions were adverse gave a better this fact, data bearing on changes in the con-
prediction of pavement life than those taken in dition of pavement components from season to
the fall. There was a high degree of correlation season were obtained from the 1960 trenching
between deflection and rutting. program (Section 2.2.3.1).
Overlays (Section 2.5.1).—A study of the The increase in strength from spring to
effectiveness of asphaltic concrete overlays in- summer of the base and subbase material was
cluded 99 flexible pavement test sections. It attributed to changes in moisture content
was clear that overlays were highly effective as rather than to changes in density. The increase
a means for extending the service life of these in strength of the embankment soil was not
pavements. accompanied by appreciable differences in
Attempts at mathematical analysis designed moisture or density.
to establish specific relationships between per- Bituminous Surface Treatment (Section
formance and overlay design were unsuccess- 2.5.7).—A limited program of tests involving
ful, because the outcome of each analysis tried surface treatments was conducted on Loop 2
proved to be highly dependent upon the as- as a part of the over-all study of pavement de-
sumptions that had to be made concerning the sign. The performance of the surface treated
mathematical model for the analyses. Further sections was inferior to that of sections with
investigation will be made by the Highway equal thicknesses of base and subbase but with
Research Board. asphaltic concrete surfaces. Their performance
Subsurface Studies in Non-Traffic Loop was improved appreciably as the thickness of
(Loop 1) (Section 2.5.2).—Loop 1 was included the underlying base and subbase course was in-
as a part of the AASHO Road Test in order creased.
to provide a group of traffic-free representative
sections (designs) of pavement that would be 2.6.2 Needed Research
continually available for a number of different In Section 1.4 there is a general discussion of
studies. research that would be desirable to improve
The condition of the subsurface components and simplify the relationships found in the
in the non-traffic loop was more adverse than in AASHO Road Test and to extend the findings
the loops where traffic was a factor. The rate of the Road Test to include other soils, mate-
and depth of heat penetration was greater in rials, and environments. In this subsection the
the pavements than in the granular shoulders. more important areas of research suggested by
There was virtually no loss of serviceability of observations of pavement performance at the
the pavements in the non-traffic loop over the Road Test are discussed.
2-yr period of the test. The Road Test performance equations in-
Embankment Pressure (Section 2.5.3) .—In clude coefficients that distinguish the relative
connection with the problem of the mechanics effectiveness of surfacing, base and subbase in
138 THE AASHO ROAD TEST, REPORT 5
the performance of flexible pavements. These At the Road Test, rutting in the wheelpaths
coefficients relate specifically to the type of as- was largely associated with transverse move-
phaltic concrete surfacing, the type of crushed ment of the component materials. Means to
stone base and the type of subbase used in the prevent such distortion in flexible pavements
experiment. They also relate, but probably to should be investigated, in the case of the sur-
a lesser degree, to the as-constructed character- facing material, methods for increasing stabil-
istics of these materials (primarily their densi- ity without increasing brittleness should be
ties) and to their environment. It is clear that sought. Extensive research into the mechanism
early research is needed in which other typical of base failure should be performed.
flexible paving materials are used in conjunc- The clear superiority of the treated over the
tion with basement soils similar to the Road untreated bases suggests the need for addi-
Test embankment material, in order that the tional experiments permitting direct compari-
coefficients for surfacing, base and subbase can sons between treated and untreated bases on
be modified for different materials where neces- actual highways. The inclusion of treated bases
sary. These studies can be effectively accom- or composite pavements in satellite and field
plished in the satellite tests mentioned in Sec- tests is indicated.
tion 1.4. Also, rough checks on the relative The relatively poor performance of the
effectiveness of different materials may be ob- gravel base, coupled with the fact that its de-
tained from the simpler field tests. flection under load was no greater than that of
Inasmuch as in the Road Test the light axle the stone base, suggests the need for additional
loads were operated at the same frequency as research into the relationship between pave-
the heavier loads, additional research is desir- ment performance and the resiliency—as well
able on the effects of very large numbers of as the strength—of the component layers.
applications of passenger loads over long Because the serviceability (and thus the per-
periods of time. formance) of flexible pavement is highly de-
The interrelationship between structure de- pendent on the longitudinal profile of the pave-
sign and load as shown in the performance ment (and because longitudinal roughness ex-
equations by the exponents of these terms may isted in the newly-constructed pavements even
be somewhat dependent upon the soil type over under extraordinary construction control at
which the structure is built. Therefore, the the Road Test), means should be sought to im-
satellite and the field tests should also include prove the uniformity of compaction and of
embankment soils different from that used in strength of the subsurface materials to mini-
the AASHO Road Test. For the same reason, mize differential settlement that may develop
tests should be conducted in environments dif- along the wheelpaths after traffic has been
ferent from that of Ottawa, Ill. placed on the pavement.
Chapter 3
a a)a .e
Test Section Na Test Section No Test Section No Test Section No. Test Section Na Test Section No
a
.0 c = = - 0
= 0
a a
.0
aa ac C
o Lone Lone ci Lone Lone Lone Lone o au Lone Lane (I) V)o .0 Lane Lone Lone Lane
I 2 I 2 I 2 I 2 I 2
0 935 936 0 805 806 3 195 196 3 643 644 3 513 514 3 353 354
22
6 933934 23 791792 3 6239240 5 6647648 66 517518 8.6393394
889 890 . 6 785 786 9 213 214 -- ZL !!! — - .
0
r6- 925 — --- !- - _i ...L
697 698 a
926 3. 3 8I D
5 6 6-i- 6 8 6 I 6 L.368
891892 6787788 221 222 2_:i.!_ 533 ±
ii ____ra
671__
797798 217218 Z--. _.L_
- 5 3
— -- --
3
0 -----.- I 8 91
2 - II — -- -
----- 6 6 683 684 6525526
12 - 909
-910
— — 0 781 —
782
—
9
—
207
—
208
9
— — 651
— 652
— —9 535
— 536 365 9
366
—
6
913 914 I
23
— 799 — — — — 3 675
676 3 529 530 3 395 396
2!_
890 m 0 793 794 956
2931 .2 3! 3!!.!! S..!L662 6649i49 86385386
3! 5206 9673674 9 549 550 9 347 348
0 905 906 cr — -- --- 232 641 642 519 520 381 382
--- I
— L 928 2 807 808
3
2 51
---- 6-'-29
— --- -- .--- III
907 908 5 _L809 810_ 203 204 0 8
650
0 92 5 — 9 653 654 2 —
—
921 6 795 796 -o 6 191 192 9 531 532 — 9 339 340
6
915 916 ———— — a — — 9 233 3 691 692
— 3
— 553
392
— 554 3 391
887 888 3 199 200
8 6 12_ 91 -
-- L.6 - 543
--
— Z2L 72.
6
0 883 884 247 248 52 4 . —
12
-6-- 911 912
Note 6 6 — 9 695 696 — 9 499 500 — 9 343 344
Shaded sections are 3 645 646 3 515 5i6 3 359 360
Subsurface Studies
Design 5
replicate sections. j7 9 II 546 I2 6 355 356
8
—
6 215 216 — 9 667 668
—9 —
495 496
— — 9 357 358
Test Section No. Cross hatched sections
are those borrowed — 9 98 Shoulder Paving-No Subbase Shoulder Paving-No Subbase Shoulder Paving-No Subbase
Lore Lo2ne Study — Design 3 Study — — Design 3 Study — — Design 3
from Design I Shoulder Paving-No Subbase
j Study — — Design 3 Test Section No. Test Section No. .0 . Test Section No.
893 8s4 - sTestSectionNo Lone Lane &2 Lone Lane Lone Lane
6
929 930 _il---- L 2 "° 'i 2 ° I a
901 902
Lone Lone 0 659 1 660 0 537 538 — 0 373 374
2 6i\
o 3 0 663 664 o 0493 494 u 0 383 384
-5 ;
TabIe 36 Designsfor Rigid ________ __ . ________
Pavement Experiments 6.!2 06 227 228 0 699 700 0 551 552 0 399 400
187 188
8
6 657 658 9!2 6 527 528
11 6 387 388
RIGID PAVEMENT RESEARCH 141
1/ 2' PER FT. SLOPE
3" CRUSHED STONE SHOULDERS
11
cedures applicable to the embankment and sub- The portland cement concrete contained
base are summarized. coarse aggregate, a natural sand, Type I port-
This subsection presents, in tabular form, land cement, water, and an air-entraining
the results of standard tests made on the agent. The coarse aggregate, an uncrushed
cement, water and aggregates used in the port- river gravel, was obtained in two sizes, size A
land cement concrete. It also includes propor- (21/-in. maximum) and size B (hA-in, maxi-
tioning data, the results of tests made on the mum). Both sizes were used in concrete pave-
plastic as well as the hardened concrete, and a ments 5 in. and greater in thickness. . Only the
summary of changes in flexural and compres- size B material was used for the 21/2 - and 3i/,
sive strengths with time. in. thick pavements. A summary of gradation
The mix designs were based on a fixed tests made on the coarse and fine aggregates is
cement factor of 6 bags per cu yd, a maximum given in Table 39.
water content of 5 gal per bag of cement, and Table 40 gives averages of the results of a
a sand-to-total-aggregate ratio of about one to number of tests made to determine physical and
three. Average air content was 4 percent. chemical properties of the portland cement and
Flexural strengths (AASHO Designation T97- chemical properties of the mixing water.
57) averaged about 650 psi at 14 days; com- The mix designs were based on a fixed ce-
pressive strengths (AASHO Designation T22- ment factor of 1.50 bbl (6 bags) per cu yd, a
57) about 3,980 psi at 14 days. maximum water content of about 5 gal per bag
Standard procedures were followed in mix- of cement, a sand-to-total-aggregate ratio of
ing and placing the concrete, which was spread about one to three. Proportioning data are
in both 12-ft lanes simultaneously. Wet straw given in Table 41. Properties of the plastic
was used in curing. concrete and the flexural and compressive
The materials used in the rigid pavement strengths of the hardened concrete at 14 days
test sections and the methods of construction are given in Table 42.
are described in detail in Report 2 along with Standard procedures were followed in the
comprehensive summaries of material control mixing and placing of the concrete. Cement
tests. Basic data concerning materials are and aggregates, batched at a central plant,
given in various data systems (see Appendix were delivered in 4-compartment batch trucks
I). A brief summary of the characteristics of to the 34—E dual drum payers. Mixing time
the materials and the methods of construction was not less than 60 sec, and all water had to
is presented in this subsection. be added within the first 15 sec. The air-
The embankment upon which the test sec- entraining agent was introduced at the mixer.
tions were constructed is described in Section The concrete was spread mechanically in
2.1.2. The subbase for rigid pavements con- both lanes simultaneously, and the operation
sisted of the same material used for subbase was continuous for the full length of a struc-
in the flexible pavements and was a locally tural section. Construction joints, when re-
available sand-gravel material modified by the quired, were formed in the transition between
addition of small amounts of fine sand and fri- sections. Pavement fabric for pavements 5 in.
able fine-grained soil. The material was pro- and greater in thickness was placed by the
duced in a washing and screening plant. The double strike-off method; for the 21,6 and 33,4
soil fines were later added and mixed with the in. pavements it was set in place and welded to
fines in a concrete mixer. In the summer of chairs prior to placing the concrete. The se-
1957 the subbase material was placed on the quence of operations included placing and
embankment as a means of protecting the spreading, strike-off and consolidation, longi-
underlying soil during the winter. Engineering tudinal floating, straightedging, belting, edg-
characteristics of the material are given in ing, and final finishing with a burlap drag. Wet
Table 4. straw was used to cure the concrete.
142 THE AAS!lO ROAD TEST, RE PORT 5
In addition to the beams and cylinders This section describes the rigid pavement
molded (In ring the paving operations for con- present serviceability index, and the i-elation-
tiol of construction, specimens were made fur ship of rigid pavement performance to design
determining Ilexural and compiessiVe strengths and load based on data from the main lactorial
at intervals over a period of two yeais. Table experiment (1)esign 1) . It also presents the
4:3 summarizes the results of tests made on results of the shoulder paving—no subbase
these Sl )ecifllens. Stu(ly (I)esign 3), and describes elements of
pavement deterioration observed (luring the
3.2 PAVEMENT PERFORMANCE traffic testing.
The concept of pavement serviceability, the !fl(/CX foi- RIfJid POV('fli('fli
1.2.1 Sc?(1Ce(lbi1f1J
derivation of the serviceability indexes, and the
rationale for the analysis of pavement per- This SuI)section contains the equations used
formance data are described briefly in Section to (letermilie the present serviceability index
1.3, and in detail in Appendixes F and G. of each rigid pavement section E(Is. 59 and
60) . Also included are Tables 44 and 45 which
give, for each section, the nurnl)er of un-
weighted applications sUstaiflC(l before the
serviceability index fell to 2.5 am-I 1.5. If the
serViceibilit7,' (lid not tOil to those levels, the
tables give the section's index value at the end
of the traffic test.
Eq. 59 was used to determine the level of
serviceability of the surviving rigid pavement
test sections every two weeks (luring the period
of traffic operation.
p 5.41— 1.0 log (1 SV) - 0.09 /C ± P
(59)
in \VI1iCh
7)
present serviceability index;
mean of the slope variance in
SV
the two wheelpaths; and
C and P - measures of cracking and
patch I ng in the pavement su r-
face. (In this equation and
throughout this report, loga-
rithms are to the base 10.)
SV was discussed in Section 1.3. Cracking,
C ( Eq. 59), is defined as the total linear feet of
Class 3 and Class 4 cracks per 1,000 sq ft of
pavement area. The length of a crack is taken
as the length of its I)roJectiOn parallel or per-
pendicular to the pavement centerline, which-
ever is greater. A Class 3 crack is defined as a
crack opened or spoiled at the surface to a
width of 1/i in. or more over a distance equal
to at least one-half the crack length, except
that any portion of the crack opened less than
% in. at the surface for a distance of 3 ft or
more is classified separately. A Class 4 crack
is defined as any crack which has been sealed.
Patching, P. is exl)ressed in square feet per
1,000 sq ft of pavement surfacing.
\Vhen it was not feasible to use the project's
-,
- -
- longitudinal p rofilometer to determine the serv-
iceability of a test section, the Bureau of Pui)lic
• Roads roughometer (Fig. 113) was used. The
-: roughometer was equipped with a special
-: - counter and opei-ated at a speed of 10 mph.
Through a study correlating the output of the
Figure 112. Paving on test tangent. roughometer with that of the profilometer, it
RIGID PAVEMENT RESEARCH 143
TABLE 38
DETAILS OF PORTLAND CEMENT CONCRETE SURFACING
TABLE 39 TABLE 40
SUMMARY OF GRADATION TESTS CHARACTERISTICS OF PORTLAND CEMENT
ON PORTLAND CEMENT CONCRETE AGGREGATES AND MIXING WATER
axle load, the number of applications of the serviceability index of a Road Test section
axle load, and the thickness of the pavement having a 5-in, thick slab (on granular sub-
slab. base). From the figure, the horizontal line rep-
The equations are represented by graphs pro- resenting a slab thickness of 5 in. intersects the
vided for the purpose of simplifying their in- "18-kip single" curve at about 400,000 axle load
terpretation (for an example see Fig2 116) and applications. (A comparison of this prediction
for displaying their predictions in conjunction with the actual performance of the six test
with the observed data (Figs. 120 through sections with a slab thickness of 5 in. subjected
123). to an 18-kip single axle load, is shown in Figure
An example of using the graph (Fig. 116)
follows. Itis desired to estimate the number of 120.)
applications of an 18-kip single axle load asso- Because of random variations in the observed
ciated with a reduction from 4.5 to 2.5 in the data, there were unavoidable differences be-
TABLE 41 TABLE 42
SUMMARY OF TEST RESULTS
PORTLAND CEMENT CONCRETE PROPORTIONING DATA
PER BAG OF CEMENT ON PLASTIC AND HARDENED CONCRETE
TABLE 43
SUMMARY OF CONCRETE STRENGTH TESTS
tween predictions from the equations and the vironment and loading history between the
actual performance of md ividual sections. highway and the Road Test.)
Thus, in using the curves, some allowance Included in the subsection are tables and dis-
should be made for the scatter of the data. cussion showing the basis for determination of
Analysis of the residuals shows that the scatter the significance or nonsignificance of the load
corresponds approximately to ±12 percent of and design variables. Also given are correla-
the slab thickness given by the performance tion indexes disclosing the degree of correlation
curves. (If comparisons are made with the found in the relationships, and mean residuals
observed performance of an actual highway in showing the degree of scatter of the observed
service, additional allowance should be made data from the predictions of the performance
to account for differences in materials, en- equations.
Loop Ld Thk 2.5-In. 3.5-In. 5.0-In. 6.5-In. 8.0-In. 9.5-In. 11.0-In. 12.5-In.
(kips) (in Surface Surface Surface Surface Surface Surface Surface Surface
R N R N R N R N R N R N R N R N
TABLE 46
ANALYSIS OF VARIANCE FOR LOG p ESTIMATES 1 WITHIN Loops
Effects3 :
Lane mean' difference 3.77 0.41 05 0.4.1 0.00
D,, reinforcement:
Lanes combined 0.01 0.03 0.00 0.00 0.07
Lane interaction S 0.00 0.03 0.05 0.00 0.02
D2, slab thickness:
Lanes combined 0.74 700 6.60 3.66 2.23
Lane interaction 0.99 0.07 0.14 0.00 0.02
D2, subbase thickness: -
Lanes combined 0.12 003 0.05 0.01 0.02
Lane interaction 0.06 0.01 0.01 0.00 0.00
Interactions among D1, D2, D3 :
Lanes combined 0.07 0.02 0.04 0.05 0.06
Lane interaction 0.01 0.03 0.02 0.02 0.01
Replicate differences:
Lanes combined 0.08 0.09 0.08 0.08 0.04
Lane interaction 0.01 0.02 0.02 0.02 0.01
2
Data from which this table arose were estimates log p as described in Appendix G.
Excludes thickest slabs in each lane.
'Mean squares for effects (underlined values considered to be significant relative to replicate differences
pooled with interaction effects.
4 Lane 2.
152 THE AASHO ROAD TEST, REPORT 5
none of the ratios was found to be significantly 66. The last two lines (Table 46) indicate that
larger than one, these interactions could be the major part of within-loop variation in log
ignored. The ratios of mean squares for D1 , p estimates was accounted for by slab thickness,
D. or D2 effects to interaction or replicate mean and that the residual variation had mean
squares were also found not to be significant. squares quite similar to the replicate difference
Only the D2 and lane effects were significant as mean squares. Other variance analyses of rigid
indicated by the underlined mean squares. Thus pavement performance data consistently gave
in the expression a,D, + a2D2 + a,D3 + a4, the same type of results with respect to pave-
both a1 and a, were assigned the value zero so ment design effects.
that D, and D, terms would not occur in the The remaining undetermined constants in
expression. For convenience it was assumed Eqs. 63 and 64 were estimated by applying the
that a2 and a., were both 1, and so pavement procedures described in Appendix G to the per-
design was characterized by the expression formance data given in Appendix A for rigid
D2 + 1. With /3 = 1, Eqs. 6and 7 are there- pavements, Design 1.
fore reduced to the following forms: A weighting function was not used for rea-
sons given in Sections 1.3.4, and the symbol, W,
B. (L1 + L2)B2 in Eq. 62 represents actual (unweighted) axle
applications. The expressions for /3 and p de-
/3=1+ (63)
(D2 + 1)8, L21 I termined from the analysis were
The last part of Table 46 shows the results of 10585 (D2 + 1) 7.35 L2328
within-loop regression analyses that were used p= (66)
to determine a value for A,, the coefficient for (L1 + L2 ) 462
log (D2 + 1) in the logarithmic form of Eq.
64. This coefficient varied from about 6.9 to Although the subbase thickness term, D3,
8.4 in the various loops, and the weighted aver- does not occur in the performance equations,
age value used for A, was 7.35 as shown in Eq. the equations do relate to pavements whose sub-
EXTRAPOLATED CURVE
I
--
10
tTi1 k:::i
I I liii __---r7I
IOU
--
IVUU
EXTRA
Figure 116. Rigid pavement performance curves from Road Test equation,
experiment design 1, for p = 2.5.
RIGID PAVEMENT RESEARCH 153
base thickness is in the range of 3 to 9 in. Data and 123 are similar except that the terminal
from Design 3, described in Section 3.2.2.2, in-. serviceability level is 1.5.
dicated that sections on 6 in. of subbase gave Points shown in Figures 120 through 123
better average performance than sections with represent only those sections whose serviceabil-
no subbase. ity fell to 2.5 or to 1.5. All remaining sections
For a particular pavement design and axle would be represented by points on the right of
load, Eqs. 65 and 66 give values for /3 and p that W = 1,114,000 applications. The number of
may be substituted in Eq. 61 if p is to be esti- such sections can be deduced from the fact that
mated from W, or in Eq. 62 if W is to be esti- at each D2 level there were four test sections in
mated when p is given. Figures 116 and 117 Loop 2 lanes and six test sections in each of
show how W varies with D2 in Eq. 62 when p the Loop 3 through 6 lanes. Although these
is fixed at 2.5 and 1.5, respectively. In each sections are not shown, their performance data
figure there are ten curves, one curve for each did influence the developed equations and
test load used in the Road Test. curves.
Figure 118 shows design requirements, com- The performance data in Appendix A, De-
puted from Eq. 62 when the final serviceability sign 1, give a minimum of 5 and a maximum of
value is p = 2.5, for a range of single and tan- 10 (p. log W) pairs for each test section. When
dem axle loads at three levels of load applica- p is fixed at 3.5, 3.0, 2.5, 2.0 and 1.5 there can
tions. Similar relationships are shown in Fig- be as many as five log W observations, and
ure 119 for a final serviceability level of 1.5. when log W is fixed at t = 11, 22, 33, 44 and
Figures 120 and 121 show the correspondence 55 index days there can be as many as five ob-
between the individual curves of Figure 116 served values for p. Corresponding to each
and performance data from Appendix A, De- observation, log W, or p, is a calculated value,
sign 1, for each of the ten traffic lanes. Each log W or , obtained from the performance re-
point represents the observed number of appli- lationships (Eqs. 61, 62, 65 and 66). Differ-
cations at which a test section had a service- ences between calculated and observed values
ability level of 2.5. Horizontal deviations of are the residuals A lo'W = log W - log W and
the points from the curves represent prediction p =P - p. Absolute values of these resid-
errors or residuals when Eq. 62 is used to pre- uals are summarized in the upper part of Table
dict the life of a section (to p = 2.5) whose de- 47. For each lane and for all lanes, the number
sign and load values are specified. Figures 122 of residuals of each type as well as mean ab-
154 THE AASHO ROAD TEST, REPORT 5
I- I I
SINGLE AXLE VEHICLE
---TANDEM AXLE VEHICLE
—
--
— -I
.-''-1.000.000 APPLICATIONS
o
IO0.000IPPLICAT ONS
I10.00.
I ONS
I
00 4 8 12 16 20 24 28 32 36
SINGLE AXLE LOAD. KIPS
0 8 16 24 32 40 48 56 64 72
TANDEM AXLE LOAD. KIPS
Figure 118. Single-tandem axle load relationship from Road Test equa-
tion, experiment design 1, for p = 2.5. All curves extrapolated
except 1,000,000 applications.
solute residuals are given. The mean absolute minus either one or two mean residuals on
residual for log W is 0.17 and for p, 0.24. either side of the performance curves. This
Log W residuals are horizontal deviations distribution supports the statement that the
from the performance equation curves and are observations, in about, nine out of ten cases,
thus of special interest. The second part of agree with the performance curves to within
Table 47 shows a rather low correlation index two mean residuals. In other words, there is
of, 0.16 for the log W residual analysis and also approximately 90 percent confidence4 that log
gives 0.22 as the root mean square residual. W will be found between log"W - 0.34 and
The low correlation index is due in part to the log"W + 0.34. In terms of the design inde&D2,
relatively narrow spread in log W values. The this band corresponds approximately to D2 ±
general nature of the A log W distribution is 0.11 (2 + 1) where D, is obtained by entering
indicated by the information that 58 percent
and 87 percent of all A log W are contained, * Table 47 includes the root mean square residual so
that twice this value can be used to set limits with
respectively, within bands formed by plus or approximately 95 percent confidence.
10
I
SINGLE AXLE VEHICLE
---TANDEM AXLE VEHICLE
-
I,1000,000 APPLICATIONS
00.0004PLICATIONS
:
2 OOOAPLIATIONS
4 8 12 16 20 24 28 323t
SINGLE AXLE LOAD. KIPS
0 8 16 24 32 40 48 56 64 72
TANDEM AXLE LOAD. KIPS
Figure 119. Single-tandem axle load relationship from Road Test equa-
tion, experiment design 1, for p = 1.5. All curves extrapolated
except 1,000,000 applications.
RIGID PAVEMENT RESEARCH 155
the performance graphs with specified W and if the fit is to be judged ivholly adequate. Since
reading off D2 on the vertical scale. For the the performance equations were developed
Road Test conditions and range of variables it across lanes and loops, it is expected that the
follows that rigid pavement design require- average residuals will be more than one-half
ments are estimated by the performance equa- the average replicate difference, but how much
tions to within about 12 percent of the slab greater cannot be determined in the absence of
thickness. replicate lanes and loops. In the Road Test
The last part of Table 47 summarizes log W performance analyses, it has been supposed
and p differences that were observed between that a satisfactory model and fit is indicated
replicate test sections. Altogether there were whenever mean absolute residuals are about
36 pairs of replicate sections in Design 1, and equal to replicate mean differences. Table 47
the mean replicate difference in p and log W is shows this comparison as 0.24 vs 0.31 for p
0.31 and 0.12, respectively. and 0.17 vs 0.12 for log W. It is quite possible
In six pairs, one replicate was out of test and that other models and fitting procedures may
the other replicate was not at the end of test do equally well, and, that some will represent
traffic. The in-test replicate had terminal serv- better the long-time performance of highways
iceability above 3.5 in three cases, and above in actual service.
3.0, 2.5 and 2.0 in the remaining three cases. 3.2.2.2 Subbase, Paved Shoulder Experiment.
Replicate differences in these cases were as- -This subsection presents the results of com-
sumed to be at least as large as when the first parisons of the performance of sections with
section went out of test at p = 1.5. and without subbase, and of sections with and
For whatever reasons two replicate sections without shoulder paving. The subbase was 6
do not show the same performance, it can be in. in thickness; shoulder paving was a 3-in.
expected that the performance data will deviate layer of asphaltic concrete 6 ft wide. No in-
from any fitted equation. For a particular lane
a satisfactory model and fitting procedure crease in life resulted from use of paved shoul-
should result in residuals that average to be ders. However, the results may have been
about the same as deviations of replicate ob- affected in some cases by damage to the
servations from their own mean. For two rep- shoulder by test traffic. Sections with subbase
licates, then, the average estimation error had an average life about one-third longer than
should be about one-half the replicate difference that of sections without subbase.
TABLE 47
SUMMARY OF PERFORMANCE EQUATION RESIDUALS AND REPLICATE DIFFERENCES
_1
2 1 2 1 22 100 0.21
2 6 1 22 95 0.39 19 0.21
3 1 12 1 28 120 0.19 40 0.11
2 24 2 28 110 0.34 67 0.20
4 2 18 1 28 124 0.21 37 0.13
2 32 2 28 111 0.15 64 0.15
5 1 22.4 1 28 127 0.20 40 0.18
2 40 2 28 122 0.21 46 0.21
6 1 30 1 28 134 0.30 28 0.12
2 48 2 28 131 0.23 30 0.22
All lanes 268 1174 0.24 371 0.17
AREA OF EXTRAPOLATION
U)
7.5
C,
z I-- S
-.
--.
— I
4
—J 2.5
U)
0
10 100 1000
AXLE LOAD APPLICATIONS IN THOUSANDS
30 KIP SINGLE AXLE LOAD
O NONREINFORCED
REINFORCED
10 100 1000
AXLE LOAD APPLICATIONS IN THOUSANDS
2 KIP SINGLE AXLE LOAD 12 KIP SINGLE AXLE LOAD
-. 10.0
I 111111
EXTRAPOLATED CURVE
U)
C.,
z
- wU)
in
_5.0
-o
I
I—.
2.5
/
U)
7.5
C,
z
U)
U)
Ui O•0
z 5.0
C.,
=
I-
U)
j 2.5
U)
10 100 10
I AXLE LOAD APPLICATIONS IN THOUSANDS
5U IcW bINULL #ALt. LUMLJ
I I HIM I I I o NONREINFORCED
REINFORCED
U)
2.5
U)
jarrn-.
100 10011
I1!
TABLE 48
LOG W AT p = 2.5, OR p AT END OF TRAFFIC TEsT1
Value2 of Log W or p
The 64 test sections in Design 3 listed in 3.2.3 Structural Deterioration and Deforma-
Table 36 comprised an experiment which per- tion
mitted direct comparisons of the performance This section describes elements of pavement
of sections (a) without subbase and without deterioration which may have led directly or
paved shoulders, (b) without subbase and with indirectly to a lowering of the serviceability
paved shoulders, (c) with subbase and without level and eventual failure of some of the rigid
paved shoulders, and (d) with subbase and with pavement test sections.
paved shoulders. Faulting occasionally occurred at cracks,
Complete performance data for Design 3 never at transverse joints. (All joints were
sections are in Appendix A. A portion of the doweled). There was a tendency for the crack-
data is presented in Table 48. For each section ing (per unit of surface area) in reinforced
the logarithm of the number of axle applica- sections with 40-ft panel lengths to exceed
tions corresponding to a serviceability level of that in nonreinforced sections having 15-ft
2.5 is given if the section fell to that level panel lengths. No part of the cracking of pave-
during the traffic testing, or the final value of p ments in the traffic loops was attributed solely
is given if the terminal serviceability level of to environmental changes, since no cracks were
the section exceeded 2.5. apparent in the non-traffic ioop (Loop 1).
The four comparisons of performance de- From cracking data, equations were derived
scribed above may be made within each group from which the number of axle applications
of four numbers in the body of Table 48, pro- associated with any given level of cracking can
vided the four numbers represent the same be computed for a given pavement design and
kind of data (log W or p, but not both). To load (see Eqs. 69, 70, 71 and 72). Graphs of
facilitate making the comparisons, groups the equations for a selected level of cracking
which contain data of the same kind have been are shown in Figure 127.
enclosed in rules. Several reversals of trend Longitudinal cracks tended to originate at
are apparent. To determine the average effect transverse joints near dowel bars in 2.5-, 3.5-
of the principal design variables (paved shoul- and 5-in, slabs but not in thicker pavements.
ders and subbase) corresponding numbers of Pumping of subbase material, including the
the six groups of log W data and of the four coarser fractions, was a major factor in the
groups of the serviceability data were averaged majority of the failures of sections with sub-
and tabulated in Table 49. From significance base. Pumping of embankment material was
tests made on the data, it is concluded that generally confined to those sections constructed
paved shoulders did not add to the performance without subbase and severe pumping of sub-
of test sections, while the presence of subbase base material was experienced only in the sec-
was, on the average, of benefit. Numerically, tions with the two thinner slab thicknesses in
the average benefit in increased life (Table 49) each loop. The amount of either material
was equivalent to 32 percent or about one- pumped through joints and cracks was negligi-
third. ble when compared with the amount ejected
The shoulder pavement, consisting of 3 in. along the edge. (See Figure 133.)
of asphaltic concrete, was placed directly on the 3.2.3.1 Cracking and Faulting.—Each test
embankment in the case of those sections with- section was inspected at least once each week
out subbase, and on the subbase in the case of for defects such as cracking, spalling, pop-outs,
the other sections. The shoulder paving was blowups, and faulting at joints and cracks.
damaged severely by test traffic. Although an Faulting.—Faulting at cracks sometimes
effort was made to keep the shoulders in repair, occurred in the later stages of pavement de-
there were instances where rain fell in the in- terioration, but faulting at joints was notably
terim between damage and maintenance. It is absent throughout the project. One transverse
likely that these circumstances had significant joint faulted seriously, but investigation
influence on the results of the experiment. showed that the joint had been accidentally
TABLE 49
AVERAGED DATA, DESIGN 3
sawed at some distance beyond the end of the the pavement adjacent to the blowup had been
dowels intended to protect it. Over the 2-yr removed in the course of maintenance opera-
period of the test there were no other cases of tions.
measurable faulting at joints, all of which were Classification of C racks.—Cracks were di-
doweled (see Table 38 for details of joint con- vided into four classes, depending upon their
struction). appearance, as follows: Class 1 included fine
Blowups.—The rigid pavements were con- cracks not visible under dry surface conditions
structed without expansion joints except at ap- to a man with good vision standing at a dis-
proaches to the test bridges in Loops 5 and tance of 15 ft. Class 2 cracks were those that
6. One blowup occurred in Loop 2 at the trans- could be seen at a distance of 15 ft, but which
verse joint at one end of a structural section exhibited only minor spalling such that the
with a slab thickness of 2.5 in. At both ends of opening at the surface was less than J/1, in.
this section the abutting sections had a slab Class 3 and 4 cracks were defined in Section
thickness of 5 in. Two other blowups occurred 3.2.1. Figure 124 shows cracks of various class-
at transverse joints—One after one-half (12 ft) ifications.
and the other after three-quarters (18 ft) of During each weekly survey, maps were pre-
owe
.d Ij4
;1, ,•.. .•... . 'I4i
Figure 124. Examples of the four classes of cracks in rigid pavement at the Road Test: upper left, Class 1; upper right, Class
2; lower left, Class 3; lower right, Class 4. Only Class 3 and 4 cracks entered into the determination of the serviceability index.
RIGID PAVEMENT RESEARCH 163
DIRECTION OF TRAFFIC
Figure 125. Progression of cracking in a 3.5-in. nonreinforced section with paved shoulders on 6.0 in. of
subbase, 24-kip tandem axle load.
pared showing the location and classification of Different values of C' may occur at about the
each crack, as well as the length of its projec- same serviceability level; for example, when
tion parallel or perpendicular to the centerline the 3.5-in, pavement of Figure 125 was nearing
of the pavement, whichever projection was failure (p = 1.8) C' had the value 233, while
greater. Crack lengths in each section were at p = 1.5 the 8-in, pavement of Figure 126
totaled by classes, divided by the area of the had a cracking index of only 92 (in these fig-
pavement, and recorded each index day in units ures no attempt has been made to distinguish
of feet of projected cracks per 1,000 sq ft of the various classes of cracking from each
pavement surface. These statistics are avail- other).
able in DS 4202 for each test section on each Values of C' determined for each section in
index day for the 2-yr period of traffic testing. Designs 1 and 3 on each index day are avail-
Cracking Index.—In Section 3.2.1 it was able in DS 4202. These data in part are given
stated that only Class 3 and 4 cracking entered in Table 50, where C' is given at W = 1,114,-
into the term, C, in the serviceability index 000 unweighted axle applications for all Design
from which pavement performance was com- 1 sections that survived the traffic test and at
puted. Another statistic, useful in studies of p = 1.5 for all sections that failed before the
cracking apart from the pavement servicea- end of the traffic test. Table 51 is similar ex-
bility concept, is the total projected length of cept that C' is given at a serviceability level of
all cracks, in feet per 1,000 sq ft of pavement 2.5 (instead of 1.5) for sections that dropped
area, represented by the symbol, C.* Figures to that level.
125 and 126 illustrate the relationship between The data in Table 50 permitted 27 inde-
changes in the value of C' and changes in the pendent comparisons between reinforced and
general appearance of the pavement as crack- nonreinforced sections at p = 1.5. From Table
ing progressed with load applications and age. 51, 29 such comparisons could be made at p
* In arriving at the value of C' for a section that had = 2.5. From either table, the relevant data be-
been patched, the patched area was assigned the crack- ing the same, 77 comparisons could be made at
ing equivalent of 1 ft of crack for each square foot of the end of the traffic testing.
patch. Thus, C' differs from the term C + P occurring A summary of the results of these three sets
in the serviceability index, only in that C' includes all
classes of cracks while C + P includes only Class 3 of comparisons (each of which included a test
and 4 cracks. of the statistical significance of the difference
164 THE AASHO ROAD TEST, REPORT 5
Figure 126. Progression of cracking in an 8.0-in. nonreinforced section on 3.0 in. of subbase, 30-kip single
axle load.
in cracking between reinforced and nonrein- average value of C' at p = 2.5 for the 73 sec-
forced sections) is given in Table 52. From the tions which dropped to that level (Table 51)
average data, the cracking in the reinforced was 103, and the standard deviation was 42.
sections with 40-ft panel lengths exceeded that From the average values of C' given in the
in the nonreinforced sections with 15-ft panel preceding paragraph and in Table 52, it was
lengths by 20 to 24 ft per 1000 sq ft of pave- concluded that a value of C' = 100 represented
ment area. However, the difference was not a substantial amount of structural deteriora-
significant-at p = 1.5, and was significant only tion in most cases, and that the relationships
at the 10 percent level at p = 2.5. At the end between design (including reinforcing), load
of the traffic test, when the average service- and load applications at this level of the crack-
ability level of the 154 sections involved in the ing index would constitute a useful supplement
comparisons was 4.15, the average difference in to the performance curves in Figures 116 and
cracking was highly significant, although the 117.
average values of C' were small (30 and 6). Analysis of Cracking Index.—Plots of C'
The observation that cracking in surviving versus W for individual sections suggested the
reinforced sections was frequently greater than use of the following model for the analysis of
in nonreinforced sections (other factors being the cracking index:
equal) appears to confirm the usual assumption
made by concrete pavement engineers that A0 L1 " Jl 2
C' = (67)
tensile stresses occurring during periods of de- D2
creasing temperature tend to increase with
panel length. However, no cracking occurred in in which, W is unweighted axle applications,
the Loop 1 pavements not subjected to traffic A0 , A1 and A2 are a set of constants to be deter-
(see Section 3.5.6.) ; thus, none of the cracks mined from analysis of the data given in Table
appearing in the traffic loops can be attributed 50, and the other symbols are as previously
solely to environment (temperature changes, defined.
moisture changes, subgrade restraint, etc.). The data given in Table 50, together with the
The average value of C' at p = 1.5 for the 67 corresponding W data given in Appendix A,
failed sections in Design 1 (see Table 50) was were used for determining four sets of the con-
168, and the standard deviation of individual stants A0 , A1 and A2, in accordance with the
values of C' about their mean was 71. The following procedure:
TABLE 50
CRACKING INDEX, C', AT 1,114,000 AXLE APPLICATIONS
OR WHEN p = 1.51 , EXPERIMENT DESIGN 12
2 2S 0 4 0 0 1 1 0
3 . 0 0
• 26 2 3 1 1 11
6 13 3 1 0 4 0
0 129* 183* 20 60 0 0
3 15 0
• 118* 115 7 2 1 9
6 63 115 36 8 4 0
3 12S 3 i51 1
343* 258* 60 8 27 1 34 0
6 . 8 23
286* 256* 211* 16 35 1 35 1
9 252* 235* 48 31 31 0 28 0
24T 3 126* 0
171* 216* 169* 67* 46 0 25 1
6 . 56 25
194* 373* 143* 144* 43 0 35 0
9 212* 218* 152* 155* 38 0 23 0
4 18S 3 . 63 1 .
178* 64* 71 8 38 0 39 1
0 36
112* 116* 80 0 52 0 25 1
9 77* 162* 145 26 46 0 29 1
32T 3 132 0
250* 119* 149* 154* 37 1 37 0
6 . 21 44 1 .14
391* 126* 171* .155* 61 33 1
9 131* 205* .173* 193* 45 0 30 0
5 22.4S 3 150* 0
195* 114* 30 0 38 0 4 10
6 0 17
184* 189* 47 9 15 42 8 0
9 100* 117* 44 88* 18 1 0 0
40T 3 . 179* 0
153* 122* 34 4 42 4 19 2
6 24 38
293* 123* 157* 9 24 17 13 0
9 209* 105* 82 98* 50 19 10 0
6 30S 3 6 - 0
93* 92* 200 28 15 0 4 9
6 4 33
126* 29 44 1 31 0 21
9 • 246* 12 164 0 22 0 0 0
48T 3 19 • 0
195* 68 41 30 25 0 8 8
6 . 9 41
74* 0 66 0 33 0 26 3 i-
9 • • • 58* 73* 163* 2 56 0 11 0.
1
Values with asterisk are for p 1.5.
= reinforced; N = nonreinforced.
TABLE 51
CRACKING INDEX, C', AT END OF TRAFFIC TEST
OR WHEN p = 2.51 , EXPERIMENT DESIGN 12
2 2S 0 4 0 0 1 1 0
3 0 0
26 2 3 1 1 11
6 13 3 1 0 4 -O
6S 0 106* 52* 20 60 0 0
3 15 0
109* 112* 7 2 1 9
6 63 115 36 8 4 0
3 12S 3 51 1
218* 160* 60 8 27 1 34 0
6 8 23
172* 195* 115* 16 35 1 35 1
9 140* 128* 48 31 31 0 28 0
24T 3 97* 0
121* 161* 145* 62* 46 0 25 1
6 56 25
96* 217* 113* 60* 43 0 35 0
9 118* 107* 96* 54* 38 0 23 0
4 18S 3 63 1
92* .50* 71 8 38 0 39 1
6 0 36
112* 82* . 80 0 52 0 25 1
9 64* 76* 108* 26 46 0 29 1
32T 3 112* 0
163* 88* 93* 125* 37 1 37 0
6 21 44
47* 122* 98* 84* 61 1 33 1
9 - 90* 147* 138* 98* 45 0 30 0
5 22.4S 3 . . 98* 0
121* 53* 30 0 38 0 4 10
6 0 17
99* 99* 47 9 15 42 8 0.
9 58* 73* 44 42* 18 1 0 0
40T 3 105* • 0
115* 34* 34 4 42 4 19 2
6 . . '24 . 38
179* 22* 113* 9 24 17 13 0
9 • . 188* 83* 82 54* 50 19 10 0
6 30S 3. . 6 0
93* 65* 133* 28 15 0 4 9
6 . 4 33
64* 29 44 1 31 0 21
9 . 135* 12 105* 0 22 0 0 0
48T 3 19 0
110* 55* 41 30 25 0 8 8
6 . . 9 41
74* 0 66 0 33 0 26 3
9 55* 62* 96* 2 56 0 . 11 .0
base thickness, since there was no consistent L, + 3.96 log D, ± 0.11 (72)
trend of C' with subbase thickness. If the aver- If the curves for tandem axle vehicles on non-
age so obtained was zero (i.e., no cracking in reinforced pavement for C' = 100 (Fig. 127)
any of the sections), the average was not used are compared with the tandem axle curves of
in the analysis since the logarithm does not Figures 117 and 118, rather close agreement is
exist. found. However, other comparisons of the per-
Data from the thickest pavements in formance with the cracking index equations
each loop were not used in the analysis. (This show varying degrees of divergence between
rule was also followed in the analysis of per- the two. A difference is to be expected since the
formance data, see Section 3.2.2.1.) serviceability index is heavily weighted by the
A set of the constants A,, A, and A2 was de- roughness of the pavement while the cracking
termined from the data corresponding to each index depends solely on the amount of crack-
of the following combinations of the load vari- ing and patching. In addition, analytical pro-
able, L2 (axle type), and the design variable, cedures were not the same in the two analyses.
D1 (reinforcing): Longitudinal Cracks at Transverse Joints.—
Single axle vehicles, nonreinforced pave- In the course of the weekly crack surveys, the
ment; points at which longitudinal cracks intersected
Single axle vehicles, reinforced pave- transverse joints were recorded in an effort to
ment; determine whether there was a tendency for
Tandem axle vehicles, nonreinforced such cracks to appear more frequently at one
pavement; and location than another. Figure 128 shows histo-
Tandem axle vehicles, reinforced pave- grams showing the frequency with which longi-
ment. tudinal cracks appeared within each 6-in, inter-
The resulting equations, together with the val cf transverse joint. These graphs represent
average absolute error per observation for each, only failed sections of Designs 1 and 3; and,
are given in the following in a form suitable for of those, only the thinnest pavement sections in
computing log W when C', D2 and L1 are given: each loop. To permit comparison between sec-
tions having 15-ft transverse joint spacing with
For single axle vehicles on nonreinforced pave- those having 40-ft spacing, the number of
ment: cracks actually observed has been converted to
log W = 4.70 + 0.5 log C' 2.62 log
-
the average number per panel. The data are
L1 + 4.84 log D. ± 0.26 (69) further summarized in Table 53, which shows
the average number of cracks per panel with-
For single axle vehicles on reinforced pave- out regard to the location of the crack-joint
ment: intersection.
log W = 4.95 + 0.5 log C' 2.30 log
-
Figure 128 shows that there was a pro-
L1 + 3.57 log D2 ± 0.17 (70) nounced tendency in the 2.5-, 3.5- and 5-in.
TABLE 52
COMPARISON OF CRACKING INDEX, C', FOR REINFORCED AND NONREINFORCED SECTIONS
1
Not significant. 'Average p = 4.15.
2
Significant at 10 percent level. Significant at 1 percent level.
168 THE AASHO ROAD TEST, REPORT 5
_2r.iJL_IIIflhjIIIME\PN
2.5
100,0,
-'.',I..
10.0
$ \
7.5
n
iiinm EMISSIONS
5.0
uIflhIIHuIflhIIIIINLIV
I 2.5
mill0 INERN
IIIIIIlIiIIIIIOIlIIIIIIII1 IRIlIIIIIIIIIIIIIIIIIIIIII1PJ!1I
1)
z
i-UIIIIIIII I 00 øt'
10.0
200
.100
LOOP 2
2.5 PAVEMENT o
0 0 10101 0 I 0 01 0 I 01
0 I 2 3 4 5 6 7 8 9 10 II 12 0 I 2 3 4 5 67 8 9 1011 12
= 900
z
W
M .800.
.500
(I,
I-
400
w
CO
>
.300
4
0
.200
C)
U,
(0
U,
.100
CO
C-,
4 LOOP 3
0 ILL 3.5 PAVEMENT
0 0 0 0 0 0 0 0 0 0
-J
400
a F
.300k
C"- I
I
(a I
.200— .200 -
z I H
.100 .10:
LOOP 4
0 5.0" PAVEMENT
o 0 iii0i0 4'0
.200 -
.10:
~\11w
mz 0
1111 LOOPS
6.5 PAVEMENT
.100 . .100
0 I -kl11 1 [1 [
0 I 0 I 0 l 0 IoI 0 I 0 lOI 0 I 0 I 0 l 0 I
80 LOOP 6
EENT 0 I1]i1
l°I°I°I
0 I 2 3 4 5 6 7 8 9 10 II 12 0 I 2 3 4 5 6 7 8 9 10 Ii 12
X, FEET X, FEET
REINFORCED PAVEMENT NON-REINFORCEO PAVEMENT
Figure 128. Summary of longitudinal cracks intersecting transverse joints of failed sections at time of removal
from test. Data from thinnest sections of each loop, Designs 1 and 3.. Shaded bars represent cracks intersecting
a joint within 3 in. of a point directly above a dowel bar.
170 THE AASHO ROAD TEST, REPORT 5
to
Ix
10 REINFORCED PAVEMENT
(8 SECTIONS)
. POSITION OF DOWEL BAR
5
0
- .1.1.1.1.1.
0 I 2 - 3 4 5 6 7 8 - 9 10 II 12
X.FEET
Figure 129. Summary of position of first crack in failed area of 32 failed sections where first crack was longi-
tudinal, Experiment Designs 1 and 3.
for the particular distributions but without distributions for nonreinforced sections where
success. early maintenance was required in the second
Distribution of Cracking in Failed Sections. panel, the third panel and so forth. The graphs
—In the upper left-hand graph of Figure 132 on the right present similar data for reinforced
average values of the cracking index, C', at p = sections.*
1.5 are shown by panels for the group of nonre- * In cases where no patches were applied to a section,
inforéed sections, Designs 1 and 3, that suffered first failure was determined by other means. Occasion-
ally two panels were damaged simultaneously, and in
early failure in panel No. 1. The remaining these cases panel values of C' were used in more than
graphs on the left in the figure give similar one graph in Figure 132.
25
NON RE INFORC ED
PAVEMENT
31 SECTIONS)
20 DIRECTION OF TRAFFIC
IN
- iRANSVERSE
CRA
10- - -
J-
REINFORCED PAVEMENT 0' 5' 10' IS'
IJ (30 SECTIONS) NONREINFORCED
w
0' 102030 40
REINFORCED
: :
0 5 10 15 0 5 tO 15 20 25 30 35 40
X.FEET X.FEET
Figure 130. Summary of position of first crack in failed area of 61 failed sections where first crack was trans.
verse, Experiment Designs 1 and 3.
RIGID PAVEMENT RESEARCH 171
Figure 132 shows that the panel first requir- to that level, or at W = 1,114,000 unweighted
ing maintenance usually had the highest crack- axle applications if the section survived the
ing index of all panels in the section at the time traffic test. The average pumping index for
the section was removed from test. This panel, failed sections at p = 1.5 was 134; the average
in all probability, was chiefly responsible for for surviving sections at W = 1,114,000 was
the drop in serviceability level to 1.5. The point 34. There was not a clear-cut definition of the
of origin of the first crack leading to such early value of the pumping index associated with the
failures (Figs. 129 and 130) therefore acquires serviceability level of 1.5. For example, one
additional significance. section failed with a pumping index of 5 while
3.2.3.2 Pumping .—The term "pumping," another survived with a pumping index of 209.
unless otherwise noted, refers to the ejection of Nevertheless, the fact that pumping was a very
material from beneath the, pavement and its important factor in the performance of rigid
deposition along the pavement edge. Although pavement sections is demonstrated by the sur-
fine material often was washed upward through vival curve in Figure 134.
joints and cracks, the amount of material trans- No consistent trend was found relating the
ported in this manner at the Road Test was pumping index to subbase thickness, and no
negligible when compared with the amount de- significant difference was found between rein-
posited along the edge. forced and nonreinforced sections. The pump-
In cases where pumping occurred in sections ing index at W = 1,114,000, however, did show
with subbase, the material ejected was the
subbase material (usually including the coarse
fractions) rather than the underlying soil.
Thus, pumping of embankment soil was con-
fined to those sections without subbase and to U,
those sections with 3-in, subbase that pumped z
0
N ON RE IN FOR CED
PAVEMENT
REINFORCED
PAVEMENT
so severely that all granular material had been I-
U (55 SECTIONS) (38 SECTIONS)
ejected prior to failure. w
0
Severe pumping occurred in all loops, but -J
AVERAT
to prevent pumping. By removing the concrete gust 1960, and November 1960. The instru-
from a few failed sections and sampling the ment used was the transverse profilometer
underlying material, it was observed that sub- (Figs. 38 and 39). Elevations relative to the
base material had apparently been removed by centerline were obtained at nine equally spaced
erosive action of water moving across the top points on the pavement at each cross-section
of the subbase, and that the remaining subbase location.
material was relatively undisturbed. There was Figure 136 compares the average October
no evidence, either from giadation tests or 1958 transverse profile of sections with slab
from visual inspection, that fines from the em- thicknesses at the two highest levels in each
bankment had entered the voids in the subbase loop with the average profile of these sections
layer. in October 1959 and November 1960. The edges
Inasmuch as the great majority of the sec- of the pavement, as well as intermediate points,
tions which failed pumped severely prior to apparently moved downward slightly relative
failure, many of these sections would have sur- to the centerline during this period with an ac-
vived the two years of traffic had the subbase companying deformation of the pavement.
material been stabilized effectively to resist ero- However, the data for these sections taken in
sion by water. October and February revealed a reversal of
3.2.3.3 Seasonal Changes in Transverse Pro- this movement in the winters of 1958-59 and
file.—The transverse profile of the rigid pave- 1959-60, apparently due to freezing of the sub-
ment sections relative to the centerline eleva- surface materials (Fig. 137). It is presumed
tion was determined at four locations in each that the longitudinal center joint relieved any
nonreinforced and at six locations in each re- stresses which might have developed in the
inforced section, Designs 1 and 3 (Fig. 135), pavement as a result of these deformations.
in October 1958, February 1959, June 1959, The winter heave of the pavement, measured
October 1959, February 1960, May 1960, Au- at the edges using deflection rods installed ad-
800
G 3 SEj 0
0
2 3 4 5 6 7 8 I 2 3 4 5 6 7 8 2 3 4 5 6 1 2 34 5 6
600 600 AVERAGE OF 5 SECTIONS 600 AVERAGE OF 6 SECTIONS 600 1 AVERAGE OF 5 SECTIONS
AVERAGE OF 13 SECTIONS
z 0 0
C 0
C
0 600 600
AVERAGE OF IS SECTIONS 1 AVERAGE OF 10 SECTIONS
A
1_ AVERAGE F4SECTIONS
600
1 AVERAGE OF S SECTIONS 600
1 AVERAGE OF 9 SECTIONS
0
='L'
PANEL NUMBER
0
REINFORCED PAVEMENT
PANEL NUMBER
200 - I V11
1A 200
O
I 234 56 7 8 I 2 3 4 56 78
PANEL NUMBER PANEL NUMBER
NONREINFORCED PAVEMENT
Figure 132. Distribution of cracking by panels in failed sections, Experiment Designs 1 and 3.
- -
.r.Y
A
Void in shoulder through which subbase material was Subbase material ejected from beneath pavement over-
ejecte(l from beneath pumping slab. night and deposited along edge.
' -
•
- - I
Typical pile of subbase material j)umped from beneath Embankment material pumped from beneath a pavement
pavement, showing change in gradation of material from constructed without a subbase.
fine to coarse in direction of traffic, toward reader.
- ..
.
.- . .
This void beneath pavement extended more than 5 ft Transverse cross-section of a pumping slab showing void
from edge. beneath pavement.
2 2S 0
3
6
6S 0 25* 11* 8 20 4 4
3 5* 17 10 7 5 6
12 6
6 11 34 7 13 2 2
3 12S 3 315* 62* 109 53 17 22 7 18
90 . 19
6 214* 102* 211* 83 18 17 15 18
63 24
9 204* 149* 69 88 12 17 18 18
24T 3 92* 37* 86* 73* 52 37 18 24
118* 22
6 69* 76* 82* 65 51 24 23 31
106* 87
9 88* 103* 101* 146* 45 21 30 27
4 18S 3 189* 191* 47 48 19 24 13 16
72 20
6 116* 91* 92 29 19 24 5 20
57 22
9 98* 147* 117 209 18 21 6 16
32T 3 216* 202* 89 50* 26 35 34 53
152* 29
6 202* 101* 210* 112 41 39 12 28
86* 39
9 75* 118* 116* 178* 32 30 11 27
5 22.4S 3 207* 133* 146* 33 27 22 11 23
32 27
6 104* 301* 63 47 20 52 18 2
97 31
9 193* 203* 79 122" 16 28 4 3
40T 3 91* 108* 127* 37 38 29 17 35
35 31
6 123* 111* 210* 67 61 113 22 0
47 66
9 . 77* 114" 142 98* 27 84 12 12
6 30S 3 122* 150* 18 32 19 15 4 22
83 19
6 237* 159 45 29 27 20 6 20
52 31
9 237* 168 120 59 22 12 1 3
48T 3 95* 164 52 185 26 25 6 53
44 22
6 208* 133 41 36 60 21 20, 46
83 26
9 - 123* 105* 228* 40 86 24 3 22
1
Values with asterisk are for p = 1.5.
= reinforced; N = nonreinforced.
RIGID PAVEMENT RESEARCH 175
TABLE 55
PUMPING INDEX1 FOR SURVIVING SECTIONS, EXPERIMENT DESIGN 1, W = 1,114,000
Single 2
6
12
21
-
11
-
4
79 18 16
18
22.4
30
-
84 21
59
164
:
'13
28
55
10
21 9
Tandem 24 - -
65 42 26
32
40
48
-
101 34
66
149
28
56
69
16
36 25
1 Each value is average for one to eight sections; replicate sections included in averages.
me
jacent to the pavement as references, averaged
about 0.33 in. between October 1958 and March
1959. Measurements were not made for the
90 succeeding winter period.
Figure 134. Estimated probability that a test section * This temperature distribution, or differential, is
with the indicating pumping index will survive 1,114,000 defined in Section 3.3.3 and is referred to as the stand-
axle load applications. ard differential, T.
176 THE AASHO ROAD TEST, REPORT 5
LANE
r I II POSITIONS WHERE CROSS-SECTIONS
:____f -~- /WERE TAKEN
DOWELED I I
LANE JOINT
25
'
60'
OF SECTION
REINFORCED PAVEMENT
T
DOWELED JOINT I II
LANE I I
I II
f I II
LANE2
I II I
I
70' II III
119'
ISO'
creased several fold from afternoon to early rebound deflection. The instrumentation and
morning. Strain and deflection, measured at procedures are described in Section 3.3.1.
a point on the edge at least 7.5 ft distant from Tables 56 and 57 summarize the schedule of
the nearest transverse joint, were also affected routine static and dynamic testing in the traffic
but to a lesser extent (Sections 3.3.3, 3.3.4, loops. Data were gathered by rounds; that is,
3.3.6 and 3.3.7). a measurement crew collecting either static or
The deflection of a corner of a 40-ft rein- dynamic data visited each test section of the
forced panel usually exceeded the deflection of selected factorial experiment only once until
a 15-ft nonreinforced panel, if load, slab thick- all sections in the experiment had been tested.
ness and temperature conditions were the same The data so collected were lumped together into
in both cases (Sections 3.3.5 and 3.3.7). On the one round of data. Successive rounds of data
other hand, deflections and strains measured from the same kind of measurements (dynamic
at a point on the edge 7.5 ft distant from the or static) were numbered consecutively for
nearest joint were not affected significantly by identification (breaks in the sequences occurred
panel length (Sections 3.3.4 and 3.3.6). when rounds were not completed due to equip-
Edge. strains and corner deflections were ment failure or other reasons).
found to decrease in approximately equal pro- As a rule, routine static measurements were
portions with increase in vehicular speed. An made during the daily 5-hr and 20-min break
increase in speed from 2 to 60 mph resulted in traffic, and routine dynamic measurements
in a decrease in strain or deflection of about were made during an 8-hr work shift occur-
29 percent. Though there was considerable ring during the regular 18-hr and 40-min daily
variation in the data, no consistent effect of traffic period. In both cases, the test vehicles
pavement design or load on the percentage rate regularly assigned to a lane were used for the
of reduction was found (Section 3.3.8). tests in that lane. One round of static measure-
Minor (but statistically significant) differ- ments usually required about 1 week. One
ences in the deflection under a 12-kip axle load round of dynamic measurements usually re-
were found from tests made in the various quired from 3 to 4 weeks.
loops and lanes on sections having the same It had been anticipated that some consistent
slab thickness. The differences were consid- trend of strain and deflection with rounds (that
ered to be of no practical significance (Section is, with accumulated axle applications) would
3.3.9). be observed. However, such was -not the case,
Transient effects of the test vehicles which and any..regular trends which might have been
were measured in the pavements included dy- present apparently were masked by the over-
namic edge strain, dynamic corner deflection, riding effect of daily fluctuations in tempera-
static edge rebound deflection and static corner ture and possibly by other unknown variables.
RIGID PAVEMENT RESEARCH 177
NORTH - DISTANCE FROM CENTERLINE.FEET -SOUTH
- - a a
z
_2' ____
/ I-
,/ uJ
IJ I
I,
/ -J
I Ui-
/
// --
I-
8
j,c,
Ui
OCTOBER 1958
OCTOBER 959
0
z
U
NOVEMBER 1960 £ U)
1.2
1.4
Figure 136. Transverse profile of pavement surface relative to centerline elevation before, during, and after the
two years of test traffic. Data from test sections with slab thickness at the two highest levels in Loops 2 through
6, Experiment Design 1, including replicate sections.
1.4
Figure 137. Effect of frost on shape of transverse cross-section of pavement surface. Data from same test sec-
tions used in Figure 136.
TABLE 56
SCHEDULE OF ROUTINE DYNAMIC EDGE STRAIN AND DYNAMIC CORNER DEFLECTION
MEASUREMENTS IN LooPs 2 THROUGH 6
EXPERIMENT DESIGN 1
(Data available in DS 5250)
TABLE 57
SCHEDULE OF ROUTINE EDGE AND CORNER STATIC REBOUND DEFLECTION MEASUREMENTS
IN LooPS 2 THROUGH 6, EXPERIMENT DESIGNS 1 AND 3
(Data available in DS 5280)
TABLE 58 TABLE 59
SUMMARY OF EARLY MORNING ROUNDS, STATIC SUMMARY OF MIDDAY ROUNDS, STATIC
MEASUREMENTS', DESIGN 1, LooPS 3-6 MEASUREMENTS', DESIGN 1, Loops 3-6
(Data available in DS 5280) (Data available in DS 5280)
2 May 1, 1959 0700 0.021 0.028 1 April 5, 1959 1300 0.015 0.018
5 July 11, 1959 0730 0.020 0.025 3 May 26, 1959 1130 0.017 0.022
8 Aug. 19, 1959 0730 0.018 0.023 July 22, 1959 1330 0.014 0.015
6
12 Nov. 4, 1959 0800 0.019 0.027
9 Sept. 2, 1959 1230 0.015 0.019
152 Feb. 14, 1960 0900 0.0102 0.0122
13 Nov. 25, 1959 1300 0.015 0.021
18 Mar. 30, 1960 0800 0.021 0.025
162 Mar. 4, 1960 1330 0.005' 0.007'
21 May 25, 1960 0730 0.021 0.027
17' Mar. 17, 1960 1430 0.006' 0.007'
Average, excluding winter round 0.020 0.026
0.012 20 Apr. 27, 1960 1330 0.017 0.019
Average, winter round 0.010
22 June 15, 1960 1330 0.018 0.021
'Data for thinnest pavement in each loop excluded. Average, excluding winter rounds 0.016 0.019
2
Subgrade frozen. Average, winter rounds 0.006 0.007
12
INSTRUMENTS FOR
OYNAMIC DEFLECTIONS
[
STRAIN GAGE
EDGE OF
,U1CTION BOX
EDGE OF SI1OIJLOER__ \
I)
ENLARGED VIEW OF STRAIN GAGE ASSEMBLY
ç 6 STRAIN GAGE —7
V.00I BRASS FOIL ENVELOPE Fig. 3.3-I Location of Gages, and Transvecse Placement of Vehicle
I for Dynamic Measurements.
EDGE OF PAVEMENT
I 2 - 3 4 5 6
(d)
p ..-. ,.........
-•p..-- -
4- '- - --.': .
'I a-\
- -- .
I .
.:.. ,. ..:-•-
..
- --
- -:-- I
Figure 141. Instrument van taking dynamic measurements. Device on pavement just
ahead of truck measures transverse placement.
182 THE AASHO ROAD TEST, REPORT 5
transverse joints separating the instrumented 3.3.1.2 Static Measurement s—Edge and Cor-
panels. If the centroici of this wheel was 20 ner Detleetion.—An exploratoi'y study in early
in. (-3 in. to -f- 2 in.) from the pavement 1959 indicated that the Benkelman beam could
edge, the measurements were accepted; othei'- be used successfully for measuring the deflec-
wise, they were rejected. The crew remained tion of the free edge of a rigi(i pavement test
at each test section until at least three vehicles section, provided the supports of the beam de-
had succeeded in passing at the specified dis- vice were placed on the shoulder in the position
tance. At least 3 measurements on each of 2 shown in Figures 145 and 146. On the other
strain gages (a total of six values) were av- hand, when the beam was placed on the pave-
eraged to obtain the section strain to be used ment in the position normally used in the flex-
in the analysis. Similarly, at least six measure- ible pavement program, measurements mdi-
ments of corner deflection were averaged to cated that the pavement beneath the supports
obtain a rej)resentative value of deflection for of the device sometimes deflected slightly as
the section.
the truck moved ahead, causing an error in
* This biased tolerance was selected as the result of the measured deflection. Thus, the measure-
special studies of the distribution of the placement of ment of interior deIlections at creep speed was
vehicles whose operators were attempting to drive at
the specified distance of 20 in. from the edge. excluded from the rigid pavement program.
Placement studies showed that edge and cor- order of axle load and axle type (single or
ner deflections were strongly influenced by the tandem). Measurement crews took their sta-
transverse position of the test vehicle and that tions at one or more test sections. If dynamic
accurate control of the transverse placement of measurements were taken, all trucks passed
a moving vehicle was difficult and time consum- over the section or sections being tested at
ing. As a result, a decision was made to I)!aCe either 5 or 35 mph (the most frequently used
the vehicle in position before niaking a meas- speed was 5 mph), and dynamic measurements
urement and to measure the rebound deflection were taken in accordance with the procedures
as the truck moved ahead at slow speed. Place-
ment of the vehicle for edge and corner rebound
deflection measurements is shown in Figure
145.
On each nonreinforced section, static re-
bound corner deflections were measured in
Panels 5 and 7 ( l)aflels were numbered consecu-
tively in the direction of traffic) while the
static rebound edge deflection was measured
near the strain gage on panel 5 (Fig. 147). On
rein forced sections, corner deflections were
taken in panels 4 and 6, and edge deflections
near the strain gage on panel 4. Thus, one edge
deflection and the average of two corner deflec-
tions were available to represent the section for
each round of data.
3..?.2 Strain and Deflection as Llnear Functions
of Load
To determine the effect on strain and deflec-
tion of varying the axle load, other factors re-
maining constant, a series of load studies was
conducted in Loops 4 and 6 in May, June and
September 1 959 during breaks in the normal
test traffic. Tables 61 and 62 give the loads,
vehicular speeds and test sections involved, as
well as other detailed information.
The testing procedure was as follows: The
trucks were assembled on the 1001) in random
Figure 113. Device for calibrating dynamic deflection
* Simultaneously load studies were made on the flex- recording equipment. Inverted depth gages used to
ible tangents as described in Section 2.3.5. move transformer cores through a known distance.
IOINT
LAST AXLE
jected to a series of single (or tandem) axle
loads on a given day were plotted against the
axle loads. Typical graphs are shown in Fig-
ure 148 in which the symbols are defined as
follows:
OF TRAILER = dynamic edge strain, in. X 106;
= dynamic corner deflection, in.;
EDGE OF PAVEMENT7
2d± de' = static rebound edge deflection, in.; and
d' = static rebound corner deflection, in.
REAR AXLE Inspection of plots similar to those shown
in Figure 148 suggested that a straight line
0 OENKELMAN BEAM passing through the origin would usually fit
the data with errors little greater, if any, than
------EE OF SD7 the estimated errors associated with the meas-
U uring system used. Therefore, a linear model
was fitted to the data for each test section-axle
type combination given in Tables 61 and 62.
Yi = a L1 , -I- e (73)
in which
Yi = dynamic edge strain, dynamic cor-
ner deflection, static rebound edge
deflection, or static rebound corner
deflection, measured under the ill
TRANSVERSE JOINT load;
Figure 145. Position of vehicle and Benkelman beam L1 , = the jtb axle load, kips;
for making static rebound deflection measurements.
a = a coefficient assumed to reflect the
conditions of design, temperature
described in Section 3.3.1.1. If static measure- and speed; and
ments were taken, the procedure was as de- ei = the jth residual.
scribed in Section 3.3.1.2. When measurements
under all trucks had been completed at a test The root mean square residuals were com-
section, the measurement crew moved to a new puted separately from the data corresponding
section and the process was repeated. to each line in Tables 61 and 62 and appear in
The measurements made on each section sub- the last column. The residuals were averaged
II
/
Figure 146. Measuring static rebound deflection with the Benkelman beam.
RIGID PAVEMENT RESEARCH 185
NONREINFORCFO PAVFMFNT
IPANEL1' I I
LANEI 2 3 I 4 5 6 7 8 I
1 15
r
LANE 2
L
DIRECTION OF
TRAFC
FI
PANEL I
LANE I I
2 3
_
4 5 [6
LANE 2 40'
...[.
Figure 147. Plan of test sections showing normal locations for measuring static rebound deflection.
40
OESIGP4 9.5" P4ONREINFORCED SLAB-3.0"SUBBASE
SPEED -CREEP
00
DATE -SEPTEMBER 28.1959
rmsr r2.3
z 30 .030
DESIGN -9.5" NONREINFORCED SLA8-6.0"SUBBASE
'C SPEED -CREEP
I- DATE -SEPTEMBER 28,1959
U)
rmsr .O0II
20 .020
I')
U)w
I
5)
10 .010
U
0 0
.030 .030
OESIGN-$.O" MONREIMFORCED SLAB-6.0"SUBBASE DESIGN-aS" NONREINFORCED SLAB-B.O" SUBBASE
SPEED -STATIC SPEED -STATIC
DATE -SEPTEMBER 28,1959 DATE -MAY 28,1959
rmsr = .0016 U) rmsr =0008
Ui
.020 .020
I z
1'
.010 .010
0 0
0 5 10 15 20 25 30 10 20 30 40 50
SINGLE AXLE LOAD, KIPS TANDEM AXLE LOAD. KIPS
Figure 148. Typical results from load studies.
186 THE AASHO ROAD TEST, REPORT 5
TABLE 61
LOAD STUDIES, DYNAMIC STRAIN AND DEFLECTION
Root Mean
Square Residual
Axle Axle Loads Section Si b Date Speed
Thickness (mph) For Corner
Type (kips) Loop Number ,.. Tested For Edge
Strain Deflection
(10-'in./in.) (in.)
TABLE 62
LOAD STUDIES, STATIC DEFLECTION
Si b Root Mean
Axle Axle Load Loo Section Th i ess Date Square Residual (in.)
Type (kips) Number Tested
- At Corner At Edge
puted. As expected, a0 was found to be related From the data summarized in Table 70, it
to design and load and is not considered further was concluded that usually a substantial
in this section. Tabulated average values of a1 amount (from 72 to 88 percent) of the varia-
and r2 from the different analyses are given in tion in corner deflection observed during the
Tables 66, 67, 68 and 69. 24-hr studies could be explained by the tem-
Although a1 varied considerably from section perature statistic T, while from 34 to 42 per-
to section, and from one pavement thickness- cent of the variation in edge strain and deflec-
axle load combination to another, there ap- tion could be accounted for in the same way.
peared to be no consistent trend of this coeffi- As a result, the decision was made to combine
cient with load and pavement thickness, as Eqs. 74 and 75 into the following general model
given by the weighted average values (Tables for use in the analysis of strain and deflection:
66-69). In addition, inspection of the dynamic Strain (or deflection) -
data gathered on May 12 and 13, 1960, for Axle load -
three different levels of vehicular speed (Table
64) failed to reveal any consistent trend of a1 f (design and random variables) x bolT
with speed. (76)
Im THE AASHO ROAD TEST, REPORT 5
From the definition previously given, T, the range, —10 F to +20 F. This range was there-
standard differential, was positive when the fore judged to be representative of the Ottawa
top surface of the standard 6.5-in, slab was area and was chosen for use in the graphs of
warmer than the bottom. Inspection of nearly the equations for strains and deflections given
4,400 hourly observations of T made from in the following subsections.
April 1 through September 30, 1959, revealed
that 98 percent of the values were within the 3.3.4 Dynamic Edge Strain as a Function of
Design, Load and Temperature
Dynamic edge strain data from rounds 4, 5,
TABLE 63
8 and 9 (see Table 56) gathered between April
AVERAGE ROOT MEAN SQUARE RESIDUAL and August 1959 were selected for use in de-
FOR EACH TYPE OF MEASUREMENT termining an empirical relationship between
INVESTIGATED IN SPECIAL LOAD STUDIES edge strain, design, load and temperature.
These rounds covered a representative range
Unit Average Root of spring, summer and fall temperatures and
Measurement System of Mean Square a great majority of the test sections were still
Measure Residual
in good condition, as determined by visual in-
Dynamic edge strain 10-' in/in. 2.5
spection, at the time and point of testing. Data
Dynamic corner deflection Inch 0.0014
from Loop 2 were excluded from the analysis
Static corner deflection Inch 0.0017
after it was found that strain measurements on
Static edge deflection Inch 0.0020
the 2.5-in, slabs (the only thickness instru-
mented in that loop) appeared to be inconsis-
TABLE 64
SCHEDULE OF SPECIAL 24-HOUR STUDIES OF DYNAMIC MEASUREMENTS VERSUS TEMPERATURE
TABLE 65
SCHEDULE OF SPECIAL 24-HOUR STUDIES OF
STATIC MEASUREMENTS VERSUS TEMPERATURE
.100
.090
.080
.070
60 .060
U)
50 .050
U
O
00 z
0
O~
0
40 .040
0
.03C
.02C
.010 0
DESIGN -8.0' NONREINFORCED SLA8-6.0" SUBBASE 0
OESIGN-5.0" NONREtNFORCED SUBBASE
LOAD - 30 RIP SINGLE AXLE LOAD -24 RIP TANDEM AXLE
DATE -JULY 28-29. 1960 DATE - JUNE 18-19, 1959
47 r'r .92
-10 -5 0 +5 +10 +15 +20 -10 -5 0 +5 +10 +15 +20
1, DEGREES FAHRENHEIT
.01
.05
018
.04
U)
Ui
.030 .03
______ __ 0
.020 .02
U)
'a
I
U
010 01
- DESIGN- 6.5"REINFORCED SLAB- 3.0"SUBBASE DESIGN - 6.2" REINFORCED SLAB -3.0" SUBBASE
LOAD - 18 RIP SINGLE AXLE I LOAD -IS RIP SINGLE AXLE
DATE - NOVEMBER 7-8.1929 I DATE -NOVEMBER 7-8, 1959
r1 .51 I r .92
0._ - - - . -
0
IU - C) + +10 +15 +20 -10 -5 0 +5 +10 +15 +20
T. DEGREES FAHRENHEIT T. DEGREES FAHRENHEIT
EDGE MEASUREMENTS CORNER MEASUREMENTS
Figure 149. Typical results from temperature studies.
TABLE 66
AVERAGE VALUES OF a, AND r' FROM DYNAMIC EDGE STRAIN DATA FOR PAVEMENT THICKNESS-AXLE LOAD COMBINATIONS OCCURRING IN
THE 24-HOUR TEMPERATURE STUDIES'
Wtd. avg. —0.0032(4) —0.0052 (3) —0.0021 (6) —0.0011 (5) —0.0014 (7) —0.0013 (4) —0.0021 (29)
0.34
in
'Values in parentheses are number of sections observed.
2
6-, 12-, 18-, 22.4- and 30-kip single axle loads.
0
tJ
TABLE 67
AVERAGE VALUES OF a, AND r' FROM DYNAMIC CORNER DEFLECTION DATA FOR PAVEMENT THICKNESS-AXLE LOAD COMBINATIONS OCCURRING IN
THE 24-HOUR TEMPERATURE STUDIES' 10
Wtd. avg. —0.048 (4) —0.037(3) —0.026 (6) —0.019 (5) —0.019 (7) —0.018 (4) —0.026 (29)
0.88
TABLE 69
AVERAGE VALUES OF a, AND 2 FROM STATIC REBOUND CORNER DEFLECTION DATA FOR PAVEMENT THICKNESS-AXLE LOAD COMBINATIONS OCCURRING
IN THE 24-HOUR TEMPERATURE STUDIES'
0
0 2 4 6 8 10 12 14 0 2 4 6 8 to 12 14
D2 INCHES
D2 .INCHES
Figure 150. Dynamic edge strains, showing effect of temperature and pavement thickness.
tent with the data from the other loops. (Two- For tandem axle vehicles:
thirds of the observations made in Loop 2 were
less than two-thirds, on the average, of the = 3.814
corresponding values predicted by Eq. 78.) (79)
L1 1000035T iD20 -8523
Plots of the data, together with information
furnished by the special load studies (Section Graphs of Eqs. 78 and 79 are shown in Figure
3.3.2) and the 24-hr temperature studies (Sec- 150. Residuals that are less than the average
tion 3.3.3), led to the selection of the following root mean square residual determined in the
model for use in the analysis: two analyses correspond to observations that
range from 83 to 120 percent of the predicted
A0 values. The coefficients of T were found to be
-- (77) 0.0031 and 0.0035, as compared to a value of
L1 - 10TD2 Az
0.002 determined from the special 24-hr tem-
in which e is dynamic edge strain x 10 in. per perature studies (Table 70).
in.; A0, A1 and A2 are positive constants to be
determined from the analysis; and the other 3.3.5 Dynamic Corner Deflection ds a Function
symbols are as previously defined. of Design and Load
The analysis indicated that the design vari- Of the four kinds of measurements discussed
able, reinforcing (or slab length), was not in Section 3.3.3 the least satisfactory (both
significant. The following equations resulted: from the standpoint of the time involved in
making the measurements and certain unex-
For single axle vehicles: plained reversals of trend in the data from
section to section and from round to round)
20.54 was the measurement of dynamic deflection.
-- (78)
- 1000031 T D21278 Since the same instrument van and crew were
used for measuring both dynamic deflection and
strain, it was decided early in the program that
TABLE 70
dynamic deflection would -be measured only
occasionally, and that the men and equipment
SUMMARY OF AVERAGES OF a, AND r2 FOR available would be used primarily in making
24-HOUR TEMPERATURE STUDIES strain measurements that could be accumu-
lated at a relatively rapid rate. Meanwhile,
Measurement
Average for ample deflection data were gathered by a dif-
All Thicknesses ferent crew making static rebound measure-
and Speeds
ments with the Benkelman beam (see Table 57
Point Type a1 r2 and Sections 3.3.6 and 3.3.7).
Of the dynamic deflection data available,
Edge Dynamic strain —0.002 0.34 those from round 9 (gathered in July and Au-
Static deflection —0.009 0.42 gust 1959 between 7:00 p.m. and 6: 00 a.m.)
Corner Dynamic deflection —0.026 0.88 were selected for analysis. Data from the 3.5-
Static deflection —0.013 0.72 in. and 5.0-in, pavements were excluded from
the analysis because they were inconsistent
RIGID PAVEMENT RESEARCH 193
with data from the remaining six thickness for analysis.* The dates of individual rounds
levels tested. However, they were includedin and the hours during which the measurements
the error computations. Following is the gen- were made in each round are given in Table
eral model chosen to represent the data: 57. Measurement methods and procedures are
described in Section 3.3.1.2.
= A0 D2'4 (80) The model selected for use in the analysis,
L1 based on Eq. 76 and plots of the deflection data,
is
in which d is dynamic corner deflection in A0
inches; A0 and A, are constants to be deter- (85)
mined from the analysis; and the other symbols L1 lOA,T DS AZ
are as previously defined. in which d'e is static rebound edge deflection, in
Four separate equations resulted from the in.; L1, T, and D5 are as previously defined; and
analyses as follows:
For single axle vehicles on nonreinforced pave-
ment:
- - 0.0214 .004
ment:
.002
- - 0.0256
L1 - D2 1374
TANDEM AXLE LOAD
.001
REINFORCED PAVEMENT
For tandem axle vehicles ° on nonreinforced NONREINFORCED PAVE
pavement:
= 0.00521 6 8 10 IS IA
(83)
L1 D2 0.870
02 INCHES
.003
T. _I0_______.__\
SINGLE AXLE VEHICLES TANDEM AXLE VEHICLES
.002
0.
U,
Ui
I
U
.001
-J
4 6 8 to 12 14 0 2 4 6 8 10 4
02. INCHES 02. INCHES
Figure 152. Static rebound edge deflection, showing effect of temperature and pavement thickness.
for several values of the temperature statistic analyses correspond to observations that range
T. The values of the coefficient of T, 0.0075 from 73 to 137 percent of the predicted values.
and 0.0100, found in the analyses compare The values, 0.011 and 0.015, of the coeffi-
favorably with the value 0.009 determined in cient of T in Eqs. 88-91 may be compared with
the special 24-hr temperature studies (Table the value, 0.013, found from the special 24-hr
70). temperature studies (Table 70).
Figure .153 shows graphs of the four equa-
3.3.7 Static Corner Deflection as a Function of tions for a range of values of the standard dif-
Design, Load and Temperature ferential, T. The deflection at the corner of
40-ft panels usually exceeded the deflection at
Static rebound corner deflectiOn data selected the corner of 15-ft panels other factors remain-
for analysis were taken from the same rounds ing constant, as was the case also for dynamic
as the static rebound edge deflections (see Sec- corner deflections (Section 3.3.5). On the other
tion 3.3.6). The same general model was also hand, static edge deflection, measured at a
used. As in the case of dynamic corner deflec- point at least. 7.5 ft from the nearest joint, was
tions (see Section 3.3.5) the design factor, re- found not to be significantly affected by panel
inforcing and/or panel length, was found to length (see Section 3.3.6). It is concluded that
have a significant effect with the result that for pavements constructed with different spac-
four equations were developed as follows: ings of transverse joints, comparisons of de-
For single axle vehicles on nonreinforced pave- flections can better be made at a point at some
ment: distance from a joint than at a panel corner,
unless the object of the comparison is to dif-
0.013 ferentiate between joint designs or spacings.
(88)
L1 10•0 T D2 '-' 5
3.3.8 Strain and Deflection as Functions of
For single axle vehicles on reinforced pave- Speed
ment: Studies of the effect of speed on dynamic edge
0.013 strain and dynamic corner deflection were con-
ducted in Loops 4 and 6 in August, September
L1 - 100015T D21-19 and December 1959. * Table 71 gives the loads,
nominal speeds and the number of sections of
For tandem axle vehicles on nonreinforced each thickness tested on the six study days.
pavement: Strains and deflections were measured with
0.00443 the equipment and in accordance with the pro-
cedure described in Section 3.3.1.1. In addi-
L1 - 10001IT D20900 tion, the following special procedures were
followed:
For tandem axle vehicles on reinforced pave- Several single axle vehicles of the specified
ment: loadings were assembled on Loop 4 or 6 and
0.00443 arranged in random order of load. The vehicles
L, - 100 .6151 D20767 then proceeded around the loop at one of the
selected speeds, measurements meanwhile be-
In these equations the symbol d' repre- ing taken at one of the test sections. The ve-
sents static rebound corner deflection in inches, hicles continued to circle the loop until at least
and the other symbols are as previously de- three readings of the section gages had been
fined. obtained for each axle load with the vehicles at
Residuals that are less than the average root * Simultaneously, speed studies were made on the
mean square residual determined in the four flexible tangents as described in Section 2.3.6.
RIGID PAVEMENT RESEARCH 195
the proper transverse placement (see Figure a0 = a coefficient presumed to depend upon
138). Then a new speed, selected at random, the conditions of design, load and
was assigned the vehicles, and the process re- temperature under which the data
peated until the section had been tested under were collected; and
all combinations of nominal speed and axle a1 = a constant presumed to represent the
load. Actual speeds were calculated from the speed effect.
recorded traces of the output of the gages and
averaged over the three or more readings taken For each of the 124 sets of data a value of
for each nominal speed-axle load combination. a0, a1 and f2 was computed. The coefficient, a0 ,
This procedure was repeated for each test sec- as expected, was obviously related to design
tion; the over-all time required was about 6 hr and load; a1 exhibited no consistent trends with
per loop. any known variables. The squared correlation
For each axle load-test section combination coefficient f2 averaged 0.74 for the strain data
occurring on a given day in Table 71, a table and 0.71 for the deflection data. Table 72 gives
was prepared listing simultaneous values of average values of a2 for strain and deflection
the observed data (strain or deflection) and data corresponding to each load-pavement
speed (typical graphs of the data are given in thickness combination. The mean of all values
Fig. 154). To smooth the data (as well as to of a1 computed from strain data differed from
separate the speed effect from that of load, de- the mean value computed from corner deflec-
sign, and temperature differential) a two- tion data by only 0.0004. In view of the vari-
parameter model was fitted to the data from ability of the data, this difference was not con-
each of the 124 tables representing the results sidered significant and an average value of a1
of the studies. The model, suggested by plots of 0.0026 was taken to represent the effect of
of the kind shown in Figure 154, was as fol- speed on either dynamic edge strain or dynamic
lows: corner deflection.
The effect of speed on strain or deflection can
Y = 10au+av (92) be demonstrated as follows:
in which
With a1 = — 0.0026, Eq. 92 becomes
Y = strain, in 10.6 in. per in., or deflec-
tion, in in.; Y = 10 (a, -0.00260) (93)
v = speed, mph, within the observed R is chosen to represent the estimated percent-
range of 2 to 60; age reduction in edge strain or corner de-
.004
-10.
NONREINFORCED NONREINFORCED PAVEMENT
.003
T0—
CL
In
.002 —T+I0
U
T+20
.00I
.004
REINFORCED PAVEMENT
.003
T -I0------
-T+ I0
T = + 20I___ I
0
0 2 4 6 8 10 12 14 0 2 4 6- 8 10 12 14
02 . INCHES D2 , INCHES
SINGLE AXLE VEHICLE TANDEM AXLE VEHICLE
Figure 153. Static rebound corner deflection, showing effect of reinforcing (or panel length), temperature, and
- pavement thickness.
196 THE AASHO ROAD TEST, REPORT 5
20
1
30 40 50
I
60
somewhat beyond the time when the weaker
'/ • MILES PER HOUR sections began to fail, data from all sections
Figure 154. Typical results from speed studies. of the thinnest level of pavement thickness in
each loop have been excluded from the aver-
ages. No consistent long-time trends in deflec-
flection for a rigid pavement test section of any tion level are apparent, as was also the case
design, as a single axle load increases in speed for the average deflection data in Tables 58,
from 2 mph to the speed v where v < 60 and 59 and 60.
the temperature differential T remains fixed A portion of the data from round 1 (April
during the period of acceleration. Under these 1959) was selected to represent the deflection
conditions a0 is fixed and level of the pavements prior to the development
of excessive pumping. To permit sound com-
R = 100 (1 - 1.012 x 10-0002 ) ( 94) parisons across two or more loops it was neces-
Figure 155 plots R against v showing that sary to limit the study to designs which were
an increase in speed from 2 to 60 mph resulted common to those loops. A choice was made of
in a decrease in strain or deflection of about 29 the 6.5- and 8.0-in. sections common to Loops
percent. 3, 4 and 5, and of the 8.0- and 9.5-in. sections
TABLE 71 -
SCHEDULE OF SPEED STUDIEs
4U N N
30
CC
bs C> <=
00
00 00 i3 Lt)
CC V C. C)
CC CC CC CC
cD' 0 0 0 0
q c
40 50 60 z -1 0 0 0 0
2 10 20 30
I I I I
V. MILES PER HOUR
Figure 155. Percentage reduction in edge strain or 0
corner deflection with increase in vehicular speed. 0
0
Cl)
Z
C C)CCaC ',-4CC
TABLE 73
SCHEDULE OF SPECIAL STATIC REBOUND DEFLECTION STUDIES
WITH 12-Kip SINGLE AxLE LOAD, EXPERIMENT DESIGN 1
(Data available in DS 5283)
R N R
_____
N
dicted with satisfactory accuracy from the
average of 24 dynamic edge strains measured
3 I 2 I I under the single axle load regularly assigned to
I 6 2 I I I
SECTIONS
IN
that section (Fig. 156). Similar predictions
9 I 1 I I DESIGNA could be made from static edge and corner de-
3 I 2 I I flections, but with somewhat less accuracy
2 6 2 I I I (Figs. 158 and 160).
9 I I I I One of the objectives of the Road Test was
3 2 I I 2 I I the investigation of the assumption that the
I 6 I 2 2 I I I life of a rigid pavement subjected to repeated
- 9 I I 1 I I I applications of the same load can be satisfac-
4 3 2 I 1 2 I I torily predicted if the strain or deflection of the
2 6 I 2 2 I I I pavement, before visible deterioration of the
9 I I I I I I pavement takes place, is known.
3 I I 2 I I 2 To test this assumption, performance data,
I 6 I I I 2 2 I in the form of log W to a serviceability level of
9 I I 1 I I I 2.5, were compared with data from early
5 3 I I 2 I I 2 rounds of dynamic edge strain, static rebound
2 6 I I 2 2 I edge deflection, and static rebound corner de-
9 I I I I I I flection.
3 I I 2 I The rounds selected for the analyses of per-
I 6 SECTIONS I I I 2 formance as a function of strain or deflection
IN
6
g
DESIGN B
I I I were the same as those used in the analyses of
— j--- — j--- — j— — j---- strain and deflection as function of design and
2 6 I I I 2 load. They are given in Sections 3.3.4, 3.3.6
and 3.3.7. -
RIGID PAVEMENT RESEARCH 199
TABLE 75
Ms&N DEFLECTIONS FROM SPECIAL STUDY
WITH 12-Kip SINGLE AXLE Lotr
(Data from Round 1)
Inasmuch as the performance analysis of De- Three sets of values of A0 and A, were obtained
sign 1 sections had shown that the effects of corresponding to the three kinds of measure-
the design variables D, (reinforcing) and D, ments analyzed. The resulting equations are
(subbase thickness) were not statistically sig- given in the three subsections immediately fol-
nificant (see Section 3.2.2.1), these variables lowing.
were ignored in this analysis. As a conse-
quence of experience gained in the performance 3.4.1 Performance from Dynamic Edge Strain
analysis, data from the highest level of slab Dynamic edge strain data from each group
thickness in each loop were also excluded. of six sections (Table 76) were averaged with-
Furthermore, a preliminary analysis having out regard to the temperature conditions at the
shown that the relationship between strain or time the individual measurements were made.
deflection and the performance of test sections The resulting value was paired with the value
subjected to tandem axle vehicles was not well of log W2. 5 for the same sections, if all six sec-
defined, data from the tandem axle lanes were tions in the group fell to that serviceability
omitted. Finally, Loop 2 data were excluded level during the traffic test, and plotted in
because of the small volume of data from failed Figure 156 as a solid disc. If the serviceability
sections. The resulting reduced experiment de- level of some of the group of six sections was
sign is given in Table 76. Data representing 12 above 2.5 at the end of the traffic test, then Eq.
design-load combinations involving a total of 62 was used to estimate log W at p = 2.5 for
72 test sections, went into each of the three each section, these values were added to the
analyses. observed values to arrive at an average for the
The general model chosen for the analyses six section group, and the result plotted as a
was the following: triangle. In each case where all six sections
ended the traffic test with serviceability levels
log W2.5 = A0 + A 1 log Y (95)
greater than 2.5, the value of log W2. 5 was
in which estimated for the corresponding load-slab
thickness combination from Eq.' 62 and plotted
log W205 = the logarithm of the number of as an open circle.
unweighted axle applications at
which p = 2.5, averaged over
the six sections occurring in a TABLE 76
design-load combination in Table EXPERIMENT DESIGN FOR ANALYSIS OF PERFORMANCE
76; and AS A FUNCTION OF STRAIN AND DEFLECTION
When Eq. 95 was fitted to the strain data the graph on the left shows the percentage of all
following relationship resulted: sections having an initial dynamic edge strain
falling within a specified interval that ended
log W2.5 = 13.35 - 4.66 log 6 (96) the traffic test at a serviceability level greater
than 1.5. The strain values were the individual
in which loge is the logarithm of dynamic edge round data for individual sections; therefore,
strain (in 106 in. per in.) averaged over the each section is represented by four values not
six sections occurring within a load-slab thick- necessarily all the same. Thus, the curve drawn
ness combination, and again averaged_over the through the tops of the bars can be regarded
four rounds of data analyzed; and log W2 . 5 is as the estimated probability (given a single
as defined in Section 3.4. edge strain determination) that a section will
The straight line (Fig. 156) is a plot of Eq. survive 1,114,000 applications of a single axle
96. It was concluded that the average of a load equal to that which caused the strain, and
large number of measured strains could be within the range of 12,000 to 30,000 lb. A
satisfactorily used to predict the average life similar interpretation can be made of the curve
of test sections of a given thickness subjected on the right, except that the load involved is a
to repeated applications of a single axle load tandem axle within the range, 24,000 to 48,000
of constant weight. lb.
Histograms (Fig. 157) were plotted from
strain and performance data for all sections 34.2 Performance from Static Edge Deflection
involved in Design 1, Loops 3 through 6. The Because the effect of temperature on dy-
10.00C
800C 05
600C rrA:r
4000 A VA TAA
rao
300C
VIA
MME
10,00
2000 Vi####AWA$U•Ø1
Ell
Cr
(I) 100C
01 800
ow
I0 600
Cr
400 L XIPIFU
200
!thh1muII ZQ 30 -6 40 50 60 70 80
E • STRAIN 10
(AVERAGE OF SIX SECTIONS. FOUR OBSERVATIONS PER SECTION)
Figure 156. Performance as a function of dynamic edge strain data from Design 1,
Loops 3 through 6.
RIGID PAVEMENT RESEARCH 201
100
SINGLE TANDEM
90 AXLE AXLE
z
w
0 80
w
°: 70
1
60
>
50
040
30
o 20
- 10
U
QL
0 0 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70
£.STRAIN X 100
Figure 157. Estimated probability that a test section will survive
1,114,000 axle load applications, based on measurements of dynamic edge
strain. Bars plotted from strain and performance data, Experiment
Design 1, Loops 3 through 6.
namic edge strain was found to be relatively The three symbols (a solid disc, a triangle and
small (see Fig. 150) and thus could probably an open circle) are as defined in Section 3.4.1.
be neglected without large error, it was con- Comparing Figure 158 with Figure 156 in-
sidered desirable to adjust all deflections to a dicates that generally a greater error would
common temperature condition prior to deter- result in predicting the average life of sections
mining the relationship existing between de- from static rebound edge deflection than from
flection and performance (see Figs. 152 and dynamic edge strain.
153). The adjustment of edge deflections was Figure 159 was plotted from edge deflection
made on the basis of the coefficient of the tem- data, but is otherwise identical to Figure 157.
perature statistic, T. determined in the analy- Therefore, the curves may be used to estimate
sis described in Sections 3.3.3 and 3.3.6. the probability (given a single edge deflection)
If d'e is an estimate of the deflection which that a section will survive 1,114,000 applica-
would have been observed had T been equal to tions of either a single or a tandem axle load
zero (all other factors remaining the same), it equal to that which caused the deflection, and
can be shown that, according to Eq. 86, within the range covered by the data.
d'e d'e X 10075r (97) 3.4.3 Performance from Static Corner Deflec-
tion
Eq. 97 was used to adjust all deflections in
the analysis to a common temperature condi- By procedures paralleling those described in
tion, T = 0. The adjusted values were then Section 3.4.2, the following equation connecting
treated in the same way as were the strain performance with static rebound corner deflec-
values (Section 3.4.1), and the resulting equa- tion was derived:
tion was as follows:
log W2.5 = 0.95 - 3.29 log d' (99)
log W 2.5 = 0.74 - 3.15 log d' (98) where
log d' = log t',, averaged over the
in which log ce is log d'e averaged over the six six sections occurring with-
test sections occurring within, a design-load in one load-slab thickness
combination and again over the_eight rounds of combination, and again
data being analyzed, and log W2.5 is as defined over the eight rounds of
in Section 3.4. data being analyzed; and
Figure 158 shows the graph of Eq. 98 as = d' adjusted to the tempera-
well as plotted values of the averaged data. ture condition, T = 0.
10.0C(
0
80CC
60CC
VA VVA VAVA
40CC
30CC r;vrwArarriAr4
m
rs",
z
I
200C
(I)
I.I1I.J.
-
F.!." 'it!'I.1.I1WE;I1I--
30(
20( 'U UI
MEN
10001
iiièt .02 .03
d. INCHES FOR TC
04
Figure 158. Performance as a function of static rebound edge deflection data from
Design 1, Loops 3 through 6.
TANDEM
AXLE
C .01 .02 .03 .04 .05 .06 0 .01 .02 .03 .04 .05 .06 .07
d. INCHES
Figure 159. Estimated probability that a test section will survive
1,114,000 axle load applications, based on static rebound edge deflection.
Bars plotted from edge deflection and performance data, Experiment
Design 1, Loops 3 through 6.
RIGID PAVEMENT RESEARCH 203
Equations for adjusting static rebound axle load equal to that which caused the deflec-
corner deflections to the temperature condi- tion, and within the range covered by the data.
tion, T = 0, based on Eqs. 88 and 89 are as
follows: 3.5 AUXILIARY STUDIES
For nonreinforced sections: 3.5.1 Overlays
d''c = d'c 1000112' ( 100) This subsection presents tabulated data from
a special study of 18 test sections overlaid with
For reinforced sections: asphaltic concrete at the time their serviceabil-
= d' 100015 T (101) ity level had declined to. 1.5. Less than one-
third of these overlays survived the traffic test-
Figure 160 shows a graph of Eq. 99 and plots ing without further major maintenance. The
of the data. Comparing Figure 160 with Fig- volume of data was insufficient to permit deriv-
ure 156 indicates that errors in prediction of ation of performance equations.
performance based on static rebound corner
deflections would generally exceed errors in When the serviceability level of rigid pave-
predictions based on edge strains. ment test sections fell to 1.5, the pavement was
Figure 161 shows the estimated probability either removed and the section rebuilt with
(given a single determination of static rebound flexible pavement materials, or it was overlaid
corner deflection) that a section will survive with asphaltic concrete. The latter. procedure
1,114,000 applications of a single or tandem was followed in a majority of the cases.
'0.0OC
8 OOC
600C MEN
VJAVVAA
' s
400C
300C MMMErN
" E
U)
200C Erma g,U no6
VVV
N
-.
Cr z
IOOC
Ou)
(I)
80C
22.4
60C
ZL
-<
Cr
•
I
III1!.10IIIII8Ifl. S
30(
20(
I1I.
- UI
.02 .03 .04 .05 .Ob .01
d INCHES. FOR T=0
II Vts
Figure 160. Performance as a function of static rebound corner deflection data from
Design 1, Loops 3 through6.
204 THE AASHO ROAD TEST, REPORT 5
tOO
SINGLE TANDEM
90 AXLE AXLE
w
80
U
j 70
60
50
40
,-
30
U
D
20
U
'• 10
ot
0 .01 .02 D3 .04 .05 .06 .07 JD8 9 0 .01 .02 .03 .04 .05 .06 .07 .08 .09
d INCHES
Among the overlaid sections, 18 were selected ture and plate bearing tests conducted on the
for special study involving the measurement of subbase and embankment materials beneath
areas of cracking and patching, rut depths and selected pavements in the four main loops as
slope variance. These .data were used in de- well as in the non-traffic loop.
termining the serviceability index for each in- In the non-traffic loop the CBR and elastic
dex day in accordance with the formula modulus of both materials tended to be some-
applying to flexible pavements (Section 2.2.1). what greater in summer than in spring periods,
though the moisture content and density were
Table 77 gives design information pertaining practically unchanged.
to each of the selected sections and overlays, In the traffic loops, tests made in spring 1960
as well as basic performance data. Table 78 yielded results not appreciably different from
groups data for sections of the same slab those obtained in the non-traffic loop for the
thickness subjected to the same load, and aver- same period, except that the average elastic
ages essential performance data. modulus of both subbase and embankment was
The date shown in the last column of Table slightly greater in the traffic loops.
77 is when the study of an overlay was termi- 3.5.2.1 Strength and Condition Data, Loop
nated because of the necessity for extensive 1.—A special study of the variation of subsur-
patching or the placement of an additional face strenth and condition with time in the
overlay. Only 5 of the 18 overlays, or 28 per- absence of traffic was conducted in the four
cent, survived the period of traffic testing. sections of Design 5 in lane 2 of Loop 1. All
However, sections of the Road Test were not four sections were nonreinforced, with a
overlaid until the serviceability level had fallen common slab thickness of 5 in. Two sections
to 1.5. This level is lower than the service- had a 6-in, subbase; the others, none.
ability at which most pavements in actual The sections were 120 ft in length and were
service are overlaid. divided into 5- by 12-ft subpanels. Subpanels
An analysis of overlay performance data were selected in random order for trenches
leading to the development of performance that were cut at periodic intervals (except in
equations similar to those pertaining to rigid winter) beginning in spring 1959. Plate load,
pavement test sections (Section 3.2.2) was at- CBR, moisture content, and density tests were
tempted but the results were not considered made on the subbase and the embankment. In
reliable because of insufficient data. each subpanel two CBR, moisture content, and
density tests were made in each wheelpath (3
3.5.2 Subsurface and 9 ft from pavement edge) on the subbase
This subsection presents in graphical or and embankment. The moisture contents and
tabular form the results of CBR, density, mois- densities were determined at a depth of 0 to 4
TABLE 77
BASIC DATA, OVERLAY STUDY, RIGID PAVEMENT
Crack-
Marshall Data Axle ing
Overlay Application Rut and Serviceability
(1,000's) Depth Patch- Date
Axle Original at
Loop Load Design Voids _____________ at of
End ing at End of
(kips) (in.) End of Over-
BeforeOver-
After End of Study
Thick- Date Asphalt Stabil- Flow After Study
To Over-
Study
ness Content ity Total Filled (in.) (sq f lay Study
(in.) lay 1,000 lay
sqft)
Q
2 6S 2.5-0-1 2 2/16/60 4.6 2554 14 3.76 74.80 485 249 0.05 67 1.3 2.4 1.5 6/15/60 10
6S 2.5-0-0 2 3/23/60 4.4 2357 13 6.04 63.65 568 166 0.05 133 1.1 1.4 2.9 6/15/60
3 12S 5.0-6-1 3 6/ 1/60 4.5 2433 10 7.44 58.62 734. 379 0.05 197 1.3 3.2 3.2 12/ 1/60
4 18S 5.0-6-0 3 12/ 3/59 4.4 2199 11 4.72 69.40 354 420 0 144 1.4 3.2 1.8 6/15/60 t
32T 5.0-9-1 3 1/16/60 4.6 2052 11 5.90 64.91 413 700 0.15 82 1.5 3.3 1.8 12/ 1/60
18S 5.0-3-1 3 1/16/60 4.6 2052 11 5.90 64.91 413 361 0.15 173 1.1 3.4 3.1 6/15/60
32T 6.5-6-1 3 6/27/60 4.4 2081 10 5.01 66.92 803 310 0.10 133 1.4 3.0 1.9 12/ 1/60
18S 5.0-0-0 3 12/ 3/59 4.4 2199 11 4.72 69.40 354 282 0 145 1.4 3.0 1.2 4/20/60 in
32T 6.5-9-0 3 6/ 2/60 4.4 2846 10 748 58.11 735 378 0.05 167 1.0 3.1 2.8 12/ 1/60
32T 6.5-3-1 3 6/27/60 4.4 2081 10 5.01 66.92 803 310 0.10 125 1.3 3.3 2.9 12/ 1/60
C)
5 22.4S 6.5-6-1 3 12/14/59 370 562 0.10 164 1.5 3.0 1.9 8/24/60
40T 6.5-3-0 3. 12/ 1/59 4.4 2358 11 5.56 . 65.20 349 321 0.05 139 0.8 3.4 1.6 5/ 4/60
40T 6.5-6-0 3 12/16/59 370 264 0.15 109 0.6 0.9 1.5 4/20/60
40T 6.5-9-1 3 4/15/60 4.5 1943 10 6.38 62.77 622 48 0.10 209 1.5 3.4 2.4 6/15/60
6 48T 8.0-3-1 3 4/18/60 4.6 1955 11 5.98 64.65 615 143 0.10 152 1.3 3.4 2.7 6/15/60
48T 8.0-9-1 3 4/16/60 4.6 1890 10 7.44 59.28 615 68 0 98 1.7 2.7 2.2 6/15/60
30S 8.0-0-0 3 6/ 1/60 4.5 2433 10 7.44 58.62 735 39 0.20 113 1.4 2.9 2.7 6/15/60
48T 8.0-6-1 3% 1/20/60 4.4 2357 11 5.92 64.14 416 342 0.10 71 1.6 3.0 2.1 6/15/60
'Before overlay.
206 THE AASHO ROAD TEST, REPORT 5
Cd
ci ,-I ,-I C'1 C') ,-I C') ,-4 C') This subsection deals with the vertical dis-
placement, in the absence of loads other than
gravity, observed at selected points on the sur-
face of concrete slabs. It treats only the maxi-
mum displacement of these points, occurring
0.
C') co •,)4 IL) CC' as the temperature of the air changes from a
14
* 30-in, diameter plate.
RIGID PAVEMENT RESEARCH 207
maximum to the next minimum, or from a the standard differential are also related, by
minimum to the next maximum. means, of an equation, to corresponding changes
As background information, graphs are pre- in air temperature (Fig: 174).
sented showing changes with time of the Additional equations are given relating the
temperature distribution in plastic and hard- displacement of a point on the pavement sur-
ened concrete slabs (Figs. 167, 168 and 169). face to the coordinates of the point and the
Changes in the standard temperature differ- thickness of the slab. From these equations, or
ential (see Section 3.3.3 for definition) 'are from their graphs (Figs. 177 and 179), esti-
correlated with corresponding displacements mates of corner displacements approaching
observed at the corners of panels embracing a the greatest observed at the Road Test may be
wide range of desigTls (Fig. 173). Changes in made.
10 3'+
32
C—
z
'a,
Li
0130
18 114
.112
C—
z
0
C-)
Ui
r 16 S
C-
CO
0
S •
.
S S
8
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
SUBBASE
. . SUBBASE
S
126
z
cr S C)
1)4 — 80
a-Li
cc I I II I II I I I I II
—J 40
160
I I I I I I I I I I I I I I I I II —J I I I I I I I I I I I I I I
EMBANKMENT 0
0 EMBANKMENT
W2 • 120
D
—J S C-)
S• ICO— •
. . . S
—J
Li S
au 80
C-)
S
C I I I II I I I I I I I III I 40
JAI'L APR. JULY OCT. JAN. APR. JULY OCT JAN. APR. JULY OCT JAN. APR. JULY OCT
Between
Outer Wheelpath Inner Wheelpath Wheelpaths
Sect. Design Test
Loop Moisture Dry Moisture Dry Location
No. Code2 Moisture Dry 74
Content Density CBR3 k E4
Content Density CBR' Content Density
(%) (pcf) (%) (pcf) (%) (pcf)
3 192 5R-6 7.0 135.7 - 94 7.3 134.9 - 101 7.1 139.8.. Joint
226 5N-3 7.7 133.7 2.8 107 7.8 135.2 3.0 103 6.8 139.4 Joint
cI
Mean 7.4 184.7 2.8 101 7.6 135.1 3.0 102 7.0 139.6
0
4 643 5N-3 7.1 134.2 6.7 85 7.2 132.5 4.4 85 7.2 132.6 Joint
-
5j
673 5R.-9 6.8 137.0 - - 124 7.8 137.2 117 7.3 133.9 Joint 0
673 5R-9 6.5 140.9 - 113 8.6 131.7 - 135 7.2 135.8 Panel
Mean 6.8 187.4 6.7 107 7.9 133.8 4.4 112 7.2 184.1
5 490 6.5N-6 7.9 131.8 - 78 7.2 142.4 - 97 7.9 134.4 Joint
550 6.5R-9 5.6 141.1 - 107 5.3 138.6 8.8 5.5 138.2 Joint
550 6.5R-9 7.9 135.1 - 105 6.9 137.0 - 123 7.5 130.0 Panel 0
Mean 7.1 136.0 - 97 6.5 139.3 8.8 110 7.0 134.2
6 348 8R-9 10.2 133.1 - 122 10.8 - - 113 9.0 135.8 Panel
Mean 10.2 133.1 - 122 10.8 - - 113 9.0 135.8
Over-all mean 7.9 135.3 4.8 107 8.2 136.1 5.4 109 7.6 135.9
1
Tests made from April 23 to May 25, 1960.
2
First figure is slab thickness, in inches; letter designates reinforcement; last number is subbase thickness, in inches.
'Penetration at 0.1 in.
30-in, plate.
TABLE 80
CONDITION DATA FOR EMBANKMENT, RIGID PAVEMENT TANGENTS
(from Spring Trench Program', 1960)
3 192 5R-6 14.0 112.0 1.3 76 15.1 111.5 1.3 82 16.0 109.2 Joint
226 5N-3 15.6 111.4 2.2 90 15.0 108.8 1.9 98 15.3 110.6 Joint
Mean 14.8 111.7 1.8 83 15.1 110.2 1.6 90 15.7 109.9
4 643 5N-3 17.9 104.6 1.2 70 18.3 104.7 1.5 63 17.5 104.8 Joint
673 5R-9 14.6 111.0 1.9 81 16.6 109.9 1.0 77 13.9 112.6 Joint
Mean
673 5R-9 16.8
16.4
109.8
108.5
0.7
1.3
68
73
14.4
16.4
109.2
107.9
2.2
1.6
90
77
14.4
15.3
109.8
109.1
Panel
z
5 490 6.5N-6 19.4 105.2 1.1 64 17.0 108.0 1.0 69 17.9 105.5 Joint
550 6.5R-9 15.0 111.8 2.0 84 15.5 112.3 - - 15.8 111.9 Joint
550 6.5R-9 15.9 108.6 1.6 84 15.5 110.4 2.5 96 15.8 111.8 Panel
C)
Mean 16.8 108.5 1.6 77 16.0 110.2 1.8 83 16.5 109.7
6 348 8R-9 13.4 114.4 2.4 105 12.4 114.7 2.3 97 13.0 114.8 Panel
Mean 13.4 114.4 2.4 105 12.4 114.7 2.3 97 13.0 114.8
Over-all mean 15.4 110.8 1.8 85 15.0 110.8 1.8 87 15.1 110.9
1
Tests made from April 23 to May 25, 1960.
2
First figure in slab thickness, in inches; letter designates reinforcement; last number is subbase thickness, in inches.
'Penetration at 0.1 in.
30-in, plate.
210 THE AASHO ROAD TEST, REPORT 5
Subbase
200 4.0 200
SPRING D
0
201 Q SUMMER
'5 50
U Uj
a
'U
(0 100 100
z
Lii
0
20
0 50
200
Embankment
01
1.0
4.0 r
awo
50
200
Ui
a-
'5 m 150
-J
)- 50
0
0
1959 1960 MEAN 1959 1960 MEAN 1959 1960 MEAN 1959 1960 MEAN
Figure 164. Spring and summer subsurface conditions, Experiment Design 5, Loop 1, lane 2.
than about 2 hr, in order that plots of the data 0C-00'' C'1
9)
TABLE 83
SCHEDULE OF SLAB CURLING MEASUREMENTS MADE IN Looe 1
(Data available in DS 5223)
Round rl Section Design Rep. Dates Hours Ai) Dates Hours (in.) Avg.(in.)
900 1R6 1 June 23-24, 1959' 1430-0330 0.143 June 24, 1959' 0330-1500 0.085
932 iRS 2 1330-0145 0.128 0145-1500 0.106
928 2R6 1 1515-0400 0.120 0400-1515 0.113
916 IRS 1 1600-0445 0.090 0445-1615 0.097
912 4R6 1 1600-0545 0.144 0.124 0545-1630 0.145 0.110
2 934 1N6 1 June 25-26, 1959 1545-0400 0.029 June 24-25, 1959 2330-1545 0.091
926 2N6 1 1600-0400 0.038 June 25, 1959 0430-1600 0.084
918 INS 1 1600-0345 0.059 . 0515-1600 0.099
910 4N6 1 1700-0315 0.048 0545-1700 0.074
914 4N6 2 1600-0215 0.043 0.044 0545-1600 0.077 0.087
3 896 1RO 1 July 14-15, 1959' 1445-0345 0.145 July 15, 1959' 0345-1445 0.118
906 2R0 1 1500-0530 0.135 0530-1515 0.127
922 3R0 1 1530-0515 . 0.070 0500-1515 0.088
908 3R0 2- 1545-0500 0.082 0500-1515 0.088
884 4110 1 1515-0530 0.106 0.114 0530-1545 0.106 0.110
4 936 1NO 1 July 16-17, 1959' 1500-0245 0.088 July 17, 1959' 0245-1515 0.087
924 2N0 1 1615-0545 0.100 0545-1530 0.092
890 2N0 2 1545-0515 0.127 0515-1515 0.122
920 3N0 1 1600-0545 0.118 0545-1545 0.115
886 4N0 1 1545-0530 0.092 0.100 0530-1545 0.091 0.096
3 1 936 1NO 1 Oct. 14-15, 1959 1530-0315. 0.024 Oct. 15, 1959 0315-1430 0.017
924 2N0 1 1415-0300 0.018 0300-1400 0.031
890 2N0 2 1445-0445 0.032 0445-1415 0.038
920 3N0 1 1345-0430 0.054 0430-1445 0.073
886 4N0 1 1500-0500 0.032 0.032 0500-1500 0.058 0.045
2 896 1RO 1 Oct. 16-17, 1959 1200-0200 0.040 Oct. 17, 1959 0200-1345 0.022
906 2R0 1 1300-0300 0.042 0300-1415 0.029
922 3R0 1 1330-0315 0.032 0315-1445 0.035
908 3R0 2 1330-0345 0.031 0345-1430 0.040
884 4110 1 1430-0500 0.055 0.042 0500-1500 0.026 0.028
3 900 1R6 1 Oct. 19-20, 1959 1330-0200 0.035 Oct. 20, 1959 0200-1245 0.028
932 1116 2 1300-0245 0.046 0245-1300 0.044
928 2R6 1 1315-0345 0.043 0345-1315 0.040
916 3116 1 1430-0345 0.056 0345-1345 0.054
912 4116 1 1500-0430 0.079 0.053 0430-1415 0.076 0.050
4 934 1N6 1 Oct. 21-22, 1959 1415-0015 0.046 Oct. 22, 1959 0015-1345 0.060
892 2N6 1 1415-0045 0.027 0045-1400 0.053
918 3N6 1 1430-0130 0.062 0130-1415 0.079
910 4N6 1 1430-0315 0.050 0315-1430 0.052
914 4N6 2 1445-0300 0.041 0.046 0300-1430 0.056 0.061
4 1 936 1NO 1 June 2-3, 1960 1515-0130 0.026 June 3, 1960 0130-1330 0.030
924 2N0 1 1600-0430 0.073 0430-1530 0.077
890 2N0 2 1500-0515 0.083 0515-1430 0.081
920 3N0 1 1530-0500 0.084 0500-1530 0.089
886 4N0 1 1545-0445 0.095 0.070 0445-1345 0.080 0.069
2 934 iNS 1 June 7, 1960' 0215-1430 0.059
892 2N6 1 0430-1500 0.097
918 3N6 1 0545-1515 0.128
910 4N6 1 0600-1530 0.094
914 4N6 2 0600-1545 0.090 0.095
3 896 1RO 1 June 9-10 1960' 1545-0315 0.104 June 9, 1960' 0300-1545 0.092
906 2R0 1 1630-0430 0.118 0530-1630 0.130
922 3R0 1 1530-0500 0.072 0500-1530 0.088
908 3R0 2 . - 1530-0445 0.082 0515-1530 0.088
884 4R0 1 1600-0415 0.127 0.105 0445-1600 0.123 0.111
4 900 1116 1 June 11, 1960 0300-1315 0.078
932 1116 2 June 10, 1960 2245-1245 0.075
928 2R6 1 June 11, 1960 0300-1330 0.041
916 3R6 1 0430-1330 0.048
912 4116 1 0430-1645 0.099 0.067
5 1 900 1R6 1 July 11-12, 1960 1530-0330 0.025 July 12, 1960 0330-1345 0.025
932 1R6 2 1500-0215 0.026 0215-1345 0.023
928 2116 1 1545-0330 0.033 - 0330-1415 0.035
916 3R6 1 1600-0415 0.041 0415-1500 0.043
912 4116 1 1630-0500 0.087 0.047 0500-1500 0.104 0.052
2 936 1NO 1 July 13-14, 1960 1515,0430 0.017 July 14, 1960 0430-1315 0.017
924 2N0 1 1530-0515 0.075 0515-1615 0.070
890 2N0 2 1515-0500 0.103 0500-1530 0.091
920 3N0 1 1500-0530 0.087 0530-1500 0.082
886 4N0 1 1615-0515 0.110 0.072 0515-1615 0.099 0.067
3 934 iNS 1 July 15-16, 1960' 1445-0215 0.069 July 14, 1960 0430-1445 0.046
892 2N6 1 1515-0330 0.088 0430-1515 0.083
918 3N6 1 1615-0545 0.108 0600-1615 0.113
910 4N6 1 1615-0445 0.989 0545-1615 0.098
914 4N6 2 1615-0445 0.084 0.089 0600-1615 0.094 0.085
4 896 1110 1 July 19-20, 1960' 1500-0600 0.146 July 19, 1960' 0515-1500 0.123
906 2110 1 1545-0615 0.156 0545-1545 0.150
922 3110 1 1530-0530 0.088 - 0600-1530 0.088
908 3R0 2- 1615-0615 - 0.091 0600-1615 0.096
884 4110 1 1615-0600 0.126 0.129 0600-1615 0.130 0.123
First figure is slab thickness, in inches; letter designates reinforcing; last ñgure is subbase thickness, in inches
'Data taken this date were included in analysis of curling of 40-ft (reinforced) panels.
'Data taken this date were included in analysis of curling of 15-ft (nonreinforced) panels.
214 THE AASHO ROAD TEST, REPORT 5
joint, at a second point 6 in. from the edge and occurs only momentarily, so that any point on
midway between transverse joints, and at a the surface of a concrete slab can be expected to
third point 1-ft from the centerline and mid- be in continuous motion. This was found to be
way between transverse joints. Table 84 gives the case in the studies of corner movement and
the depths at which the thermocouples were in- the curl of concrete slabs.
stalled for each of the four slab thicknesses 3.3.5.2 Instrumentation of Corner Movement
occurring in Loop 1. and Curling of Concrete Slab s.-As shown in
The thermocouples in three 2.5- and three Figure 170, sixteen small brass plugs were in-
5.0-in, panels were connected to a common stalled in the pavement surface in a 6- x 6-ft
junction box so that the temperatures in these square area in the corner region of all sections
six sections could be observed at frequent in- in Design 1, Loop 1, Lane 2, or a total of 24
tervals over any desired period of time. sections. These plugs provided reference points
Similarly, thermocouples in three 9.5- and three in the pavement surface for use in the study of
12.5-in, panels were connected to a common curling. Table 82 gives details of the designs of
junction box. Thermocouples in the remaining the 24 sections, each of which consisted of but
22 sections were connected to individual junc- one panel either 15 or 40 ft in length.
tion boxes adjacent to the sections. A reference rod and instrument housing,
Automatic equipment (Fig. 166) made it similar to those used in the traffic loops, were
possible to record on punched paper tape, at installed near the corner of each panel (Section
the end of intervals as short as 5 mm, the out- 3.3.1.1). A dial gage with 2-in, travel gradu-
put of as many as 225 thermocouples at a rate ated in thousandths of an inch (Fig. 171) was
of about one thermocouple per second. used for measuring movements of the corner
Figure 167, an example of the kind of infor- relative to the top of the reference rod.
mation furnished by the thermocouple installa- Displacements of the 16 brass plugs on the
tion, illustrates changes in temperature distri- slab surface were measured by means of an
bution with time occurring in plastic concrete. aluminum beam (Fig. 171) equipped with a
Figure 168 shows simultaneous values of air machinist's level, three movable probes, a fixed
temperatures. Apparently the concrete in this probe at one end, and an adjustable support at
12.5-in, panel took its initial set with a para- the other end. The probes were placed on 2-ft
bolic distribution of internal temperature centers to coincide with the reference plugs in
similar to one of the curves shown. Neglecting the pavement. The movable probes actuated
the effects of future changes in moisture distri- dials similar to the one previously described.
bution, one would expect that the hardened slab The procedure followed in taking one set of
would resume its initial shape only when this measurements on the reference plugs in any
particular temperature distribution occurred panel was as follows.
again. Figure 169, which shows typical temper- The beam was first placed in the position
ature distributions in hardened concrete, indi- shown in Figure 171 (designated Position No.
cates that any particular distribution usually 1) with the fixed probe resting on the plug
nearest the panel corner, and the movable
TABLE 84 probes resting on the other three plugs in a line
parallel to the transverse joint. The beam was
DEPTHS AT WHICH THERMOCOUPLES WERE INSTALLED,' leveled in this position and the readings of the
DESIGN 1, Loop 1, LANE 2 dials were recorded. The beam was then ro-
tated through 90 deg about the corner reference
Thermo- Depth Below Surface of Pavement (in.) plug, and placed parallel to the pavement edge
couple with the movable probes resting on the three
Number 2.5-In. 5.0-In. 9.5-In. 12.5-In.
Slab Slab Slab Slab plugs in a line parallel to the edge (Position
No. 2). The beam was again leveled and the
1 .26 .25 .25 .25 dial readings recorded.
2 2.0 1.5 2.0 2.0 The instrument was then moved inward over
3 4.5 2.5 4.0 5.0
the second line of plugs parallel to the edge
(Position No. 3), leveled, and the readings re-
4 6.5 3.5 6.0 8.0 corded. In the same manner measurements
5 8.5 4.5 8.0 11.0 were made on the third (Position No. 4) and
6 10.5 7.0 9.0 12.0 fourth (Position No. 5) lines of plugs parallel
7 9.0 11.5 14.5 to the edge. Finally, the beam was returned to
8 11.0 13.5 16.5 a position perpendicular to the edge (Position
9 13.0 15.5 18.5 No. 6), and readings taken on the line of plugs
10 . 17.5 20.5 farthest from the tranverse joint.
These measurements, combined with a meas-
1
Thermocouples below horizontal line in each column urment made of the distance between a point
were installed in subbase or embankment. on the instrument housing (which was attached
RIGID PAVEMENT RESEARCH 215
Figure 166. Van with equipment for recording internal pavement temperature.
216 THE AASHO ROAD TEST, REPORT 5
AIR TEMPERATURE
I- FIVE FEET ABOVE GROUND
w
I
Z
LLi
90
80
70 L..
500 16 . 17 18 19 20 21 22 2300
HOURS
JULY 9. 1958
2L
9.1
1300 HOURS 1500 HOURS 2200 HOURS 0600 HOURS .7's 1100 HOURS
J
L
5.0 INCH PAVEMENT
91
7c
65
1300 11.11.1 Iuu 0100 0500 0900 300
L AUGUST 12. 1959 AUGUST 13.1959
TIME. HOURS
2 •S :i('
12.5 INCH PAVEMENT
90
I-
85
z
U
a
I
80
0,
U
U
a
0 75
U
0
70
00
TIME. HOURS
Figure 169. Typical changes in temperature distribution in summer for four thicknesses of pavements. 'Curves represent
average of observations on three test sections.
9
218 THE AASHO ROAD TEST, REPORT 5
TRAIN GAG/\
O
0
PLUGS FOR
CURL MEA5URMFNT
,- RCFERENCE ROD
6. 2' 2 2'
+
TLI •T
0 = =
JOINT
Figure 170. Layout of gages, plugs and reference rod for strain and curl measurement
in Loop 1.
A comparison of the three curves (Fig. 172) minimum before t2. By this method the values
—air temperature vs time, the temperature dif- of AT shown at the bottom of Table 82, under
ferential in a 6.5-in, slab vs time, and the "Up Movement" were obtained.
corner displacement of a 12.5-in, slab vs time— For each observed negative value of m, AT
indicates that the motion observed at the panel was selected in a manner analogous to that just
corner was closely related to the other two described, as follows:
variables. Inasmuch as the temperature statis- Within each round the beginning and ending
tic .T had proved to be useful in explaining vari- times of the period during which each panel
ation in load defiections with time (Section corner was moving downward were determined.
3.3.3), this variable was chosen for an investi- From these data the earliest time (t3 ) at which
gation of the correlation of corner movements any panel corner started its downward move-
with changes in temperature conditions. ment, and the latest time (t4 ) at which any
The particular value of AT to be associated panel corner ended its downward movement,
with each observed positive value of m was were found. AT was defined as the net change
selected in the following manner. in T occurring over a period beginning at the
Within each round the beginning and ending time that T was lastat a minimum before time
times of the period during which each panel t and ending at the time when AT was last at
corner under observation was moving upward a maximum before t. Values of AT obtained
were determined. From these data, the earliest by this procedure are given at the bottom of
time (for example, t) at which any of the Table 82 under "Down Movement."
panel corners started its upward movement, These definitions of AT insured that at least
and the latest time (t2 ) at which any panel a portion of the temperature history prior to
ended its upward movement, were found (see the time of observation would be included in
center curve, Fig. 172, for an example). The the temperature statistic and that none of the
statistics, AT, was defined as the net change in temperature history occurring after the period
T occurring over a period beginning at the time of observation would be included.
that T was last at a maximum before time t,, The results of investigation of correlation
and ending at the time when AT was last at a between AT and m are shown in Figure 173,
RI(1D PAVEMENT RESEARCH 219
in which rn. was plotted against AT for eight as in all eight graphs the lines passed close to
combinations of slab thickness and panel the origin, values of .T equal but opposite in
length. The movement data are averages ACIOSS sign are apparently associated with maximum
subbase thickness of the values given in Table corner displacements that are approximately
82. The straight lines were fitted to the data equal but opposite in direction, at least within
by the method of least squares; the root mean the range of values observed.
square residual for each case is shown. Table In the study of corner movements, an effort
85 gives the slopes of these lines. was made to restrict the observation periods to
According to Figure 173 some degree of cor- (lays (luring which large fluctuations in tem-
relation exists between maximum movement at perature were likely to occur, since large (lis-
panel corners and the temperature statistic, placements were judged to be of greater l)1ac-
T, within the observed range of T. except in tical interest than small ones. For this reason,
the case of the 2.5-in, slabs, where the correla- the relationship of ni. to T is not well defined
tion appears to be poor or absent. Inasmuch by the data for small values of T.
Figure 171. Curl beam in position 1 for measurement of vertical movement, with
(upper left) close-up of probe near adjustable SHpj)Ort and (upper right) instrument
for measuring vertical movement at panel corners.
III VERTICAL DISPLACEMENT, INCHES. OF PANEL
CORNER RELATIVE TO ITS POSITION AT TIME OF
FIRST OBSERVATION
T . AIR TEMPERATURE, DEGREES FAHRENHEIT
DEGREES FAHRENHEIT
- 0 0
0 0 UI -0 -0 0) — a) ID
0 jO 0 0 0 0 0
flI I I I I I
0
3 C i—m+.I40 INCHES—
•
rn
2 0 .
ID
o-- N
0 .
-24° F-
--b-
N • S
-10
P1 P1
• rI.
I_I —I •• I.
o z -
•
0 OD
CD
•
• 0 --
C- S ____________
-, C
0 2
rn 0
2 --
-CD
0
0
m
0' 10 S
1 0I
0
•
• S
S
• - S—I
0 0 • I'l
2
ID o
1I 00 _•
mII3 INCHES _______
RIGID PAVEMENT RESEARCH 221
NONREINFORCED SECTIONS
00 .100 100
E0502.5INCHSLAB 5.0 INCH SLAB . 9.5 INCH SLAB 2.5 INCH SLAB
rmSr.028 \.050- rmsr .019 rmSr.OI2 rmSr.009
30 -IS 15 30
______
_____________
-30-IS 15 30
I _
-30
_____
-IS 15 30 -30 IS 30
REINFORCED SECTIONS
100 .100
\ or
..
\ E° I
25INCPI-SLAB 50 INCH SLAB 9.5 INCH SLAB I00[ 12.5 INCH SLAB
050 rmSr .OIO .054 rMS
\ •O5O rmsr .026 rrnSr .OI8 \
t0
_.050 . -.O5O \s\, • 050
C.,
E
..IOO[ \ _.ido[ \ -.100 7.100
\
Figure 173. Maximum movement at panel corner vs corresponding change in T, Experiment Design 1, Loop 1,
data from special studies of corner movements averaged across subbase thickness, replicate sections included.
Round 4 excluded because of absence of temperature data.
The two top curves (Fig. 172), as well as for amplitudes of the air temperature-time
theory*, indicate that successive amplitudes of curve) for a six-month period in 1959 were
the T vs time curve are related to the corre- plotted as shown in Figure 174.
sponding amplitudes of the air temperature vs The convention adopted for determining the
time curve. Pairs of corresponding amplitudes signs of AU and T was as follows: An ampli-
(T for amplitudes of the T-time curve and iU tude AU corresponding to a period during
which the air temperature U was increasing
* See, for example, Thomlinson, J., "Temperature was given a positive sign. During the same (or
Variations and Consequent Stresses Produced by Daily approximately the same) period, T was de-
and Seasonal Temperature Cycles in Concrete Slabs."
Concrete and Constructional Engineering, Concrete creasing, and the amplitude AT was given a
Publications, Ltd., London, England. negative sign.
TABLE 85
SLOPES OF LINES IN FIGURE 173, DATA FROM SPECIAL
STUDIES OF CORNER MOVEMENTS
Figure 174 suggested the following model for lyzed separately (results are in DS 5226). How-
representing AT as a function of AU: ever, because m varied over a wide range
(Table 86) and the larger movements were of
ATI =A 3\/(AU) 2 (102) the greater practical interest, it was decided to
in which J ATJ is the absolute value of AT; summarize the results of the study of curling
A is a constant to be determined from the by an analysis of data averaged over experi-
analysis; and AU is as previously defined. ments yielding relatively high value of m.
The analysis, which minimized the sum of Furthermore, the results of the analyses of
the squared deviations of the observations from load deflections at panel corners (Sections 3.3.5
the values predicted by the model, resulted in and 3.3.7) suggested that there might be a dif-
a value of A = 3.10. The graph of the model ference between the shape assumed by a 15-ft
with this value of A is shown in Figure 174. nonreinforced panel at a given value of T and
The values of AU and AT in dashed-line boxes that taken by a 40-ft reinforced panel at the
were not used in the regression analysis. same value of T, with corresponding differences
to be expected in the motion of points on the
3.5.3.4 Typical Curling of Slabs in Corner slab surfaces. It was decided, therefore, that
Region.—In the course of the special studies of the summary analysis would be made in two
curling (Table 83), there were 14 instances in parts—one of selected experiments made on
which maximum upward displacements were nonreinforced sections and the other of selected
determined at panel corners and 16 in which experiments made on reinforced sections. The
maximum downward displacements were meas- experiments chosen for analysis are given in
ured. In each instance (hereafter referred to Table 86. Further details of these experiments,
as an experiment) Table 83 gives the average including the value of AT associated with each,
value of m for the four first-replicate sections are given in Table 87.
involved, and Table 86 summarizes these aver- Figure 175 plots the average values of m
ages. (Table 86) against the corresponding values of
As a preliminary step in the investigation of AT. The solid lines were fitted to the data by
curling, each of the 30 experiments was ana- the method of least squares. For comparison
50
40
4,
S •
p0 ;..e.:,e:..
• • I. • I
'— — -,
..
.._1 • L_J
20 &mlr.
..
d'....
P. S.
0
I..
, . • .
S
.
,S. • S
S.• •
10 ..
S. S
I • •
I—
I.- _. ..• ..
00
S S
S
—30
• —40
C') aOC
REPLICATE SECTIONS LISTED IN TABLE 79 N N CO N CO N ..' .2
I I 2)0 0 -.-
a) I I
Movement (in.) II
a I ' LC) 00000
0cc o cc 0
0) ,
C' 0
Rnd. Sub Reinforced LI) II) ' CO 0C C
Nonreinforced 0 00000
CED *91
-4 Er
Figure 175. Analysis of curling of concrete slabs, average (solid lines) over slab and subbase thickness (Sec.
3.5.3.4). Dashed lines from special studies of panel corners (Sec. 3.5.3.3).
each of the 80 reference points within an ex- 2. In experiments involving one level of re-
periment were plotted against time and curves inforcing, all values of m associated with one
drawn through the points. Next, the times at level of slab thickness and with one location on
which these curves were at maxima and minima the panel were averaged.
were determined. If the times of maxima and
minima for the 16 points within a panel dif- These processes resulted in two sets of data,
fered, the average times were used to define one for nonreinforced and one for reinforced
maxima and minima for the panel. Smoothed panels, with each set consisting of four sub-sets
readings of the dial gages at the time estab- corresponding to the four levels of slap thick-
lished for panel maxima and minima were taken ness (2.5, 5.0, 9.5 and 12.5 in.). Each sub-set
from the curve for each point. From linear com- in turn, was made up of 16 averaged values of
binations of the smoothed readings for all 16 m corresponding to the 16 reference points on
points on a panel, values for m were calculated a panel.
for each point. The value of AT for each experi- Analysis.—A linear model, whose terms were
ment was determined in accordance with the mutually orthogonal polynomials in the coordi-
procedure described previously. nates x and y of the reference points (Fig. 176)
The conclusion was stated in Section 3.5.3.3
that equal values of AT opposite in sign re- and the slab thickness D2, was used to represent
sulted in movements at panel corners that were each set of data. The model consisted of a total
equal in magnitude but opposite in direction. of 64 terms (16 reference points times four
In the processing of data from the selected ex- thickness levels) with coefficients determined
periments (Table 87), it was assumed that the by the data. Data from replicate sections were
same rule applied to other points on the surface used to determine experimental error, and only
of a panel. In accordance with this assumption coefficients that were found to be significant at
and with the decision stated earlier to combine the 1 percent level were retained in the equa-
the data from certain experiments yielding high tions resulting from the two analyses.
values of m, the data were processed prior to
analysis as follows: For nonrein forced sections:
1. In experiments involving downward mo-
tion at panel corners, the sign of each value of m X 103 = 24.9 -- 0.729x - 0.652y +0.00635x2
m was reversed, so that the signs of rn would be + 0.00743y2 + H 1 (0.2945 - 0.00544x
consistent in all experiments. + 000108y - 0.0000979y2) - H 2 (0.00599
RIGID PAVEMENT RESEARCH 225
(-I
H
+
.0I
+
.03
.04
6' 6'
SLAB THICKNESS 9.5 INCHES
-07 ----
SLAB THICKNESS 12.5
____ INCHES
.02 .01
.03 .02
.04
.06 .05
.04
, .07 \\\C .03
N .09 \ .07
.06
9. \
DOWELLED TRANSVERSE JOINT DOWELLED TRANSVERSE
.150 POINT B
.100
.050
O •' ®
hI
-.050
D. INCHES
k
L,
.10
.08
.07
.03
.02
DOWELLED TRANSVERSE
.01
-101
.09
.08
07 .
08
.04 .03
1
6 I
SLAB THICKNESS 9.5 INCHES SLAB THICKNESS 2.5 INCHES
101
.01
.02
.02
-\\ N .03 T
.03 .04
\.OR
.07
.06
100 .100
950 .050
6 ;
C I I I .- o--
2.5
+
.050 .050
I
U
z
a - .I0c
.100
e
$ :
.050 - cI
I I I I
9.5 2.5 2.5 5.( 9.5 12.5
-.O50
02 . IN CHES
Figure 181. Comparison of results of corner movement study (solid circles) with
predictions from equations for curling of concrete slabs (open circles).
There was little evidence, however, from the followed in the experiments. Results are pre-
studies of corner movements and the curling of sented in graphs of the major and minor
concrete slabs that the displacements at panel principal stresses in a 36-sq ft region bounded
corners relative to the earth were substantially by the pavement edge and a transverse joint
and consistently greater in the case of the re- (Figs. 187 through 190).
inforced than in the nonreinforced panels (see An equation is given from which may be
Tables 82 and 86). estimated the critical stress, in terms of slab
thickness and axle load, caused by a single-axle
3.5.4 Load Stresses in Surface of Concrete vehicle traveling near the edge of the pave-
Slabs ment.
Surface strains resulting from the appli- Of the three pavement design variables, only
cation of a rapidly oscillating load (6 cps) slab thickness had an appreciable effect on
were measured in a series of experiments in the measured strains. For a constant axle load the
non-traffic loop. These measurements served greatest tensile stress occurred when the two
as a means for estimating the stress in the loaded areas, each of which simulated the con-
upper surface of concrete slabs caused solely tact area of a dual tire, were nearest the
by load (as distinguished from stress resulting transverse joint, for all slab thicknesses. The
from environmental changes). The pavement greatest compressive stress occurred at a point
design variables were reinforcing (and/or on the pavement edge with the loaded areas 4
panel length), subbase thickness and slab thick- to 6 ft distant from the joint (Figs. 187 through
ness. 190, Table 90).
Four positions of the load (which simulated For a constant axle weight and slab thick-
that of a single axle load in the traffic loops) ness, it was estimated that the maximum com-
were investigated. These were chosen to repre- pressive stress at the edge due to edge loading
sent four successive positions (relative to a exceeded, in absolute value, the maximum
transverse joint) of a vehicle traveling with its tensile stress due to corner loading by 51 to
outer dual wheel centered on a line parallel to 112 percent. The exact percentage depended
and 1 ft distant from the pavement edge. upon the thickness of the slab (Table 90).
This subsection describes the instrumenta- Between October 9, 1959, and November 2,
tion, field procedures and analytical methods 1960, a series of eight experiments, designed to
furnish information regarding the distribution
TABLE 88 of load stress in the surface of concrete slabs,
was conducted on the sections comprising De-
EXPERIMENT DESIGN FOR SPECIAL STUDIES OF LOAD sign 1, Loop 1, lane 2, with the exception of
STRESSES IN SURFACE OF CONCRETE SLABS the 2.5-in, pavements. The latter were excluded
because the equipment used for applying load
Sub- Number of Sections was too heavy to be placed on pavements of
base 5.0-In. 9.5-In. 12.5-In. that thickness. Table 88 gives the designs of
Thick- Slab' Slab' Slab' the 18 test sections involved in each experi-
ness ment. -
(in.) N R N R N R A rapidly oscillating load was applied to the
pavement through two wooden pads on 6-ft
centers, each approximating the contact area of
a typical dual tire assembly used in Loop 4
'N = nonreinforced; R = reinforced. (Fig. 182). Each load cycle was intended to
RIGID PAVEMENT RESEARCH 229
COCOc COCO
c'
COU)U)CO U)COU)CO
.7 ,-Ic,-4CD
0 II I I II
NCENTER OF J49J COCOCOCO
c',-i -4',-4
_f_.io LOWOED AREA
(____
4C)c,-( I01,.4
C)
U)
CO
C-ICO
- DOWELLED TRANSVERSE C)4O C-I
CO
12 •13 JOIN7 4 0) U)-ICO 'C''-
4-' COCOCO,-4
01-4',,U)
C-I IN '5'C)C') I
IO)ICO cq I01'5'
/ I 0)CII ILr)I-I
' -4L0 0001CO
'I,CON 00000)1-I
'C
*
4C')CO'5' U)CONOO
Figure 182. Numbered points show the several load oZ
positions used in special strain studies.
230 THE AAS110 ROAD TEST, REPORT 5
for the concrete. No gages were required at the opposite components. The dead weight neces-
intersection of joint and edge as the strain sary to prevent the upward components from
there was assumed to be zero. Figure 183 lifting the truck from the pavement was pro-
shows gages being installed; Figure 184 shows vi(led in the form of concrete blocks resting
a finished installation. As in the case of the on a platform located directly above the rotat-
traffic loop installations, cables placed under- ing weights. The load was transmitted through
ground connected gages to a junction box inverted A-frames which could be folded up-
adjacent to the panel. ward against the side of the vehicle when not
The vibrating loader was nlouflte(l on a truck in use. Contact with the pavement was made
(Fig. 185). The essential parts were two ad- through wooden pads (Section 3.5.4) . Strain
justable weights rotating in opposite directions gages mounted on each member of the two A-
in a vertical plane in such a nmnner that all frames (Fig. 186) provided a means for
dynamic force components except those in a calibrating the device on the project's elec-
vertical direction were balanced by equal and tronic scales and for obtaining a continuous
fl I i
Figure 183. Two of the several steps in installation of strain gages in Loop 1.
Figure 184. A finished installation of strain gages in Loop 1. There were 18 installa-
tions of this type. Ramp protected cables from truck-mounted vibrating loader.
RIGID PAVEMENT RESEARCH 231
7
77
record of loading while a pavement gage output possible to determine the amplitude of each,
was being recor(led. Drawings of the vibrating as well as the nature (tension or compression)
loader are available upon request (DS 5200). of the strain. The nominal amplitude of the
In normal operation the load was varied load applied to 3-in, pavements was 12,000 lb.
sinusoidally with time, at a frequency of 6 eps, For the 9.5-in, pavements it was 22,400 lb and
from a minimum value of about 500 lb on each for the 12.3-in. pavements 30,000 lb. These
contact area to a maximum value which de- loads corresponded to the single axle loads in
pended upon the thickness of the pavement Loops 3, 5 and 6.
being tested. The measured strain also varied 3.5.4.2 Field Procedures and Data Process-
sinusoidally with time, very nearly in phase ing.—Field Procedures.—The normal proce-
with the load, and of course at the same fre- dure in taking the data was to begin at the
quency. From examination of simultaneous westernmost factorial section on Loop 1 and
traces of the load wave and strain wave it was proceed eastward. Usually all load positions
Figure 186. Close-up of strut on vibrating loader, showing location of strain gages
that provided continuous record of loading.
232 THE AASHO ROAD TEST, REPORT 5
selected for a particular round were completed that round-load position-slab thickness com-
on a section before measurements were made bination.
on the next section. Thus, for one load position within an experi-
With the load in one of the selected positions, ment, the processing resulted in three sets of
the recording equipment mounted in a van data corresponding to the three levels of slab
(similar to that shown in Fig. 141) was thickness (5.0, 9.5 and 12.5 in.), with each set
switched to each of of the 33 pavement gages - consisting of 33 averaged strain gage readings.
in succession. The output of each pavement As the third step in processing, each such set
gage was recorded as an oscillating line on was converted from strain gage readings to
paper tape, adjacent to the record from the magnitude and direction of major and minor
load gages. The over-all time required to com- principal strains at the 15 gage points on a
plete the measurements associated with one panel, employing standard techniques based on
load position on one panel, including the time elastic theory (see Appendix E). These data
required to set up the vibrating loader, was are available in DS 5205.
about 30 mm, of which about 5 min were spent As the final step before analysis, each prin-
in recording the strains. cipal strain was divided by the corresponding
Data Processing.—The first objective of each load in accordance with the assumption that
experiment was to derive by statistical tech- strain is directly proportional to load. Thus, as
niques a pair of empirical equations for each a result of the four-step processing of data, the
load position, of the following general forms: only remaining independent variables to be
considered in the analysis of strain were the
Major principal strain = a function of pave- coordinates x and y of a gage point and the
ment design, load, thickness, D2, of the slab.
and the coordinates 3.5.4.3 Typical Stress Distributions.—Analy-
of the gage point. sis of Strains.—The three sets of data corre-
sponding to each round-load position combina-
Minor principal strain = a function of pave- tion given in Table 89 (except combinations
ment design, load, involving Round 2, and load positions 12, 13
and the coordinates and 14 in Round 4) were analyzed using statis-
of the gage point. tical procedures. The strain data were repre-
sented by a linear model whose 48 terms (3
slab thicknesses by 16 combinations of x and
(The coordinate system used was that shown in y) were mutually orthogonal polynomials in
Figure 176.) x, y and D2. As a result of the elimination of
The second objective was to compute from reinforcing and subbase thickness as inde-
Eqs. 105 and 106 and the appropriate plane pendent variables, there were six Sections with-
stress equations linking stress and strain the in each round-load position-slab thickness com-
estimated value of major and minor principal bination whose variation in strain furnished a
stresses at closely spaced points in the pave- measure of residual effects. The residual
ment surface within the 36-sq ft area of ob- effects, in turn, were used to determine the
servation. statistical significance of each coefficient. (The
Minor variations in load were observed dur- coefficients from each analysis with significant
ing the 5 min spent in reading the output of terms indicated are available in DS 5211.) Of
the 33 pavement gages. It was assumed (based the 48 original coefficients the 16 terms of the
on the special load tests conducted in the traffic highest order were discarded, and of the re-
loops, Section 3.3.2, and on several pilot experi- maining 32 only those that were found to be
ments on these test sections) that strain was significant at the 1 percent level were used in
proportional to load. Therefore, as the first the calculations.
step in processing the data, the reading of each Distribution of Principal Stresses.—The
gage was adjusted to a value corresponding to analyses of data from load positions 1, 2, 3 and
the nominal load (12,000, 22,400 or 30,000 lb) 4 of Round 7 resulted in four pairs of equations
assigned to the test section. (one for each load position) like the pair given
An examination of the adjusted values in- as Eqs. 105 and 106. Principal strains pre-
dicated that variations in the strain on panels dicted from these equations were converted to
at the same level of. slab thickness but at dif- principal stresses in accordance with the
ferent levels of reinforcing and/or subbase formulas from elastic theory given in Appendix
thickness were small and apparently random in E, using values of Young's modulus and Pois-
nature. Therefore, within each round and for son's ratio for the pavement concrete deter-
the same load position, the readings of gages mined in the Road Test laboratory. The stresses
with the same coordinates, x and y, installed so determined were used in plotting contours
on panels of the same slab thickness (irrespec- of principal stress (Figs. 187, 188, 189, and
tive of subbase thickness and reinforcing) were 190). All stresses are recorded in pounds per
averaged to obtain a set of data representing square inch, with the usual sign convention—
RIGID PAVEMENT RESEARCH 233
6' __ 6'
Th
SLAB
THICKNESS
\ \ 180
90 5.0 INCHES
\ \ K!10
'4
70
60 ___
30 50 60 I
I 6o50I
010 \30\50\70
___________________________________
DOWELLED TRANSVERSE JOINT
I4
6KIPS
T
1:
SVERSE J0INT
to--- 4
-- fl
6'
6'
- -'
( \
-U . 80 (
A6°
70
1I
I 9.5 INCHES
. 20
25 )
,
60
o 80'\ 42 \
o
-15
rf 1
T
__________________________
70 J
DOWELLED TRANSVERSE JOINT
6' _________________________
Z - r
_____
- - DOWELLED TRANSVERSE JOINT
"• I' _ 6' '" ___________________
6'
I 5
0-5 lo—
.
12.5 INCHES
70 -101
50
oio30 40
\\\
40 A
3°I
-I5 20 __
05
I I5KIPS I_
.7—
T DOWELLED TRANSVERSE JOINT - I DOWELLED TRANSVERSE JOINT
Figure 187. Contours of major and minor principal stresses for corner loading (load position 1).
234 __________I THE AASHO ROAD TEST, REPORT 5
6'
- - _
20
I I
SLAB
THICKNESS
[\
60
40 (IIII
40 (
20
N U 5.0 INCHES
___
\
20 20 Z.
0140000
' H -20
r -
20 66185
80 60
60 90\
-20
I
DOWELLED TRANSVERSE JOINT-"
6 L DOWELLED TRANSVERSE JOINT
6
——— — — - — — — — — — —(~~
/1140 — 0
`-
•
9.5 INCHES C
70 -40
io -60
Ih 14-H .I20 8\
1 1.
/10
20
30
401
j360
I II I
f -40
I Ii I
))JIII
DOWELLED TRANSVERSE JOINT-/ DOWELLEDTRANSVERSEJOINT
—I' 6 J
6 6' 61
614
1'
20 I
\. I
-
40
i
T •
C
j30 12.5 INCHES
20 ,, -.-.. __\
60
-I0 -80
Li.20 I4'
A
I
Z_T
30 • -60
20 10 -i
)
10
20 3D
40
T• _ ----____
40
I
DOWELLED TRANSVERSE JOINTJ DOWELLED TRANSVERSE JOINT
6. _-
Figure 188. Contours of major and minor principal stresses for load position 2.
RIGID PAVEMENT RESEARCH 235
__ I -
4030 -20 -
SLAB
0
THICKNESS
k— I4--.4
ZT I00 T
50 INCHES -60
80 ,//
40 6 - :160 -140 20
0 0 20
-80 -60
2'
-20
20
6'
30
2O
1 10
0 *
Z
(D638 U
U- 4
I4
'AW T1 -80 -6 i rI
I/
1
Ip
60
IRS
r \
9.5 INCHES 60 -120
50 /
/ 20 30 40
0 -80
-60
-40
-20
0 I
p
DOWELLED TRANSVERSE JOINTJ I Da.WELLED TRANSVERSE
6'
3021
120
-
is -20 I
U
'N
U
\
I
S SKIPS
T1
j
12.5 INCHES -140 /
-120 -00 /
-80 ID
1015 20 25 30
-60
40
-20
35
0.
___ p •-'
DOWELLED TRANSVERSE JOINT-" I DOWELLED TRANSVERSE JOINT
I 4- I' - 6
Figure 189. Contours of major and minor principal stresses for load position 3.
236 THE AASHO ROAD TEST, REPORT 5
6'
113,3
6KIPS -
C 6KA
SLAB
THICKNESS 64
ll
5.0 INCHES o
N
50
40
\ 30(r0 5
\20 /i
I4
IP7y 30 .S1
160 z
- 7 N . ;~ ZJ
-60
30 too
b 80
• -6D
9.5 INCHES
/
-4D
7I i
-20
o I
I4 4 -
-
I ZK:
IQ
/ -so
35 -140
25
80
I5 = 2.5 INCHES IS
0
20 'U
E
35
40
/ 25
/ /is20
I
/50// /
Figure 190. Contours of major and minor principal stresses for load position 4.
RIGID PAVEMENT RESEARCH 237
TABLE 90
MAXIMUM TENSILE AND COMPRESSIVE STRESSES FOR A 1-Ku' SINGLE AXLE LOAD
(Data from Design 1, Loop 1, Lane 2) -
10.041
SLOPE X 155 1N. PER ONE PERCENT INCREASE
tent was held at a constant value near zero.
IN MOISTURE CONTENT. Figure 193 shows averaged results; the arrows
= III X I15 6 IN. PER IN. PER ONE PERCENT INCREASE indicate the order in which the successive sets
IN MOISTURE CONTENT.
of measurements were made.
The third study was similar to the second,
except that the beams were in a saturated con-
dition (Fig. 194). A comparison of Figures
193 and 194 shows that the dry specimens
OF DRYING DAYS
usually increased their length by a slightly
greater amount than the saturated specimens
for the same increase in temperature.
A comparison of the estimated coefficients of
25/'
10.039 expansion (Figs. 192 and 193) suggests that
adding 0.5 percent moisture to dry specimens
would produce the same unit increase in linear
IN. PER ONE PERCENT
= 355 X I0
INCREASE IN MOISTURE CONTENT
dimensions that occurred when the tempera-
7'
. 355 X Id 6 IN. PER IN. PER ONE PERCENT
ture was increased by about 35 F.
INCREASE IN MOISTURE CONTENT At the time beams were sawed from Loop 1
10.038
sections for testing at the Road Test laboratory,
additional beams were forwarded to the Port-
land Cement Association Laboratory at Skokie,
Ill. The laboratory measured changes in the
length of these beams occurring (a) with a
change of temperature, relative humidity held
constant; and (b) with a change in relative
10.037'
0 0.5 1.0 1.5 humidity, temperature held practically con-
MOISTURE CONTENT.-PERCENT stant. Results for temperature changes at
constant humidity are given in Table 91; length
Figure 192. Effect of moisture content on length of changes associated with an increase in tempera-
concrete beam. Each point is average of measurements
on six specimens. ture differed from those resulting from a de-
crease in temperature.
In the test involving a change in humidity,
The Road Test laboratory found that length the beams were first allowed to reach constant
decreased with moisture loss at a rate as high weight in a moist room adjusted to 100 per-
as 355 x 106 in. per in. per 1 percent decrease cent relative humidity at 69 F. Relative
in moisture content. The Portland Cement As-
sociation reported a unit decrease in length of
229 x 106 associated with a decrease of 50 per- 300
cent in the relative humidity of the air si,ir-
rounding the specimens.
Three laboratory studies were made at the 250
Road Test for the purpose of determining coef-
ficients of expansion of samples of concrete 0 INCREASE IN TEMPERATURE
sawed from nonreinforced sections in Design
5, Loop 1, lane 2. Six 5- by 5- by 14-in, beams 200
The major effort of the staff was directed sections were provided to permit direct per-
toward fulfillment of the first objective—to formance comparisons between sections with
find the relationship between pavement per- and without subbase, and between sections with
formance on the one hand and pavement de- and without shoulder paving. The subbase was
sign, loading and number of load applications 6 in. in thickness; shoulder paving was a 3-in.
on the other. layer of asphaltic concrete 6 ft wide. No in-
Pavement Performance (Section 3.2.2).—An crease in life resulted from use of paved
innovation at the Road Test that strongly shoulders. However, the results may have been
influenced the direction taken by the research affected in some cases by damage to the shoul-
was the development of a definition of pave- der by test traffic. Sections with subbase had
ment performance in terms of the trend in an average life about one-third longer than
pavement serviceability (momentary ability to those without subbase.
serve traffic) with increasing number of load Structural Deterioration (Section 3.2.3).—
applications. The serviceability of each sec- Inspections of the pavements were made
tion at the Road Test was determined at regu- weekly and after each rain. Faulting occa-
lar intervals from the roughness, cracking and sionally occurred at cracks, never at transverse
patching existing in the section (Section 3.2.1 joints (all joints were doweled). Longitudinal
and Appendix F). cracks tended to originate at transverse joints
Most of the major findings relative to the near dowel bars in 2.5-, 3.5- and 5-in, slabs but
first objective were based on data from the not necessarily in thicker pavements. No part
312-section main factorial experiment, and are of the cracking of pavements in the traffic loops
embodied in equations defining the relationship was attributed solely to environmental changes,
between slab thickness, axle load, number of since no cracks appeared in the non-traffic loop
applications of the load, and the corresponding (Loop 1).
reduction in pavement serviceability (Eqs. 61, From cracking data, equations were derived
62, 65 and 66). Given any three of these from which the number of axle applications
factors, the fourth may be found from the equa- associated with any given level of cracking
tions, or from graphs of the equations provided (per unit of surface area) can be computed
for that purpose (Figs. 116 and 117). for a given pavement design and load (Eqs.
Because of random variations in the ob- 69, 70, 71 and 72). Graphs of the equations
served data, there were unavoidable differences for a selected level of cracking are shown in
between predictions from the equations and the Figure 127.
actual performance of individual sections. Pumping of subbase material, including the
Analysis of these differences shows that the coarser fractions, was a major factor in the
scatter corresponds approximately to ± 12 majority of the failures of sections with sub-
percent of the slab thicknesses given by the base. Pumping of embankment material was
performance curves. Figures 120 through confined to those sections constructed without
123 show comparisons of observed data with subbase. The amount of either material
predictions from the equations. If comparisons pumped through joints and cracks was negligi-
are made with the observed performance of an ble when compared with the amount ejected
actual highway in service, additional allow- along the edge (Fig. 133).
ances should be made to account for differ- Strain and Deflection as Functions of Design,
ences between the highway and the Road Test Load, Temperature and Speed (Section 3.3).—
in materials, environment and loading history. Strains were measured at a point on the edge
The effect on performance of the pavement of the pavement 7.5 ft from the nearest joint.
design variable, reinforcing (and/or panel Deflections were measured at the same point
length), was not significant; consequently, this as well as at panel corners. All measurements
variable was excluded from the performance were made with the center of the contact area
equations (see Table 68 for differences in de- of the outer dual wheels of the test vehicles
sign between reinforced and nonreinforced approximately 20 in. from the pavement edge.
sections). It should be noted that the trans- Both static (vehicle stationary) and dy-
verse joints were dowelled in both the rein- namic (vehicle traveling at 35 mph) deflections
forced and nonreinforced slabs. were measured at panel corners. Edge deflec-
The effect on performance of varying the tions were of the static type. Edge strains
thickness of the subbase between 3 and 9 in. were determined with the vehicle moving at
was not significant, and this variable also was 35 mph. Test vehicles were normally those
excluded from the performance equations. regularly assigned to the section, but a series
However, in a separate experiment, the per- of special studies was conducted involving
formance of sections on 6 in. of subbase was other loads, as well as different speeds.
found to be superior to that of sections with The general level of deflection measured at
no subbase. approximately the same time of day over a
Subbase—Paved Shoulder Experiment (Sec- period of several mQnths did not change ap-
tion 3.2.2.2).—Sixty-four nonreinforced test preciably with increasing number of load ap-
RIGID PAVEMENT RESEARCH 241
plications (Sections 3.3 and 3.3.9). Other range of thickness (2.5 to 12.5 in,.) investigated
factors being equal, strains and deflections were (Figure 181).
directly proportional to load ('Section 3.3.2). Load Stresses in the Surface of Concrete
Twenty-four hour studies of the effect of Slabs (Section 3.5 .4).—Surf ace strains result-
fluctuating air temperature showed that the de- ing from the application of a rapidly oscillat-
flection of panel corners, under vehicles travel- ing load (6 cps) were measured in a series of
ing near the pavement edge, at times increased experiments in the non-traffic loop. These
several fold from afternoon to early morning. measurements served as a means for estimating
Edge strains and deflections were affected to the stress in the upper surface of concrete
a lesser extent (Sections 3.3.3., 3.3.4, 3.3.6 and slabs caused solely by load (as distinguished
3.3.7). from stress resulting from environmental
The deflection of a corner of a 40-ft rein- changes).
forced 'panel usually exceeded that of a 15-ft Equal loads were applied on two areas
nonreinforced panel, if load, slab thickness and separated by a distance of 6 ft. Each area simu-
temperature conditions were the same in both lated the contact area of a dual tire of the
cases (Sections 3.3.5 and 3.3.7). Edge deflec- type used in the traffic ioops. Four positions
tions and strains were not affected significantly of the load were investigated. These were
by panel length and/or reinforcing (Sections chosen to represent four successive positions
3.3.4 and 3.3.6). (relative to a transverse joint) of a vehicle
An increase in vehicular speed from 2 to 60 traveling with its outer dual wheel centered
mph resulted in a decrease in strain or deflec- on a line parallel to and 1 ft from the pavement
tion of about 29 percent. Design and load had edge. The pavement design variables were
no consistent effect on the percentage reduc- reinforcing (and/or panel length), subbase
tion (Section 3.3.8). thickness, and slab thickness.
Prediction of Performance from Strain or Of the three pavement design variables, only
Deflection (Section 3.4).—The average life to a slab thickness had an appreciable effect on
serviceability level of 2.5 of sections of the measured strains.
same slab thickness could be predicted with For a constant axle load, the greatest tensile
satisfactory accuracy (Fig. 156) from the aver- stress occurred when the two loaded areas were
age of 24 dynamic edge strains measured under nearest the transverse joint, for all slab thick-
the single-axle load regularly assigned to that nesses. The greatest compressive stress oc-
section. Similar predictions could be made curred at a point on the pavement edge with
from static edge and corner deflections, but the loaded areas 4 to 6 ft from the joint (Figs.
with somewhat less accuracy (Figs. 158 and 187 through 190, Table 90).
160). For a constant axle weight and slab thick-
Subsurface Studies (Section 3.5.2) .—In the ness, it was estimated that the maximum com-
non-traffic loop (Loop 1) CBR and plate bear- pressive stress at the edge due to edge loading
ing values for both the subbase and the em- exceeded, in absolute value, the maximum
bankment tended to be somewhat higher in the tensile stress due to corner loading by 51 to
summer than in spring periods. Plate bearing 112 percent. The exact percentage depended
values determined in spring 1960 in the traffic upon the thickness of the slab (Table 90).'
loops were slightly greater than those measured An equation is given from which may be
at the same time in the non-traffic loop. estimated the critical stress, in terms of slab
Curling of Concrete Slabs (Section 3.5.3).— thickness and axle load, caused by a single-
A study was made in the non-traffic loop (Loop axle vehicle traveling near the edge of the pave-
1) of the maximum displacements of points on ment (Eq. 107).
concrete slabs occurring as the temperature of Moisture and Temperature Coefficients of
the air changed from a maximum to the next Expansion (Section 3.5.5).—Linear coefficients
minimum, or from a minimum to the next of thermal expansion of the portland cement
maximum. concrete, as determined in the Road Test la-
During periods of continuously changing air boratory and by the Portland Cement Associa-
temperature, points on the surface of the con- tion ranged from 4.6 x 10 to 5.6 x 10 per
crete slabs were in continuous vertical motion deg F. The Road Test laboratory reported de-
(Fig. 172). Vertical displacements of panel creases in length with moisture loss as high
corners exceeding 0.1 in. were frequently ob- as 355 x 106 in. per in. per 1 percent decrease
served (Table 82). in moisture content. The Portland Cement
Displacements of panel corners can be esti- Association reported an average unit decrease
mated from corresponding changes in ambient in length of 229 x 10 associated with a de-
temperatures (Table 85 and Fig. 174). The crease of 50 percent in the relative humidity of
rate at which corner displacement changed the air surrounding the test specimens.
with increase in slab thickness varied con- Serviceability Changes, Non-Traffic Loop
siderably—even changing signs—within the (Section 3.5.6) .—Twenty-four test sections in
242 THE AASHO ROAD TEST, REPORT 5
and biweekly determinations of the servicea- way engineering practices, were underdesigned
bility index from slope variance revealed no for the assigned loading. The majority of these
significant changes in serviceability over the 2- failures followed severe pumping involving the
yr period of the test. ejection of subbase material from beneath the
slab. There was evidence that the material was
3.6.2 Needed Research removed from the upper surface of the sub-
In Section 1.4 there is a general discussion of base by the erosive action of water moving
research that would be desirable to improve towards the pavement edge. These factors,
and siihplify the relationships found in the taken together, suggest the desirability of ad-
AASHO Road Test and to extend the findings ditional experiments permitting direct com-
of the Road Test to inëlude other soils, mate- parisons of performance between relatively
rials, and environments. The more important thin slabs constructed on typical granular sub-
areas of research suggested by observations of bases, and slabs of the same design built on
pavement performance at the Road Test are subbase materials resistant to erosion by water.
discussed in this subsection. Subbase thickness in the range from 3 to 9
The constants appearing in the rigid pave- 'in. did not significantly affect pavement per-
ment performance equations would possibly formance at the Road Test. However, the ques-
have taken on significantly different values had tion of the effect of thickness on performance
the Road Test embankment soil, subbase mate- in cases involving more stable subbase mate-
rial or environment been substantially differ- rials—particularly if such materials are resist-
ent. Therefore, if Road Test findings are to be ant to erosion by water—probably can be
applied in areas where such differences exist, answered with precision only through further
additional experiments of the satellite type research.
(Section 1.4) would be useful in assisting engi- The failure of the subbase material to drain
neers to adjust these constants to correspond laterally at a rate sufficient to prevent pump-
with local conditions. ing suggests the need for further research
At the Road Test failure was confined largely directed toward development of more positive
to sections which, according to normal high- criteria for selection of granular subbases.
Appendix A
Pavement Performance Data
This appendix includes the basic performance data for all test sections.
At least five sets of coordinates on the serviceability-applications history
curve are given for each section or subsection in the AASHO Road Test.
For those sections that failed prior to the end of. test traffic the logarithms
of the number of applications at which the smoothed serviceability history
curve crossed serviceability levels of 3.5, 3.0, 2.5, 2.0, and 1.5 are given in
the columns in the left half of the table. For the sections whose service-
ability level did not fall to p = 1.5, the columns on the right give the
serviceability levels at 22-week (11-index day) intervals throughout the
test traffic period.
For the main factorial sections only, coordinates for additional points
are given where appropriate. For example, for a section that failed, the
principal performance data would be the five points (applications at
p = 3.5, 2.5, ... 1.5) on the left of the table. If this section survived, say
30 index days, the serviceabilities at 11 and 22 index days also are given
in the right of the table. Conversely, where complete (5-point) data show
on the right, additional points are given on the left where appropriate. In
certain of these cases, apparent inconsistencies will be noted between infor-
mation on the two sides of the table. These are due to the techniques used
in smoothing and rounding off the data for the tables—in case of doubt
the complete (5-point) set of data should be used. The tables are given in
the following order:
Flexible Pavement Page
Factorial Experiment—Design 1
Weighted Applications 244-248
Unweighted Applications 249-253
Paved Shoulder Experiment—Design 2
Weighted Applications 254-257
Unweighted Applications 258-261
Base Type Experiment—Design 4
Weighted Applications . 262-265
Unweighted Applications S 266-269
Surface Treatment Experiment—Design 6 270-271
Serviceability Trends
Non-Traffic Loop 1, lane 1 . 272
Rigid Pavement
Factorial Experiment—Design 1
Unweighted Applications 273-277
Paved Shoulder Experiment—Design 3
Unweighted Applications 278
Serviceability Trend
Non-Traffic Loop 1, lane 1 279
243
244 THE AASHO ROAD TEST, REPORT 5
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
INDEX DAY
11 22 33 44 55
APPLICATIO}S THROLH INDEX DAY - THOUSANDS
75 290 401 944 1226
SENVICEABILITY TREND LEVEL LOG APPLICATIONE THROtH INDEX DAY STRUCTURE
3.5 3.0 2.5 ?.0 1.5- 4.876 5.462 5.604 5.975 6.089 DESIGN
LOOP LANE SECTION LOG WESGHTED APP. TO SENVICEABILITY LEVEL* SFBVICEABILITY TREND VALUE ON INDEX DAY** D1 D2 D3
FACTORIAL EXPERIMENT-DESIGN 1
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
(Continued)
INDEX DAY I
11 22 33 44 55
APPLICATIONS THROIIH INDEX DAY - THOANDS
75 290 401 944 1226
sVICEABILrTY ThD LEVEL LCX k"LICATIONS THROiH INDEX DAY SYR0CTURE
3.5 3.0 2.5 . ?.O 1.5 4.876 5.I.62 5.604 5.975 6.089 . DES
LOOP LANE SECTION LOG WEXGHTED APP. TO SwICEABILrrY LETL* SEVICEABILITY TREND VALUX ON INDEX DAY" Di D2 D3 SETION
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
(Continued)
INDEX DAY
U 22 33 44 55
- APPLICATIONS T1fflOH INDEX DAY - THOANDS
75 290 /.01 91.4 1226
svICEAILrrY TRD LP1. LOG APPLICATIONS THR0B INDEX DAY scu
3.5 3.0 2.5 20 1.5 4.876 5.1.62 5.604 5.975 6.9 DIGN
LOOP LANE STION LOG WEIGHTED APP. TO SFVICEABILrtY LEVEL' S!BYICEABILTrY TREND VALUE ON INDEX BA?' Di D2 D3 SWrION
FACTORIAL EXPERIMENT-DESIGN 1
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
(Continued)
INDEX DAT
II 22 33 1.1. 55
AMICATIONS THROH INDEX DAY - TBOI35AJ11S
75 290 1.01 94.4 1226
SIRVICEABILrrY TRD LEV1 LOG kFPLIGATIORS TWO(H INDEX DAT SThUCTURE
3.5 3.0 2.5 2.0 1.5 4.876 5.4.2 5.604 5.975 6.089 . DESIGN
LOOP LANE SCfION LOG WEXGHTED APP. TO SEiVICEABII1TY LEVEL' SEBVICEABILITT TRZND YALUY ON INDEX DAT" Di 02 D3 SkCTION
(x.m)
485 3732 4059 4322 4450 4494 3 3 04 485
1
451 3811 4378 4494 4534 4604 3 3 08 451
1
4865 4972 5008 5020 5020 34 - 3 3 12 415
1 415
429 4258 4570 4691 4764 4777 3 3 12 429
1
449 4450 4494 4534 4570 4604 3 6 04 449
1
419 4534 4570 4635 4691 4741 3 6 08 419
1
487 4112 4494 4812 5520 5592 23 21 3 6 12 487
1
413 4664 4834 4875 4890 4905 25 3 9 04 413
1
4494 4777 4855 4947 5020 23 3 9 08 471
1 471
5053 5558 37 31 28 29 29 3 . 9 12 441
1 441
1 411 3987 4093 4403 4494 4534 4 .3 04' 411
481 4812 4875 4890 4919 4933 28 4 3 08 481
1
443 4604 4875 5042 5118 5150 30 4 .3 12- 443
1
473 4494 4801 4875 4875 4890 23 .4 6 04 473
1
455 '4834 4997 5103 5155 5170 34 4 6 08 455
1
453 4777 4933 5053 5118 5140 32 4 6 08 453
1 425
1 425 5063 5390 5626 5735 5794 37 29 27 4 6 12
437 4664 4834 4947 5118 5134 28 4 9 04 437
1
417 4890 5123 5347 5585 5622 35 23 16 . 4 9 08 417
1
477 5166 5901 6086 38 36 33 29 24 4 9 12 477 -
1
439 4691 4777 4789 4812 4823 5 3 04 439
1
421 5042 5118 5129 5134 5145 38 . . 5 3 08 421
1
1 479 5160 5577 5809 5890 5937 41 31 29 5 3 12 479
1 423 4960 5074 5112 5123 5129 37 5 6 04 423
469 4905 5238 5664 5824 5873 36 28 26 ' 5 6 08 469
1
445 5204 5776 5900 6066 39 34 33 22 20 5 6 12 445
1
475 5603 5735 5801 5817 5839 40 37 35 5 9 04 475
1
1 483 5461. 5590 5621 5626 5746 36 35 26 5 9 04 483'
447 5735 5927 40 39 37 29 27 5 9 08 447
1
427 5874 40 39 37 35 35 5 9 12 427
1
486 3255 3352 3431 3556 3607 3 3 04 486
2
2 452 4351 4494 4570 4604 4664 3 3 08 452
416 4604 4834 4985 5118 5190 28 3 3 12 416
2
430 4534 4812 4875 4905 4919 25 3 3 12 430
2
450 4494 4570 4604 4664 4717 3 6 04 450
2
2. 420 4403 4635 4764 4777 4777 3 6 08 420
2 488 4789 5585 5610 5627 5655 34 31 26 3 6 12 488
414 4570 4834 4890 4905 4919 26 3 9 '04 414
2
472 4534 4845 4997 5112 5160 29 3 9 08 4.72
2
2 442 4717 5074 5306 5735 5876 32 23 21 3 9 12 442
2 412 4093 4139 4258 4322 4378 4 3 04 412
2 482 4635 4855 4890 4919 4933 26 4 3 08 482
444 4855 5031 5112 5129 5140 34 4 3 12 444
2
2 474 4845 4890 4905 4919 4933 30 4 6 04 474
2 456 4604 4875 5074 5190 5546 30 16 4 6 08 456
454 4570 5103 5301 5372 5631 33 17 19 4 6 08 454
2
426 4789 5042 5372 5590 5646 33 24 19. 4 6 12 426
2
438 4801 4985 5110 5160 5180 33 4 9 04 438
2
418 4865 5112 5223 5342 5590 35 19 4 9 08 418
2
478 5166 5866 38 36 33 26 30 4 9 12 478
2
440 4947 5020 5020 5031 5042 36 . 5 3 04 440
2
422 4960 5103 5165 5204 5 24 3 37 5 3 08 422
2
480 5084 5440 5810 5900 39 30 25 17 19 5 3 12 480
2
424 4933 5093 5140 5165 5185 37 5 6 04 424
2
2 470 4920 5200 5585 5903 35 26 25 17 19 5 6 08 470
446 5545 5832 5944 39 35 34 24 24 5 6 12 446
2
2 476 5180 5404 5474 5589 5846 38 26 20 5 9 04 476
2 484 4875 5579 5612 5650 5786 35 32 26, 5 9 04 484
448 5604 5868 5991 39 37 35 25 26 5 9 08 448
2
2 428 5665 5876 40 37 37 31 32 5 9 12 428
248 THE AASHO ROAD TEST, REPORT 5
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
(Continued)
IWDEN BA!
11 22 33 44 55
APPLICITIONS T10R INDEX DAY - TH0ANDS
75 290 401 944 1226
SVICEABILITY TREND LEVEL LOG APPLICATIONS TRROR INDEX DA! STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.876 5.462 5.604 5.975 6.089 DESIGN
LOOP LAJE STION LOG UG11ED APP. TO SENVICEABILPrY LEVL1 slVIcKABxtrr! TREND VAL1)! ON INDZ DAY** 01 D D sicrioi
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, UN WEIGHTED APPLICATIONS
INDEX DAY
11 22 33 4.4 55
APPLICATIONS THROUGH INDEX DAY - THOUSANDS
80 233 /.45 807 1111+
SERVICEABILITY TREND L1VEL LOG APPLICATIONS THROUGH INDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 DENIGN
LOOP LANE SECPION LOG UNWEIG}?I'ED APP. TO SERVICEABILITY SERVICEABILITY TREND VALUE ON INDEX DAY** O 02 03 SECTION
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, UN WEIGHTED APPLICATIONS
(Continued)
INDEX DAY
11 22 33 4-4 55
k"LICATIONS THROR INDEX DAY - TH0ANtE
80 233 41.5 807 1.111.
sIczLBILrrT TRD LVEL LOG APPLICATIONS THR01H INDEX DAY 9raucTuR.E
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.61.2 5.907 6.047 DESIGN
LOOP LANE sriow LCG UNVEIGHTED APP. TO 8VICLBILITT L1* SPRVICEABILTrY TRD VALU! ON INDEX DAY' D1 02 D3 SECTION
(in.)
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, UN WEIGHTED APPLICATIONS
(Continued)
INDEX DAY
11 22 33 41. 55
LPFLIGATIONS ThROH INDEX DAY - TRO1EANI
80 233 445 807 1114
SENVICEABILITY TREND LEVE LOG APPLICATIONSTNEOE INDEX DAY CTRUCTIW-E
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 DESIGN
]O0P LANE SNDION LOG EJNWESGHTEI) APP. TO SENVICEABILITY LEVEL SENVICEABILITY TREND VALUE ON INDEX DAY** D1 02 03 SECTION
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, UN WEIGHTED APPLICATIONS
(Continued)
INDEX DAY
11 22 33 4.4 55
APPLICATIONS THRODOR INDEX D! - TH01NDS
80 233 4.45 807 1114
svIcEA.BI1..rrY TREND LEVEL LOG APPLICATIONS THR0R INDEX DAY STRUM ORE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 DENIGN
LOOP LAXE SPrION LOG UWGED APP. TO SENVICEABILI'rr LEVEL' SERVICEABILTrY TREND VALUE ON INDEX DAY" D 02 03 SECrION
(in.)
FACTORIAL EXPERIMENT—DESIGN 1
FLEXIBLE PAVEMENT, UN WEIGHTED APPLICATIONS
(Continued)
INDEX DAY
11 22 33 1.1. 55
APPLICATI0P THR01H INDEX DAY - TEOANI
80 233 1.45 807 111.4
svIcFABILrry TREND LEVEL LOG APPLICATIONS TRR0H INDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.61.2 5.907 6.047 DEIGN
LOOP LANE S!TI0N LOG UNWG}F1'ED APP. TO SVICEABIL1?Y LEVEL' SERVICEABILITY TREND VILU! ON INDEX DAY" Di 02 03 SECTION
(x.xxx)
INDEX DAY
11 22 33 41. 55
APPLICATIONS THROU3H INDEX DAY - THOUSANDS
75 290 401 944 1226
SERVICEABILITY TREND LEVEL LOG APPLICATIONS THROIXH INDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.876 5.462 5.604 5.975 6.089 DE3IGN
LOOP LANE SECTION SUBSECTION* LOG WEIGWPED APP. TO SERVICEABILITY LEVEL" SERVICEABILITY TREND VALUE ON INDEX DAY*** Di 02 D3
(x.xcx)
3 1 177 1 4222 4403 4494 4570 4635
3 1 177 2 4126 4494 4717 4764 4777
3 1 177 3 4093 4121 4255 4450 46'4
3 1 177 4 4 13 9 4351 4494 4534 4654
3 1 179 1 4 11 2 4322 4450 4494 4534
3 1 179 2 4126 451 4494 4635 4691
3 1 179 3 4103 4351 4534 4741 4523
3 1 179 4 4139 4450 4494 4570
3 1 175 1 4812 5042 5155 5213 5218
3 1 175 2 4865 5031 5123 5165 5175
3 1 175 3 4933 5112 5204 5420 5562
3 1 175 4 4933 5175 5455 5557 5610
3 1 183 1 3936 4960 5134 5233 5549
3 1 183 .2 5150 5372 5599 5627 5636
3 1 183 3 5175 5262 5390 5624 5631
3 1 183 4 2954 4777 4960 5112 5213
3 1 173 1 4890 5053 5112 5123 5134
3 1 173 2 4.905 5053 5112 5134 5155
3 1 173 3 4890 5053 5112 5129 5134
3 1 173 4 4865 5042 5112 5118 5134
3 1 181 1 5020 5093 5118 5123 5129
3 1 181 2 5118 5165 5170 .5180 5105
3 1 181 3 5165 5170 5175 5180 5105
3 1 181 4 5020 5074 5118 5134 5145
3 2 178 1 3732 4033 4037 4037 4037
3 2 178 2 3846 4033 4063 4082 4093
3 2 178 3 3908 4042 4071 4079 4082
3 2 178 4 3255 3352 3431 3498 3607
3 2 180 1 3694 3987 4059 4077 4082
3 2. 180 2 3694 4033 4033 4037 4037
3 2 180 3 3811 3987 4055 4079 4093
3 2 180 4 4077 4093 4258 4 45 0 4494
3 2 176 1 3130 4081 4664 4764 4777
3 2 176 2 4075 4789 4875 4890 4890
3 2 176 3 4534 4034 4875 4890 4905
3 2 176 4 4403 4801 4875 4890 4590
3 2 184 1 3732 4664 4875 4090
3 2 184 2 5008 5155 5213 5262 5286
3 2 184 3 4570 4890 4997 5042 5093
3 2 184 4 408.4 4664 4845 4960 5020
3. 2 174 1 4764 4875 4890 4890
3 2 174 2 4604 4875 4985 5008 5020
3 2. 174 3 4789 4933 5053 5112 5123
3 2 174 4 4789 4919 5008 5020 5042
3 2 182 1 4570 4875 4960 5020 5031
3 2 182 2 4905 5020 5042 5063 5074
3 2 182 3 4741 4960 5008 5020 5031
3 2. 182 4 5008 5020 5031 5053 5014
*AVERAGE WIDTH OF PAVED SHOULDER, SUBSECTION 1 - 1 Fl; SUBSECTION 2— 3 Fl; SUBSECTION 3 - 5 Fl; SUBSECTION 4 - 7 P7.
*I8LA IF TREND DID NOT INCLUDE SERVICEABILITY LEVEL.
***BLAHE IF TREND REACHED 1.5 BEFORE INDEX DAY.
APPENDIX A 255
INDEX DAY
II 22 33 4.4 55
k"LICATIONS THROjER IND2 DAY - TROtAN06
75 290 /.01 94.4 1226
r3.5 sRvIcEAaIuTy TREND LF1
3.0 2.5 2.0 1.5 4.876
LOG APPLICHIONS THR0R INDEX DAY
5.462 5.604 5.975 6.069
STRUCTURE
DI0N
Loop LANY SEXi'ION SUTION LOG WGlED APP. TO SBVICEABILITY SERVICEABILITY TREND VALUE ON INDE DAY' Dl 02 03
INDEX DAT
U 22 33 1.4 55
APPLICATIONS THROLH INDEX DAT - TROUEANT
75 290 401 944 1226
SRVxCEABILrn TREND LTEL L(X APPLICATIO8 TRR0tH INDEX DAT STRUCI'VRE
3.5 3.0 2.5 2.0 1.5 4.876 5.462 5.604 5.975 6.089 DIGN
100P LANE STION SUECTION' LOG WGNEED APP. TO SVICEAsILrrT Lw/EL' SVICEABILITT TREND VALUE ON INDEX DAT** Di D2 03
(x.m)
1 435 1 4139 4403 4494 4570 4604
1 435 2 4322 4494 4534 4570 4604
1 435 3 4131 4322 4450 4494 4494
1 435 4 3773 4093 4112 4 258 4378
1 407 1 3811 4042 4139
4086 4258
1 407 2 4117 4322 4635
4494 4764
1 407 3 4079 4139 4403 4494 4494
1 407 4 4 18 3 4378 4534 4635 4741
1 431 1 4351 4664 4801 4865 4890
1 431 2 4403 4741 4801 4855 4075
1 431 3 4427 4717 4801 4834 486
1 43.1 4 4117 4322 4534 4717 4812
1 405 1 4093 4139 4494 4 635 4717
1 405 2 4093 4139 4291 4427 4534
1 405 3 3607 4112 4494 4741 4777
1 405 4 4 11 2 4291 4534 4664 4764
1 433 1 4865 5031 5103 5118 5123
1 433 2 5031 5084 51125118 5123
1 433 3 5093 5118 5123 5129 5134
1 433 4 4801 5008 5118 5123. 5134
1 409 1 4905 5020 5053 5084 5103
1 409 2 5053 5118 5123 5129 5134
1 409 3 4604 4947 5074 5112 5.118
1 409 4 4855 5063 5112 5118 5123
2 436 1 3130 3255 3431 3498 3556
2 436 2 3130 3255 3431 3498
2 436 3 3255 3498 3694 3773
2 436 4 3255 3431 3556
2 408 1 3255 3431 3498 3607 3694
2 408 2 3352 3498 3556 3653 3732
2 408 3 3352 3498 3556 3653 3732
2 408 4 3431 3556 3653 3773 :5773
2 432 1 4351 4764 4855 4875 4890
2 432 2 4050 4494 4691 4789 4823
2 432 3 4131 4570 4717 4789 4823
2 432 4 4 10 3 4322 4635 4764 4834
2 406 1 4183 4691 4834 4875 4890
2 406 2 4075 4107 4378 4741 434
2 406 3 4183 4741 4823 4890 490
2 406 4 4042 4093 4139 4534 4664
2 434 1 4855 5020 5031 5042 5053
2 434 2 4960 5020 5031 5042 5053
2 434 3 4823 5020 5042 5053 5074
2 434 4 4378 4865 5074 5118 5123
2 410 1 4947 5042 5084 5140 5145
2 410 2 5112 5118 5123 5129 5134
2 410 3 5020 5031 5053 5063 5084
2 410 4 5008 6031 5042 5063 5084
APPENDIX A 257
TNDEX DAY
11 22 33 44 55
AWUCATIOND TND0H INDEX DAY - THOIAN08
75 290 401 91.1. 1226
I S](VICEABILITY IREND LEVEL LOG APPLICATIOND THROIEHINDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.876 5.1.62 5.604 5.975 6.089 DNDIGN
LOOP LQP sET1N S1JTION0 LOG WEXG}IrED APP. TO SwiCEABILrrY LEVEL** SBVICEABILI1Y TREND VALUE ON INDEX DAY*** Di 02 D3
INDEX DAY
11 22 33 44 55
APPLICATIONS THROUGH INDEX DAY
80 807
- THOURANUR
1114
233 445
SERVICEABILITY TRENDLEVEL LOG APPLICATIONS THROUGHINDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 DEDIGN
LOOP LANE SECTION SUBSECTION* LOG UNWEI6HrED APP. TO SERVICEABILITY LEVEL** SERVICEABILITY TREND VALUE ON INDEX DAY" D1 D2 03
'AVERAGE WIDFH OF PAVED SHOULDER, SUBSECTION 1 - 1 FT; SUBSECTION 2 - 3 Fr; SUBSECTION 3 - 5 Fr; SUBSECTION 4 - 7 Fr.
'1'BUNE IF TREND DID MIT INCLUDE SERVICEABILITY LEVEL.
"BLANK IF TREND REACHED 1.5 BEFORE INDEX DAY.
APPENDIX A 259
INDEX DAY
II 22 33 44 55
-
APPLICATIO!G TNDOI3E INDEX DAY THOLNI
80 233 41.5 807 1116
SVTCEABIL1TY TREND LEVEL LOO APPLICATIONS THR0H INDEX DAY STRUCrURE
3.5 3.0 2.' 2.0 1.5 4.901 5.368 5.64.8 5.907 6.047 DENION
LOOP LANE SEUFION SUBSEUFION LOG UMiGI11ND I.??. TO SV10EABJLITY LNNELU SRvICEABILrrY TREND VALUE ON INDEX DAY*** riD2 D3
(x.x) (in.)
(x.w)
637 1 4006 4516 4766 4842 4842 04
637 2 4842 4842 4842 4842 4846 04
637 3 4293 4842 4842 4842 4846 04
637 4 4336 4790 4806 4828 4842 04
609 1 3060 3311 3397 3469 3531 04
609 2 4859 4806 4842 4842 4842 04
609 3 4094 4507 4705 4806 4842 04
609 4 3885 4389 4609 4813 4846 04
635 1 4901 4940 4979 5006 5006 04
635 2 4887 4921 4968 4990 4994 04
635 3 4897 4952 4985 4998 5006 04
635 4 4918 4957 6006 5014 5026 04
613. 1 4883 4887 4891 4895 4897 04
611 2 4883 4911 4934 4934 4940 0.4
611 3 4878 4904 4928 4934 4934 04
611 4 4885 4895 4901 4901 4904 04
639 1 4990 4998 5002 5006 5010 04
639 2 5014 5026 5026 5026 5026 04
639 3 4998 5014 5051 5056 5056 04
639 4 4885 4946 4990 4994 04
613 1 4931 4990 4994 4994 4998 04
613 2 4990 4998 5002 5010 5014 04
613 3 4918 4990 4994 4998 5002 04
613 4 4911 4940 4957 4968 4979 0.4
638 1 4774 4846 4874 4861 4881 04
638 2 4870 4870 4874 4874 4878 04
638 3 4.58 4862 4862 4862 4666 04
638 4 4056 4846 4846 4850 4650 04
610 1 3204 3311 3397 3469 3531 04
610 2 3204 3317 3469 3531 3531 04
610 3 3531 3676 3764 3808 3808 04
610 4 3204 3317 3397 3469 3531 04
636 1 4934 4940 4946 4952 4957 04
636 2 4940 4957 4968 4974 4979 04
636 3 4928 4990 5014 5035 5051 04
636 4 4893 494() 5010 5035 5040 04
612 1 4918 4934 4940 4946 4946 04
- 612 2 4934 4940 4952 4957 4963 04
612 3 4940 4957 4968 4985 4994 04
612 4 4934 4952 4963 4974 4985 04
640 1 4979 5030 5040 5046 5051 04
640 2 4998 5006 5014 5022 5035 04
640 3 4911 5006 5022 5035 5035 04
640 4 4928 5030 5035 5035 04
614 1 4952 4994 4998 5002 5010 04
614 2 4885 4998 5010 5022 5026 04
614 3 4994 5002 5014 5022 5026 04
614 4 4914 4994 4998 5002 5006 04
260 THE AASHO ROAD TEST, REPORT 5
INDEX DAY
11 22 33 4.4 55
LPPLICATIONS TBROLU INDEX DAY - TROANr$
80 233 445 807 1114
SNDVICE&BXLITE TRENDLEVEL LOG LPPLICATIONS TRRON INDEX DAY STRUCruRE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 D5IGN
LOOP LAIN SNDfION SUBSPION' LOG UMJEIGND APP. TO STICEABILITY SwICEABILrrY TREND VALUE ON INDEX DAY** D1 P2 D3
(x.)
1 435 1 4766 4828 4846 4854 4858
1 435 2 4806 4846 4850 4854 4858
1 435 3 4747 4806 4842 4846 4846
1 435 4 3764 4551 4705 4790 4821
1 407 1 3808 4129 4413 4766 4790
1 407 2 4716 4806 4846 6862 4881
1 407 3 4363 4766 4828 4846 4846
1 407 4 4774 4821 4850 4862 4878
1 431 1 4813 4860 4887 4899 4904
1 431 2 4828 4878 4887 4897 4901
1. 431 3 4835 4874 4887 4893 4899
1 431 4 4716 4806 4850 4874 4889
1 405 1 4659 4766 4846 4862 4874
1 .405 2 4586 4766 4796 4835 4850
1 405 3 3585 4705 4846 4878 4883
1 405 4 4705 4798 4850 4866 4881
1 433 1 4899 4946 4985 4994 4998
1 433 2 4946 4974 4990 4994 4993
1 433 3 4979 4994 4998 5002 5006
1 433 4 4887 4934 4994 4998 5006
1 409 1 4908 4940 4957 4974 4985
1 409 2 4957 4994 4998 5002 5006
1 409 3 4858 4918 4968 4990 4994
1 409 4 4897 4963 4990 4994 4998
2 436 1 3060 3204 3397 3469 3531
2 436 2 3060 3204 3397 3469
2 436 3 3204 3469 3676 3764
2 436 4 3204 3397 3531
2 408 1 3204 3397 3469 3585 3676
2 408 2 3311 3469 3531 3633 3716
2 408 3 3311 3469 3531 3633 3716
2 408 4 3397 3531 3633 3764 3784
2 432 1 4813 4881 4897 4901 4904
2 432 2 4190 4846 4870 4885 4891
2 432 3 4747 4654 4874 4885 4891
2 432 4 4683 4806 4862 4881 4893
2 406 1 4774 4870 4893 4901 4904
2 406 2 4336 4694 4821 4878 4893
2 406 3 47 , 74 48.78 4891 4904 4904
2 406 4 4129 4488 47.66 4850 4866
2 434 1 4897 4940 4946 4952 4957
2 434 2 4921 4940 4946 4952 4957
2 434 3 4891 4940 4952 4957 4968
2 434 4 4821 4899 4968 4994 4998
2 410 1 4918 4952 4974 5010 5014
2 410 2.: 4990 4994-4998 5002 5006
2 410 3 4940 4946 4957 4.963 4974
2 410 4 4934 4946 4952 4963 4974
APPENDIX A 261
INDEX 0*1
11 22 33 1.4. 55
-
LPPUCC10G THROB INDEX Di! THOtLNI8
80 233 445 807 111.4
SVIC*BIL! TRD LEVEL WG AMJCATIONS TEROUGHINDEX DAT STROCrURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.61.8 5.907 6.047 DTGN
LOOP Lii! STIOP S1ThSiTTON' LOG UWG) APP. TO 5VICEA3IL1TT LFL" SWCE*BXL17Y TRD VALUE ON INDEX DAY' D D2 03
SPECIALBASES EXPERIMENT-DESIGN 4
FLEXIBLE PAVEMENT, WEIGHTED APPLICATIONS
INDEX DAY
11 22 33 1.4 55
APPLICATIOP&S THRONDH INDEX DAY - THOUSANDS
75 290 401 944 1126
*BPSIC DESIGN OF SECTIO?E - 3—INCH SURFACE, VARIABLE THICKNESS BASE, 0—INCH SUBBASE.
*BLANE IF TREND DID NOT INCUJDE SERVICEABILITY LNJEL.
***BLANE IF TREND REACHED 1.5 BEFORE INDEX DAY.
APPENDIX A 263
I NDEX DAY
U 22 33 64 55
APPLICATIONS TOH INDEX DAY - TMOANDS
75 290 1.01 91.4 1126
L)P LANE STI0N SUBSrrI0N L(X WGI8 APP. TO SENVIC*BIUTY jjypj" SVICEABILY TREND VALUE ON INDEX DAY"
rl
264 THE AASHO ROAD TEST, REPORT 5
INDAD( DAY
11 22 33 '.h
A.PFLICATIONS THROWN IND( DAY - THOURAJDS
75 290 601 944 1126
AVURACE BASE
SVICEABILTTY TREND LEL LLG APPLICATIOI€ THROUGHINDEX DAY T}UCKNESS OF
3.5 3.0 2.5 2.0 1.5 4.876 5.462 5.60.4 5.975 6.O9 EACH CL'3SEXTI0
LOOP LANE S3rI0N SUP1ITION LOO WEIG8?ND APP. TO SWICE&BILITY LEVEL" SwICEABILrry TREND VALUE ON INDEX DAY***
INDNE DAY
11 22 33 44 55
&PPLICATIONS T8ROIXR lEONE DAY - T)AJr
75 290 401 94.4 1126
INDE( DAY
1.1 22 33 14 55
APPLICATIONS THROIJ3H INDEX DAY - I
80 233 /45 807
AVER/CE BASE
SERVICEABILITY TREND LEVEL LOG APPLICATIONS TEROIJGH INDEX DAY
THICKNESS OF
3.5 3.0 2.5 2.0 1.5 1.901 5.368 5.618 5.907 6. EACE SUBSECrION
LOOP LANE SEVION SUBSECTION* LOG UMIEIGEffED APP. TO SERVICEABILITY LLVEL** SERVICEABILITY TREND VALUE ON INDEX DAY***
1 169 1
5741 5789
'
Z 1 0Z On 1 250 0
9500
1 169 2 5061 5179 5647
1 169 3 4889 4924 4952 4990 5010 6500
1 169 4 4901 4908 4911 4918 4928 3500
1 105 1 37 34 33 19 19 12 50 0
2 4895 5026 5683 5728 5746 9500
1 105
1 105 3 4842 4874 4891 4908 4918 6500
1 105 4 4842 4846 4846 4850 4854 3500
1 171 1 5744 5751 5758 5760 5760 12500
1 171 2 5010 5237 5718 5734 5741 9500
1 171 3 4918 4985 5026 5071 5090 6500
1 171 4 4842 4883 4885 4887 4891 3500
103 1 5018 5179 5715 5726 5746 12500
1
1 103 2 4887 4921 5010 5030 5035 9500
1 103 3 4854 4874 4881 4883 4883 6500
1 103 4 4846 4850 4858 4862 4806 3500
1 167 1 34 36 36 36 37 9875
1 167 2 39 39 40 39 39 7625
1 167 3 40 41 41 40 41 5375
1 167 4 32 18 23 18 17 3125
1 101 1 38 38 38 37 37 9875
1 101 2 41 41 41 41 42 7625
1 101 3 41 40 41 40 39 5375
1 101 4 4488 5071 5250 5487 5489 3125
2 170 1 5 6 83 5736 5767 795 5836 12500
2 170 2 4911 5040 5109 5169 9500
2 170 3 4828 4897 4940 4994 6500
2 170 4 4854 4878 4881 4881 4883 3500
2 106 1 5080 5751 5795 5830 5865 12500
2 106 2 4269 4921 4990 5040 5095 9500
2 106 3 4705 4842 4883 4893 4934 6500
2 106 4 4842 4846 4850 4854 4858 3500
2 172 1 5585 5709 5760 5765 5769 12500
2 172 2 4924 4990 5035 5080 5090 9500
2 172 3 4828 4862 4883 4893 4904 6500
2 172 4 4846 4850 4854 4858 3500
2 104 1 5104 5723 5732 5734 5741 12 50 0
2 104 2 4914 4963 5002 5030 5035 9500
2 104 3 4457 4866 4883 4893 4901 6500
2 104 4 4457 4806 4842 4842 4842 3500
2 168 1 S 37 39 39 39 39 9875
2 168 2 37 38 40 38 40 7625
2 3 43 41 40 39 38 5375
168
2 168 4 4056 4928 5056 5237 5882 3125
2 102 1 40 40 39 39 39 9875
2 102 2 40 41 41 40 40 7625
2 102 3 34 32 34 27 27 5375
2 102 4 3808 4862 4998 5080 5154 3125
*B/$IC DESIGN OF SECIO}S - 3—INCH SURFACE, VARIABLE THICKNESS BASE, 0—INCH SUBBASE.
**BUc IF TREND DID NOT INCLUDE SERVICEABILITY LEVEL.
IF TREND REACHED 1.5 BORE INDEX DAY.
APPENDIX A 267
IMM DI!
11 22 33 44 55
APPLICCIOM THROtE IND= DAY - THOlAHI)
80 233 445 807 1114
LOOP LANE SL'TION SIJEShITION LOO WMESG}fFED APP. To S0WICEABILTTY LPVEL' SFJWICEABILTTY TREND VALOR ON INDEX DL!*U*
(x.xxx)
4 1 567 1 37 38 36 27 25 14250
4 1 567 2 4885 4940 5030 5 19 4 5715 10750
4 1 567 3 4866 4889 4901 4901 4904 7250
4 1 567 4 4881 4881 4881 4881 4883 3750
4 1 561 1 39 41 38 38 39 1 425 0
4 1 561 2 4899 5022 5683 5709 5741 10750
4 1 561 3 4828 4874 4893 4918 4934 7250
4 1 561 4 4609 4828 4850 3750
4 1 565 1 14250
4 1 565 2 4968 5030 5046 5061 5076 10750
4 -1 565 3 4901 4904 4908 4911 4914 7250
4 1 565 4 3061 3531 3716 3808 3848 3750
4 1 559 1 5071 5355 5583 5647 5651 . 14250
4 1 559 2 4821 4911 4974 5006 5022 10750
4 1 559 3 4821 4883 4 90 4 4934 4940 7250
4 1 559 4 4056 4619 4782 4821 4850 3750
4 1 563 1 37 40 41 40 39 9000
4 1 563 2 56.94 5760 5790 5832 5887 7000
4 1 563 3 4885 5123 5208 5291 5432 5000
4 1 563 4 3716 4766 4813 4862 4914 3000
4 1 557 1 38 35 35 28 27 9000
4 1 557 2 5237 5633 5692 5729 5746 7000
4 1 557 3 4918 4934 4979 4994 4994 5000
4 1 557 4 4747 4850 4878 4897 3000
4 2 568 1 39 39 39 29 29 1 425 0
4 2 568 2 4891 4979 5066 5149 5625 10750
4 2 568 3 4835 4887 4897 4901 4904 7250
4 2 568 4 4774 4858 4881 4881 4843 3750
4 2 562 1 39 42 40 29 30 14250
4 2 562 2 4891 4957 5127 5438 5529 10750
4 2 562 3 4821 4874 4895 4904 7250
4 2 562 4 3311 3633 4006 4598 4813 3750
4 2 566 1 . 14250
4 2 566 2 4897 4968 5035 5051 5066 10750
4 2 566 3 4813 4874 4899 4928 4940 . . 7250
4 2 566 4 3311 3676 3885 4245 4774 3750
4 2 560 1 4921 5026 5095 5127 5164 14250
4 2 560 2 4963 4994 5002 5010 5014 10750
4 2 560 3 4828 4866 4887 4899 4914 7250
4 2 560 4 3311 4609 4774 4806 4828 3750
4 2 564 1 40 43 42 42 42 9000
4 2 564 2 5247 5647 5721 5748 5765 7000
4 2 564 3 4911 4990 5046 5100 5159 -5000
4 2 564 4 3060 3633 4032 4893 3000
4 2 558 1 39 36 36 25 25 9000
4 2 558 2 4963 5080 5213 5510 5540 7000
4 2 558 3 4889 4946 5018 5066 5080 5000
4 2 558 4 4782 4858 4887 4901 4901 . 3000
*BASIC DESIGN OF SESFIONS - 3-INCH SURFACE, VARIABLE THICKHSSS BASE, 4-INCH SUBBASE.
**BLANE IF TREND DID NOT INCLUDE SERVICEABILITY LNJEL.
***BNE IF TREND REACHED 1.5 BAFORE INDEX DAY.
268 THE AASHO ROAD TEST, REPORT 5
INDEX DAY
'Ill 22 33 44 55
APPLICATIONS T0H INDEX DAY - THOUSANDS
80 233 445 801 11.4J
loop LINE SION sIJRSNDrION' Loo UNXXGIED APP. TO SVICEIBILITY LP.V SWTCEABILrT TREND VALOR ON INDEX DAY*
*RASIC DESIGN OF SEC1IONS - 3-INCH SURFACE, VARIABLE THICENESS BASE, h-INCH SUBBASE
**B[,iJ'y IF TREND OIl) NOT INCLUDE SERVICEABILITY LEVEL.
***BLA)C IF TREND REACHED 1.5 BEFORE INDEX DAY.
APPENDIX A 269
INDEN DAY
11 22 33 44 55
AFFLICATIONS THROH INDEX DAY - THO1.ND
80 233 445 807 1114
SVICEABILTrT TREND LtL LOG APMCATIONS THROt}i INDEX DAY AVDRME BASE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.04
EACH ON
WOP LANE SL'TI0N SUDTION* LCG UPMG}fFED APP. TO SDBVICEABILITY LWEL" SRVTICEABILrrY TREND VALUE ON INDEX DAY***
(x.w) (x.x) (x.ux)
6 1 287 1 41 40 41 33 34 17000
6 1 287 2 41 40 36 29 27 13000
6 1 287 3 4887 4901 4904 4904 4909 9000
6 1 287 4 4737 4650 4887 4897 4901 5000
6 1 279 1 41 42 42 41 41 1 700 0
6 1 279 2 41 32 33 18 19 13000
6 1 279 3 4870 4904 4946 5006 5071 9000
6 1 279 4 4881 4883 4857 4889 4891 5000
6 1 285 1 44 43 41 42 39 16125
6 1 285 2 43 43 42 42 41 12375
6 1 285 3 5018 5118 5379 5605 867 8625
6 1 285 4 4897 4990 5010 5026 5026 4875
6 1 283 1 39 37 37 36 34 16125
6 1 283 2 41 42 40 41 40 1 237 5
6 1 283 3 5090 5367 5769 5841 5929
1 8625
6 283 4 4974 5026 5066 5118 5159 4875
6 1 289 1 39 39 37 31 34 11750
6 1 289 2 5217 5342 5422 5521 5622 9250
6 1 289 3 4862 4891 4911 4934 4952 6750
6 1 289 4 4842 4846 4846 4850 4850 4250
6 1 281 1 39 39 48 32 33 11750
6 1 281 2 5432 5584 5701 5797 6038 9250
6 1 281 3 4952 5006 5046 5080 5123 6750
6 1 281 4 48.06 4862 4887 4899 4901 4250
6 2 288 1 42 44 43 41 40
2 17000
6 288 2 42 43 42 39 37
6 2 13000
288 3 4931 5002 5056 5189 5661
6 2 4 9000
288 4850 4893 4921 4968
2 5000
6 280 1 40 39 39 38 38 17000
6 2 280 2 42 39 39 33 35 13000
6 2 280 3 4889 4918 4974 5026 5774 9000
6 2 280 4 4850 4:887 4897 4901 4904 5000
6 2 286 1 40 40 48 38 35 16125
6 2 286 2 40 39 38 37 37 12375
6 2 286 3 : 39 25 28 21 19 8625
6 2 286 4 4895 4952 5010 5030
2 4875
6 284 1 41 38 38 39
2 35 16125
6 284 2 41 43 43 41 41
6 2 12375
284 3 42 36 37 31 : 30
6 2 4 8625
284 4928 5090 5203 5233 5256 4875
6 2 290 1 42 44 42 37 32 11750
6 2 290 2 41 32 33 20 21 9250
6 2 290 3 4904 4957 5022 5080 5090 6750
6 2 290 4 4881 4581 4883 4883 4583 . 4250
6 2 282 1 43 43 40 36 32 1 175 0
6 2 282 2 40 37 29 23 24 9250
6 2 282 3 4990 5051 5104 5159 5184
6 2 6750
282 4 4934 4940 4940 4946 4946 4250
BASIC 0NEIGN OF SECTI0 - 4-INCH SURFACE, VARIABLE THICKNE3S BASE, 4-INCH SUBBASE FOR BITUuIN0US AND CEMENT BASE AND
8-INCH STONE BASE.
270 THE AASHO ROAD TEST, REPORT 5
INDEX DAY
U 22 33 44 55
-
APPLICATIOND fHROUGH INDEX DAY THOUSANDS
1226
75 290 401 944
SERVICEABILITY TREND LEVEL LUG APPLICATIOND THROUGH INDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.876 5.462 5.604 5.975 6.089 DEDIGN
LOOP LANE SEOrION LL wG}rrED APP. TO SERVICEABILITY LEVEL* SERVICEABILITY TREND VALUE ON INDEX DAY** D 02 03 SEOTION
INDEX DAY
11 22 33 44 55
APPLICATIOND THROIZH INDEX D! - THOUSANDS
80 233 445 807 1114
SVICEABILITY TREND LEL LOG APPLICATI0E THRO1HINDEX DAY STRUOURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.648 5.907 6.047 DHEIGN
LOOP LANN SCTTON LCX UNWEIGHTED APP. TO SERVICEABILITY L.XTEL* SERVICEABTLTTY TREND VALUE ON INDEX DAY 'D 02 D3 SECrION.
FACTORIAL EXPERIMENT-DESIGN 1
RIGID PAVEMENT, UNWEIGHTED APPLICATIONS
INDEX DAY
U 22 33 1.4 55
80 233 807
-
LPPUCATIO6 TIiROIEH INDEX DAY THOAN1
4.65 111.4
sENVIcKABILrTT TREND LEVEL LOG LPPLICATIO!6 TER0R, INDEX DAY PTRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.64.8 5.907 6.0/.7 DEDIGN '**
LOOP LANE SEDFION LOG UNWEIGHTED APP. TO SENVICEABILITY LEVEL' SENVICEABILTTY TREND VALUE ON INDEX DAY" SEDFION
Di 02 03
FACTORIAL EXPERIMENT—DESIGN 1
RIGID PAVEMENT, UN WEIGHTED APPLICATION
(Continued)
INDEX DAY
11 22 33 44 55
LppLIcjTIoNs TOLR INDEX DAY. - TROLN1
80 233 1.45 807 1114
LOG APPLICATIONS THROR INDEX DAY STRUCrW.E
SRVICEABILITY TREND LEVEL DEIGN
3.0 2.5 2.0 1.5 4.901 5.368 5.642 5.907 6.047
3.5
LCG UWGffEI) APP. TO SVICEABILITY LgJa' SENVICEABIUTY TREND VALUE ON INDEX DAY" D1 02 D3 SION
LOOP LA10 SFrION
41 39 0 35
1 195 5441 5467 5489 5495 5497 0 35
1 239 5432 5441 5447 5455 5461 43 40
5284 5329 5388 5450 5461 47 27 0 35
1 - 213 37 0 50
1 225 43 41 41 41
245 6047 44 43 42 41 35 0 50
1 46 42 43 38 31 0 50
1 221 5989 42 37 0 50
1 219 45 44 42
217 48 47 45 46 44 0 65
1 42 41 40 40 39 0 65
1 193. 45 45 44 41 0 65
1 249 . 45
46 46 44 44 42 0 65
1 207 . 45 46 44 0 80
1 201 46 45
235 46 46 45 44 43 0 80
1 44 42 42 42 40 0 80
1 185 1 35.
1 209 5382 5412 5429 5438 5444 47 37
205 5174 5237 5405 5432 5435 45 26 1 35
1 43 1 35
1 .231 5435 5463 5482 5495 5510 45
251 5955 6042 40 39 39 38 28 1 50
1 45 45 44 40 1 50
1 203 . 45
191 5849 5851 5855 5856 5860 44 44 43 1 50
1 45 43 39 33 1 50
1 233 5996 46
4
199 45 45 46 42 1 65
1 45 43 45 43 1 65
1 247 46
237 47 46 42 47 45 1 65
1 . 46 45 44 44 44 1 65
1 241 44 45 43 1 80
1 211 47 46
215 46 45 45 46 42 1 80
1 44 44 43 41 1 80
1 197 46
2 196 5484 5489 5495 5497 5502 41 41 0 35
240 5280 5303 5310 5314 .5 321 43 0 35
2 0 35
2 214 5291 5332 5370 5422 5473 46 25
226 5674 5755 5825 5836 5848 46 42 36 0 50
2 46 45 43 43 28 0 50
2 246 5987 6034
2 222 5736 5882 5910 5933 5954 46
45
41 38 26 0 50
2 220 5813. 5845 5853 5863 5887 43 41 0 50
218 47 45 46 44 42 0 65
2 42 41 42 40 0 65
2 194 43
2 250 . 46 44 44 43 41 0 65
2 206 45 42 44 42 40 0 65
2 202 46 45 45 45 43 0 80
2 236 46 46 45 43 43 0 80
166 43 43 43 44. 42 0 80
2 41 1 35
2 210 5432 5435 5438 5441 5444 46
206 5329 5367 5415 5432 5438 47 30 1 35
2 41 1 35
2 232 5397 5435 5441 5452 5468 45
2 252 5951 5995 6018 6029 6041 39 40 40 38 1 50
2 204 5973 5994 6008 6013 6019 45 44 45 43 1 50
192 5706 5734 5751 5769 5799 43 43 40 1 50
2 45 44 39 1 50
2 234 5651 5874 5889 5893 5899
200 46 44 45 45 41 1 65
2 44 42 45 43 1 65
2 246 46
238 47 44 41 42 41 1 65
2 45 44 46 44 1 65
2 242 46
2 212 45 44 43 42 41 1 80
2 216 44 44 43 44 40 1 80
2 198 . 45 44 44 44 43 1 80
APPENDIX A 275
FACTORIAL EXPERIMENT-DESIGN 1
RIGID PAVEMENT, UNWEIGHTED APPLICATION
(Continued)
I NDEX DAY
11 22 33 44 55
LPPLICHIONS TI0H INDEX DAY - TEOLEANI
80 233 4.45 807 1114
SIRViCEABILTY TREND LEVEL LOG APPLICATIONS THROR. lEONE DAY STRUCTE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.649 5.907 6.01.7 DEIGN 1**
LOOP LANE SI0N LOG UWGHTED APP. TO SENVICEABILrrY LEVPZ' SENVICEABIL1'T! TREND VALUE ON INliNE DAY" D2 03 sEOrIow
(x.xxx) (x.x)
4 1 643 5540 5681 5609 5849 5855 46 45 31 0 50 3 643
4 1 647 5452 5475 5495 5512 5547 45 44 0 50 6 .647
4 1 677 5412 5435 5444 5455 5463 45 43 0 50 9 677
4 1 649 47 46 46 44 38 0 65 3 649
4 1 697 45 44 45 45 44 0 65 6 697
4 1 655 46 45 45 46 43 0 65 6 655
4 1703 5984 6021 46 44 43 44 30 0 65 9 703
4 1 671 46 45 46 45 44 0 80 3 671
4 1 6.87 45 45 45 46 45 0 80 3 687
4 1 683 47 47 47 46 44 0 80 6 683
4 1 651 46 45 44 45 43 0 80 9 651
4 1 675 44 42 43 43 42 0 95 3 675
4 1 701 47 45 46 47 45 0 95 6 701
4 1 689 45 43 43 42 41 0 95 9 689
4 1 681 5529 5549 5577 5595 5618 46 44 1 50 3 681
4 1 661 5487 5502 5510 5512 5512 45 43 1. 50 6 661
4 1 673 5746 5758 5762 5767 5772 45 44 39 1 50 9 673
4 1 641 45 46 46 42 38 1 65 3 641
4 1 705 46 45 45 43 36 1 65 3 705
4 1 685 6042 46 45 45 44 34 1 65 6 685
4 1 653 5983 6000 6015 6037 46 46 44 44 18 1 65 9 653
4 1 691 46 45 44 42 39 1 80 3 691
4 1 669 47 46 46 47 44 1 80 6 669
4 1 707 45 43 42 43 39 1 80 6 707
4 1 695 47 46 46 46 43 1 80 9 695
4 1 645 45 45 41 . 41 40 1 95 3 645
4 1 665 46 46 45 46 45 1 95 6 665
4 1 667 47 47 46 48 48 1 95 9 667
4 2 644 5480 5502 5516 5525 5534 46 44 0 5.0 3 644
4 2 648 5376 5412 5463 5512 5516 46 37 0 50 6 646
4 2 678 5438 5444 5450 5455 5461 45 42 0 50 9 678
4 2 650 5603 5633 5671 5694 5836 46 45 28 0 65 3 650
4 2 698 6032 45 45 45 46 34 0 65 6 698
4 2 656 5931 5950 5965 5976 5999 46 45 45 41 0 65 6 656
4 2 704 5816 5843 5851 5855 5858 45 44 44 0 65 9 704
4 2 672 45 44 43 42 41 0 80 '3 672
4 2 688 45 45 45 44 42 0 80 3 688
4 2 684 46 46 45 44 42 0 80 6 684
4 2 652 44 44 44 4.3 41 0 80 9 652
4 2 676 42 41 42 41 40 0 95 3 676
4 2 702 46 45 45 45 42 0 95 6 702
4 2 690 45 44 44 44 42 0 95 9 69.0
4 2 682 5401 5438 5447 5458 5482 45 40 1 50 3 682
4 2 662 5149 5189 5222 5233 5243 43 1 50 6 662
4 2 674 5559 5575 5585 5598 5610 44 44 1 50 9 674
4 2 642 5901 5926 5953 44 45 41 34 26 1 65 3 642
4 2 706 5867 5885 5891 5895 5899 46 45 46 1 65 3 706
4 2 686 5887 5891 5895 5897 5901 46 46 46 1 65 6 686
4 2 654 5956 5971 5985 6001 6015 44 46 44 44 1 65 9 654
4 2 692 46 46 46 44 40 1 80 3 692
4 2 670 45 46 45 46 44 1 80 6 670
4 2 708 45 44 45 44 38 1 80 6 708
4 2 696 46 45 45 45 42 1 80 9 696
4 2 646 45 45 44 43 40 1 95 3 646
4 2 666 45 45 45 46 43 1 95 6 666
4 2 668 46 46 46 47 46 1 95 9 668
276 TEE AASHO ROAD TEST, REPORT 5
FACTORIAL EXPERIMENT—DESIGN 1
RIGID PAVEMENT, UNWEIGHTED APPLICATION
(Continued)
NDEX cAY
11 22 33 44 55
A?FLICATI06 TIOIOUUIi INDNE DAT - TN0ANDS
80 233 1.1.5 807 1114
svIcEKBIi.rrY TREND LEV1 L40 APPLICATIO?6 TMRON INDFI DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.61.8 5.907 6.047 DESIGN '*
LOOP LANE STION 1.40 UWGF?TED LI'?. TO S0WICEABILrTY LEVEL' SSRVICEABILrrY TREND VALUE ON INDNE DAT" Di 02 D3
(x.,ccx) (in.)
FACTORIAL EXPERIMENT-DESIGN 1
RIGID PAVEMENT, UN WEIGHTED APPLICATION
(Continued)
T ND
1EX -DAY
11 22 33 1.4 55
APPLICATI0? THR0UR INDED DAY - THOUSANDS
80 233 445 807 1111.
sENvIcEABILrry TREND LEVEL LOG AJ'PLICATIO?6 TRR0iH INDEX DAY STRUCTURE
3.5 3.0 2.5 2.0 1.5 4.901 5.368 5.61.8 5.907 6.01.7 OIC
WOP LANE SECrION LCE UNWEIGWTED Al?. TO SVICEABILI?Y LEVEL SERVICEABILITY TREND VALUE ON INt)EX DAY' D 0 03 SECTION
INDEX DAY
11 22 33 41. 55
APPLICATIOHR THROIN INDEX DAY .- TH0USAN1
80 233 445 807 1114
SERVICEABILTI'Y TREND LEVEL LOG APPLICATIONE THROUGH INDEX DAY STRUCTURE
3.0 2.5 2.0 1.5 4.901 5.368 5.642 5.907 6.047 DESIGN***
3.5
LOOP LANE SECTION LOG UNWEIGHTED APP. TO SENVICEABIIJTY LEVEL* SERVICEABIliTY TREND VALUE ON INDEX DAY D1 D2 D3
(x.xxx)
- 0 = NONREINFORCED; 1 = REINFORCED
D2 - SLAB THICKNENS
- SUBBASE THICKNESS
MISCELLANEOUS DATA
DFLFZTIONS OF LOOP 1 FLEXIBLE SEDTIOHS
USED IN SEASONAL WEIGHTING FUNCTION FLEXIBLE PAVE]ENT
TIME - APPLICATIONS CLIMATIC DATA (THOUSANDTHS OF INCHES) - WEIGHTED APPLICATIONS
z
(DESIGN; SURFACE - BASE - SUBBASE) g
)-1 -.
SO sO
C.) C.) b.', Z • C) Z • P. Z P. C.) P-P C) Q P-P Z C-. I
.4
I I
sO
.-1
I I
1
I I
r-1
I Z
0= C-
C') C) C) C)
P-I
-)
-
C.) C?
0 P-I 0 Z X1 P. CC. E C) C\ I1 Cfl 115 115 115 11) Cl.
8 = Replicate
• , I . -.
SPRING CREEP SPEED DEFLECTION DATA;
MEAN OF TESTS—MARCH 9 AND MARCH 31, 1959
10 in. in. 1073 in. 1073 in. 10 31n. 1073 in. 1073in. 10 in. 10 in. 10 in. 10 in.
in. io- in.
130* 3-0-0 76 70* 3-0-4 342* 182* 132* 4-3-4 85* 110* 62* 4-3-8 70* 347* 49
1-0-0 65* 80 4-3-12 28. 60 28
1-0-4 110* 3-0-4 56 44 3-0-8 80* 120* 4-3-8 50 48
1-3-0 90* 3-0-8 52 66 3-0-12 52 77 46 4-3-12 40 76 45 4-3-16 42 92 30
1-3-4 77 3-3-0 72* 78* 3-0-12R 48 56 40 4-6-4 71 90 58 4-3-16R 46 97 36
1-6-0 62 3-3-4 46 38 3-3-4 80* 120* 78* 4-6-8 44 78 46 4-6-8 50 88* 31
1-6-4 44 3-3-4R 63 72 3-3-8 70 86 57 4-6.-8R 40 62 43 4-6-12 62 3.1.1 48
2-0-0 70* 3-3-8 51 68 3-3-12 36 45 29 4-6-12 32 50 30 4-6-16 37 59 26
2-0-4 56 3-6-0 74 55 3-6-4 75* 112* 90* 4-9-4 54 97 50 4-9-8 64 109* 38
2-3-0 32 3-6-4 54 56 3-6-8 57 . 66 46 4-9-8 40 71 42 4-9-12 40 82 33
2-3-OR 40 3-6-8 44 37 3-6-12 34 48 33 4-9-12 28 44 26 4-9-16 27 56 28
2-3-4 32 4-0-0 69 65* 4-0-4 74 83 70* 5-3-4 51 95 44 5-3-8 52 96 33
2-3-4E 28 4-0-4 67 66 4-0-8 32 44 40 5-3-8 40 60 34 5-3-12 37 76 '27
2-6-0 24 4-0-8 70 74 4-0-12 34 38 30 5-3-12 29 48 25 5-3-16 30 65 26
2-6-OR 28 4-3-0 59 46 4-3-4 65 79 61 5-6-4 46 72 40 5-6-8 44 87 32
2-6-4 26 4-3-4 54 56 4-3-8 49 62 46 5-6-8 29 57 28 5-6-12 32 64 26
2-6-/.E 26 4-3-8 39
.-
R = REPLICATE
* = ESTIMATED VALUE
19
Appendix D
TEST PROCEDURES AND EQUIPMENT
FLEXIBLE PAVEMENT DEFLECTION tires were opposite the contact point. When
the vehicle had moved well past the contact
All deflections in the flexible pavement point of the probe arm and the dial hand had
experiment were measured with either the come to rest, the final reading was recorded.
Benkelman beam or the linear variable differen- The difference between the initial and load
tial transformer (LVDT). Schematic drawings reading multiplied by two was the deflection.
of the Benkelman beam and the LVDT are In the other procedure, termed "creep speed
shown in Figures 1-D and 2-D. rebound," the probe was inserted between the
tires to a distance of about 11/2 ft, a load
Benkelman Beam Deflections reading taken as the wheels passed the probe
and a final reading taken with the load out of
Two procedures were employed in tests with range of influence. Here the difference in the
the beam. In one, termed "creep speed normal," two readings multipled by two was the de-
the probe arm of the beam was inserted be- flection.
tween the dual tires of a loaded test vehicle to
a distance of about 41/2 ft, lining up the arm
by eye in such a position that rubbing of the LVDT Pavement Deflection,s
probe arm and the tires would not occur. While The LVDT equipment consisted of a small
the truck was standing, the buzzer was turned transformer tubing, a steel reference rod and
on and the initial reading of the dial taken. a perforated steel plate. The perforated plates
The vehicle was then moved slowly forward were placed at the pavement layer interfaces
(creep speed) until it was at least 10 ft past and in the embankment during construction.
the tip of the beam. While the vehicle was After the pavement was completed, a .hole was
being moved forward, a maximum dial reading formed down to each plate and a section of
was recorded which occurred when the dual tubing was inserted. The lower part of the
* For description of deflection tests made on rigid tubing was of a special flexible type. The
pavements see Section 3.3.1. upper part, extending to the top of the surface
Beam guide
Cap
BALL BEARING
PIVOT Spring
Latch
ALUMINUM PROBE BEAM Steel tube
late brazed to
Washer
legs and bolted
to beam Spring
6 4 611
A—A
ASPHALTIC CONCRETE
NLVDUNI
ALUMINUM CAP
RUBBER GROMMET
TEEL ROD D.
STAINLESS STEEL
6 BELOW SURFACE
LVDT unit was removed and the hole was 18 16, 32, 48
sealed with a special tyl)e of stopper. 24 10,20,28
An instrument van contained all the neces- Base 12 16,32,48
sary electronic equipment for operating the 18 16,32,48
24 10, 20,28
L\T DTs and recording the deflection. Figures 18 10, 20, 28
3-D and 1-D show the exterior and interior of Subbase
24 10,20,28
the van. 30 5, 10, 15
Measurements were made with the loaded Embankment 18 5, 15,25
test vehicle driven at a selected speed so that 24 5,15,25
the center of the dual wheels passed, as closely 30 5, 10, 15
as possible, over the transformer. Either a
continuous record of the vertical movement of (1)) RIGW PAVEMENT
the transformer relative to the plate was re-
corded on graph paper or the maximum verti- Surface 12 35, 70, 105
cal movement was recorded on punched tape. Subbase 24 5, 10, 25
30 5, 10, 15
Plate Load Tests Embankment 18 5, 15, 25
Basic equipment consisted of: (1) reaction 24 5, 10,15
:10 5, 10,15
trailer; (2) hydraulic ram and jack; (3) vari-
ous sizes of steel sl)aCel'S for use in trenches
of different depths; (4) a 12-in, diameter Plate and pressures used for all routine testing.
Used on 3-0-0, 3-6-0 and 3-0-8 design sections.
cylindrical steel loading frame cut out on two Instructions were that total gross deflection should
sides to allow use of center deflection dial; not exceed 0.20 in.
(5) spherical bearing block; (6) 1-in, thick Used on 3-6-8, 3-6-16 and 3-0-16 design sections.
AI'1'INI)1X 1) 285
steel plates, 12, 18, 24 and 30 in. in diameter; checked periodically, was used to convert gage
(7) 16-ft long aluminum reference beam. A pressures to load in pounds. The load was ap-
schematic diagram of the apparatus is shown plied to the plates through the 12-in, diameter
in Figure 5-D. steel loading frame and the spherical bearing
The reaction trailer was of the flat-bed type, block. I)eflection was measured with a dial
having no springs and four sets of dual wheels gage (Fig. 4-D).
on the rear. For the tests on the AASHO Road The weight of the loading frame and plates
Test a cantilever beam protruding from the was allowed to act as a seating load for which
rear of the trailer was used as a reaction. The no correction was made.
distance load to rear wheels was 8 ft. A maxi- Tests were made in trenches about 3 to 4
mum reaction of about 12,000 lb could be ob- ft wide. The procedure provided for the ap-
tamed with a 17,000-lb loaded rear axle. plication and release of three different psi
A standard hydraulic ram was used to apply loads and for measurement of the downward
the load. A calibration curve, which was and upward movement of the plates. The loads
Ago
- --
.LLmI4 J
I 1jJ':
~
trf_
..
SR-4 GAGES
FACE PLATE 7
6"
7— MERCURY CHAMBER
Ili
FLEXIBLE RIM
GAGE CHAMBER
DIAPHRAGM
COVER PLATE
_V FLEXIBLE TUBING
STAINLESS STEEL
RIBBON BRAND
TERMINAL STRIP
(In some early tests, an automatic compaction pavements. At each location, when the pave-
hammer with a pie-shaped foot was used.) ment was completed, a hole was formed in the
After soaking four days the specimens were structure to the level of the perforated plate.
penetrated with a 1.95-in, diameter piston and A rod with a sheet-metal screw soldered to one
the load-deformation curve was plotted. This end was turned into one of the holes of the
curve was corrected for initial curvature and a perforated plate and protective flexible tubing
"corrected" load determined at 0.1- and 0.2-in. placed in the hole so that the rod moved with-
penetration. CBR values reported were the out restriction. The top of the rod extended to
corrected unit loads, in psi, expressed as a within approximately 2.5 in. of the pavement
percentage to the values 1,000 at 0.1-in, pene- surface. \Vhen not in use the holes were capped
tration and 1,500 at 0.2-in, penetration. That is, with a special stopper.
Corr. unit load x 100 Figure 6-D shows a typical installation.
CBR = Throughout the course of the test frequent
1,000 measurements were made from the pavement
surface to the top of these rods. Changes in
in which CBR is measured in 0.1-in., and cor- these measurements indicated changes in the
rected unit load in psi at 0.1-in, penetration. thickness of the structure to the level of the
Surcharge weights were 10 lb for specimens plate. An electronic device used to make these
of subbase and base and 25 lb for specimens measurements recorded the data electronically.
of soil, for both soaking and testing.
Field In-Place CBR Test.—The test pro- PRESSURE CELLS, FLEXIBLE
cedure used followed closely that described in PAVEMENT
"Suggested Method of Test for Moisture- The pressure cell used at the Road Test on
Density Relationships and California Bearing the embankment surface utilizes SR-4 strain
Ratio of Soils," submitted by Corps of Engi- gages as transducers to record the pressure
neers, U. S. Army, in Procedures for Testing transmitted to the face of the unit. The gages
Soils, American Society for Testing Materials, were cemented to a flexible diaphragm mounted
Philadelphia, April 1958. in the interior of the cell. A schematic diagram
All tests on the AASI-IO Road Test were run of the cell is shown in Figure 7-D.
on freshly exposed surfaces of the particular Pressures were measured with the loaded
layer being tested, in the natural moisture wheel (single axle vehicle) stopped at 6-in, in-
condition. A 30-lb surcharge weight, including tervals from points 2 ft ahead and 2 ft behind
the 10-in, diameter steel ring, was used the location of the cells. In tests using tandem
throughout. Fine silica sand was used to pro- axle loads, observations began 2 ft ahead of the
vide smooth seating, in addition to any leveling first wheel and were continued at 6-in, intervals
required to provide a level test area. Load- to a point 2 ft behind the second wheel of the
penetration curves were plotted and "cor- assembly.
rected" CBR values computed as described.
LAYER THICKNESS CHANGES DENSITY TESTS
During construction of the flexible pavement, Crushed Stone Base
perforated plates were placed 6 to 8 ft below Tests on the crushed stone base were made
the embankment level, on the surface of the with the nuclear density surface probe built on
embankment, and on the surface of the sub- the project. The calibration curve was deter-
base and base in transverse lines across the test mined by compacting samples of base into
288 THE AASHO ROAD TEST, REPORT 5
square steel boxes 1 cu ft in volume and about were generally excavated 4 in. in diameter and
8 in. deep, and plotting the Count-wet density 4 to 8 in. deep depending on the layer thickness.
relationship. Wet densities were converted to Moisture content and weight of material were
dry densities by use of moisture contents deter- determined from material excavated.
mined from oven-dried samples.
Seating of the surface gage was no major Embankment Soil
problem during construction because the steel
wheel rollers left a smooth, flat surface on the Embankment soil densities were taken with
base layer. In subsequent testing on Loop 1 drive-tube samplers of thin-wall tubing, (%2
and the main loops, however, this did become in. thick), cut into 37/8-in. lengths and beveled
a problem since a very rough surface was pro- at one end. They were connected to a drop
duced as the fines in the base stuck to the sur- hammer by a pin for ease of removal. During
facing layer when it was removed. To over- construction two tube samples constituted one
come this problem which led to large errors in test; however, often during traffic period one
determining density, a procedure was developed tube sample was considered as a test. Moisture
whereby a thin layer of minus No. 10 stone tests were taken from the density samples.
dust at about 85 percent moisture was tamped
over the test spot to fill air gaps and provide Maximum Density Tests
a good seat for the gage. Maximum density for granular materials
was determined from full curves of at least
Subbase, Gravel Base, and Cement-Treated three moisture-density points. Soil maximum
Base densities were obtained from one-point tests
Tests on these materials were made with based on full curves. A full discussion is Con-
Rainhart No. 171 rubber balloon volumeters tained in AASHO Road Test Report 2, "Ma-
modified to include pressure gages. Three terials and Construction," Chapter 2 and
pounds air pressure was standard usage. Holes Appendix B.
Appendix E
FORMULAS FROM ELASTIC THEORY USED IN CONNECTION WITH
SECTION 3.5.4
Formulas used for converting gage readings = major principal stress, psi;
to principal strains and for converting principal = minor principal stress, psi;
strains to principal stresses are as follows:
A. Symbols appearing on the formulas are Positive values of stresses and strains indicate
defined as follows: tension.
6, eb and r., are the readings of gages a, b and
B. The strain components E., and yo, were
c, respectively, at a rosette gage point (Fig. obtained from gage readings by the following
1-E). formulas:
Eo = strain parallel to x-axis; At rosette gage points:
= strain parallel to y-axis; =
= shear strain in x-y plane; 1
61 = major principal strain;
EY = — (— E + 2E, + 2)
= minor principal strain; 2
= inclination of major principal strain you (E — Eb)
to x-axis, measured counterclockwise
from x-axis; At gage points along transverse joint:
= inclination of minor principal strain = gage reading
to x-axis (çL i = 02 + 900);
Ey =
E = Young's modulus, psi (value used
herein was the dynamic modulus for YEY =0
concrete pavement at the Road Test,
= 6.25 x 106 psi);
At gage points along edge:
= Poisson's ratio (value used was 0.28 = gage reading
for concrete pavement at the Road to =
Test) Yxu =0
C. E l, 62 and çi were obtained from 6,
Ey and y by the following formulas:
1. At rosette gage points:
V
1 1
Ci = (z + Ey) + V (Cx — e)2 + Yxy2
>
w
' 1\I
a
I
E2 = — (E
cfr (or 4 2)
1
2
+ Cy)
90
- tan-1
' fxy
degrees.
_____
2 + Yxy
2
0.
Whether the last formula yielded pl or 2 was
x determined by testing the value given by the
TRANSVERSE JOINT formula against the following relationships:
I
If y.,y >0,0 < 01 <900
If-y_,y.< 0, —900 <01 < 0.
If y,y = 0 and E> Ey, 4 1 = 0 0 .
Figure 1—E. Rosette gage with nomenclature and If = 0 and E., < Ey, 01 = 900.
coordinate system used in connection with strain data
described in Section 3.5.4. If Ixy = 0 and E. = Ey, does not exist.
289
290 THE AASHO ROAD TEST, REPORT 5
The relative performance of various pave- ticular application of the pavement service-
ments is a function of their relative ability to ability-performance system, it is necessary to
serve traffic over a period of time. There have set down some fundamental assumptions upon
been no widely accepted definitions of perform- which the system is based.
ance that could be used in the evaluation of There is a statement attributed to D. C.
various pavements or that could be considered Greer, State Highway Engineer of Texas:
in the design of pavements. In fact, design "Highways are for the comfort and conveni-
systems in general use in highway departments ence of the traveling public." A reasonable
do not include consideration of the level of inference from this simple statement is that the
performance desired. Design engineers vary only valid reason for any road or highway is
widely in their concepts of desirable perform- to serve the highway users. Another opinion
ance. By way of example, two designers are is that "a good highway is one that is safe and
given the task of designing a pavement of cer- smooth."
tain materials for certain traffic and environ-
ment for 20 years. The first might consider. The opinion of a user as to how he is
his job to be properly done if not a single crack being served by a highway is by-and-large sub-
occurred in 20 years while the second might be jective. There is no instrument that can- be
satisfied if the last truck that was able to get plugged into a, highway to tell in objective
over the pavement made its trip 20 years from units how well it is serving the users. The
the date of construction. There is nothing in measurement of damage to goods attributed to
existing design manuals to suggest that either rough roads may provide an exception to this
man is wrong. This is simply to demonstrate rule but one of minor importance since a road
that any design system should include con- rough enough to damage properly packed and
sideration of the level of serviceability to traf- properly suspended goods would be classed sub-
fic that must be maintained over the life of the jectively so low, by all users that little could
road. How long must it remain smooth and be gained by an objective measure.
how smooth? There are, however, characteristics of
One popular design system involves the highways that can be measured objectively
determination of the thickness of slab required which, when properly weighted and combined,
in order to hold certain computed stresses be- are in fact related to the users subjective
low a certain level. It is clear that cracks will evaluation of the ability of the highway. to
occur if a pavement is overstressed, but no- serve him.
where can be found any reference to the effect The serviceability of a given highway
of such cracks on the serviceability of the pave- may be expressed by the mean evaluation given
ment. Engineers will agree that cracks are it by all highway users. There are honest dif-
undesirable, and that they require maintenance, ferences of opinion even among experts making
but the degree of undesirability seems to have subjective evaluations of almost anything. Thus
been left dimensionless. It may be apparent there are differences of opinion as to which
that one pavement has performed its function automobile in a given price range is best, dif-
of serving traffic better than another, but a ferences among judges of a beauty contest, and
rational answer to the question, "How much differences as to which bank, broker, grocery
better?" has not been available. store, or bar to patronize. Opinion as to the
To provide dimensions for the term "per- serviceability of highways is no exception.
formance" a system has been devised that is Economic considerations alone cannot explain
rational and free from the likelihood of bias these differences.
due to the strong personal opinions of groups Therefore, in order for normal differences
or individuals. It is easily conceivable that of opinion to be allowed with the smallest aver-
such a system could be adopted by all depart- age error for each individual highway user,
ments thus providing for the first time a na- serviceability, may be expressed in terms of
tional standard system for rating highways the mean evaluation of all users.
and pavements.
Before discussing the derivation and a par- Performance is assumed to be reflected
by the serviceability trend of a pavement with
* An adaption of a paper given at the 39th Annual increasing number of axle load applications.
Meeting of the Highway Research Board. It is assumed that the performance of a pave-
291
292 THE AASHO ROAD TEST, REPORT 5
ment can be described if one can observe its consideration. Among these were surface fric-
serviceability from the time it was built to the tion and condition of shoulders.
time its performance evaluation is desired and Test sections at the Road Test were as short
can plot this serviceability record against the as 100 ft—too short for a satisfactory subjec-
traffic the pavement has served. The traffic tive evaluation of their ability to serve traffic
history must include the number of axle loads (most highway users consider a high-speed
and their magnitude sustained by the pave- ride over a pavement necessary before they
ment. will rate it). Thus, objective measurements
that could be made on the short sections had
USE OF THE SERVICEABILITY- to be selected and used in such a way that
PERFORMANCE SYSTEM pavements only 100 ft long could be evaluated
as though they were much longer.
A typical example of the system which has
been in actual field use at the AASHO Road Definitions
Test, is described in this section. Definitions
and detailed steps in the development and use of To fulfill the requirements of the Road Test
a performance index for evaluation of the Road rather ordinary terms were given specific defi-
Test pavements are included. It is emphasized nitions as follows:
that this case is only one of many possible ap- Present Serviceability —the ability of a spe-
plications of the principles involved. It related cific section of pavement to serve high-speed,
to the performance of the pavements only, yet high volume, mixed (truck and automobile)
it would have been easy to extend the system to traffic in its existing condition. (The definition
provide a measure of the sufficiency of the en- applies to the existing condition; that is, on the
tire highway, including grade, alignment, ac- date of rating, not to the assumed condition the
cess, condition of shoulders, and drainage, as next day or at any future or past date.) Al-
well as characteristics of the pavement itself. though this definition applies to the Road Test
and may apply to any primary highway system,
Purpose the system could easily be modified for use with
city streets, farm roads, etc. Obviously, service-
The principal objective for the AASHO Road ability must be defined relative to the intended
Test calls for significant relationships between use of the road.
performance under specified traffic and the Individual Present Serviceability Rating -
design of the structure of certain pavements. an independent rating by an individual of the
To fulfill this objective an adequate and un- present serviceability of a specific section of
ambiguous definition of pavement performance roadway made by marking the appropriate
was required. None was available. point on a scale on a special form (Fig. 1—F).
Special Considerations This form also includes provision for the rater
to indicate whether or not the pavement being
In addition to the four primary assumptions, rated is acceptable as a primary highway. For
certain special considerations relating to the the Road Test application, the rater was in-
specific requirements of the Road Test were structed to exclude from consideration all feat-
included. Inasmuch as the project was designed ures not related to the pavement itself, such as
to provide information relating to the pave- right-of-way width, grade, alignment, and
ment structure only, certain aspects of normal shoulder and ditch condition.
pavement serviceability were excluded from Present Serviceability Rating (PSR) —the
mean of the individual ratings made by the
members of a specific panel of men selected for
the purpose by the Highway Research Board.
5- This panel was intended to represent all high-
Very Good way users. It included experienced men, long
Acceptable ? associated with highways, representing a wide
Good
variety of interests, such as highway adminis-
Yes 3
tration, highway maintenance, a federal high-
F—H Fair
way agency, highway materials supply (cement
No 2-
Poor
and asphalt), trucking, highway education,
Undecided automotive manufacture, highway design, and
Very Poor highway research.
0- Present Serviceability Index (PSI) —a
mathematical combination of values obtained
Section Identification Rating
from certain physical measurements of a large
Rater Date Time Vehicle number of pavements so formulated as to pre-
dict the PSR for those pavements within pre-
Figure 1—F. Individual present serviceability rating scribed limits.
form. Serviceability Trend.—a continuous graph of
APPENDIX F 293
serviceability plotted against axle load applica- raters to ride over the section at high speed
tions. without being influenced by the condition of
Performance —the serviceability trend of a pavement at either end.
section of pavement with increasing number of Field Rating—The members of the panel
axle load applications. are taken in small groups to the sections that
are to be rated. They are permitted to ride
Formulation of a Present Serviceability Index over each section in a vehicle of their choice
A minimum program for the establishment, (usually one with which they are familiar), to
derivation and validation of a PSI (or any simi- walk the pavement and to examine it at will.
lar index that may be considered for another Each rater works independently—there is no
purpose) is as follows: discussion among the raters. When he is satis-
fied as to his rating, he marks his rating card
Establishment of Definitions—There must and turns it in to a staff representative. The
be clear understanding and agreement among group then moves on to the next section. Each
all those involved in rating and in formulation group takes a different route to reduce the
and use of the index as to the precise meanings possibility of bias over the day (raters may
of the terms used. Exactly what is to be rated, rate differently in the afternoon than in the
what should be included, and what excluded morning, therefore, the groups are scheduled
from consideration?. so that some sections are rated by one or two
Establishment of Rating Panel—Because groups in the morning and the same sections
the system depends primarily on the subjective by the other groups in the afternoon). It has
ratings of individuals, great care should be been found that, near metropolitan areas, sec-
taken in the selection of the persons composing tions with satisfactorily different characteris-
the rating group. Inasmuch as serviceability is tics can be found close enough together so that
defined as the mean opinion of this group, it is the raters can travel routes containing about 20
important that the raters represent highway sections per day. When rating present service-
users, and they should be selected from various ability of a pavement, raters have found it help-
segments with divergent views and attitudes. ful to ask themselves "How well would this
Orientation and Training of Rating Panel road serve me if I were to drive my own car
—The members of the panel are instructed in over roads just like it all day long today?"
the part they are to play; they must understand Here again, of course, serviceability is related
clearly the pertinent definitions and the rules to the intended use of the road, primary high-
of the game. It has been found worthwhileto way, city street, farm road, etc.
conduct practice rating sessions where the Replication—It is necessary to determine
raters can discuss their ratings among them- the ability of the panel to be consistent in its
selves. When they make their official ratings ratings. The Road Test panel rated many sec-
they must work independently with no oppor- tions twice, first on one day and again on
tunity for discussion of the ratings until the another day near enough to the first so that the
entire session has been completed. section did not change physically, yet remote
Selection of Pavements for Rating—Be- enough so that all extraneous influences on the
cause ratings are to be made of the service- raters would be in effect. In general, it might
ability of pavements, a wide range of service- be expected that replicate ratings would differ
ability should be represented among the more when separated by several months than
pavements that are selected for rating. More- when separated by only one day. For this
over, there should be among the sections reason, the replication differences observed in
selected pavements containing all of the various the Road Test rating sessions are perhaps to
types and degrees of pavement distress that some degree an underestimate of replication
are likely to influence the serviceability of differences in a larger time reference. The
highways. Before a field rating session, engi- difference between repeated ratings on the
neers study the highway network in the area same section is a criterion for the adequacy of
under consideration (200 mi or less in diame- a present serviceability index derived from
her, for example) and pick sections of road-
way so that a reasonable balance is obtained measurements.
among obviously very good, good, fair, poor Validation of Rating Panel—Because the
and obviously very poor sections. The Road panel is intended to represent all highway
Test system was based on four rating sessions users, it is necessary to test its ability to do so.
in three different states; 138 sections of pave- To a limited extent such validation was ob-
ment were studied. About one-half were flexi- tained for the Road Test panel by selecting
ble pavement; the other half, rigid. The Road other groups of users and having them rate
Test panel agreed that the minimum desirable some of the same sections that had been rated
length of a pavement to be rated was 1,200 ft; by the panel. One such group consisted of two
however, in a few cases shorter sections were commercial truck drivers who made their rat-
included. This length was sufficient for . the ings based on the rides they obtained when driv-
294 THE AASHO ROAD TEST, REPORT 5
ing their own fully-loaded tractor-semitrailer these, or by any number of other summary
vehicles. Another group was made up of ordi- statistics involving variance of the record,
nary automobile drivers not professionally as- power spectral density analysis, etc. Trans-
sociated with highways. For the sections in- verse profile may be summarized by mean rut
volved, these studies indicated that the ratings depth, variance of transverse profile, etc. The
given pavements by the Road Test panel were variance of rut depth along the wheel paths is
quite similar to those that were given by the also a useful statistic. Cracking occurs in dif-
other user groups. Of course, if a greater num- ferent classes of severity as do other measures
ber of sample groups had been studied, more of surface deterioration. Measurements in any
positive statements could be made as to how of these classes may be expressed in one unit
well the panel represented the universe of all or another.
users. Derivation of a Present Serviceability
Physical Measurements—If it is practica- Index—After obtaining PSR's and measure-
ble for the panel to rate all roads in the area ment summaries for a selection of pavements,
often enough, no measurements need be taken. the final step is to combine the measurement
Analyses may be based on the PSR itself. Since variables into a formula that "gives back" or
it was not possible for the panel to rate the predicts the PSR's to a satisfactory approxi-
Road Test sections (ratings were desired every mation. Part of this procedure should consist
two weeks), it was necessary to establish a PSI in determining which of the measurement
or index that would predict the panel's ratings. summaries have the most predictive value and
To accomplish this, measurements of certain which are negligible after the critical measure-
physical characteristics of the pavements were ments are taken into account. The technique of
necessary. To determine which measurements multiple linear regression analysis may be used
might be most useful, the members of the panel to arrive at the formula, or index, as well as to
were asked to indicate on rating cards which decide which measurements may be neglected.
measurable features of the roadway influenced For example, a longitudinal profile summary
their ratings. It was apparent that present may be sensitive to faulting so that faulting
serviceability was a function primarily of longi- measurements need not appear in the index
tudinal and tranverse profile with some likeli- formula whenever this profile measure is in-
hood that cracking, patching, and faulting cluded.
would contribute. Therefore, all of these char- The decisions as to which terms should be in
acteristics were measured at each of the 138 the serviceability formula and which terms
sections that were rated by the panel. Several should be neglected may be made by comparing
other objective measurements could have been the lack of success with which the formula
added to the list if other phenomena were per- gives back the ratings with a pre-selected cri-
mitted consideration by the established rules of terion for closeness of fit, such as the Panel's
the game. Skid resistance, noise under tires, replication error. There is no justification for
and shoulder and ditch conditions might be in a formula that can predict a particular set of
this category. ratings with greater precision than the demon-
Measurements fall rather naturally into two strated ability of the panel to give the same
categories: those that describe surface defor- ratings to the same pavements twice. There-
mation and those that describe surface deteri- fore, the multiple linear regression analysis
oration. Of course, phenomena in the second will yield a formula that will combine certain
category may or may not influence measure- objective measurements to produce estimates
ments in the first category. Measures of sur- of the panel's ratings to an average accuracy
face deformation will reflect the nature of no greater than the panel's average ability to
longitudinal and transverse profiles, or may repeat itself.
represent the response of a vehicle to the pro-
file, as does the BPR roughometer. Supple- Performance
mental profile characteristics, such as faulting The serviceability index is computed from a
will ordinarily be measured. Present and past formula containing terms related to objective
surface deterioration will be reflected through measurements that may be made on any section
measures of cracking, spalling, potholing, of highway at any time. At the AASHO Road
patching, etc., and may include phenomena Test, these measurements were made and the
whose influence on present serviceability rat- index computed for each test section every
ings range from negligible to appreciable. two weeks. Thus a serviceability-time history
Summaries of Measurements—There are is available for each test section beginning at
many different ways to summarize longitudinal the time test traffic operation was started. The
and transverse profiles. For example, longi- present serviceability values range in numeri-
tudinal profile may be expressed as total devi- cal value from 0 to 5 (Fig. 1—F).
ation of the record from some base line in To fulfill the first Road Test objective of
inches per mile, number of bumps greater than finding relationships between performance and
some minimum, some combination of both of pavement structure design, some summariza-
APPENDIX F 295
tion of the serviceability-time history is im- ments for each section. Summaries for all
plied. Performance may be said to be related evaluations of the 74 flexible pavement sections
to the ability of the pavement to serve traffic are given in Table 1—F, and corresponding
over a period of time. A pavement with a low evaluations for the first 49 rigid pavements are
serviceability during much of its life would given in Table 2—F.
not have performed its function of serving Although the panel members had indicated
traffic as well as one that had high service- that rutting in flexible pavement must influ-
ability during most of its life even if both ulti- ence serviceability, the first three rating ses-
mately reached the same state of repair. sions did not include pavements with rutting
Performance, at the Road Test, was defined severe enough to contribute significantly to the
as the trend of serviceability with increasing pavement serviceability. Since severe rutting
load applications. Analysis of performance occurred at the Road Test it was necessary to
was based on mathematical models for express- assemble the panel for a fourth session in which
ing the serviceability trend in terms of design, sections with severe rutting were rated. Re-
load, and number of load applications. The analysis of the data from all four sessions then
procedures for analysis are discussed in Ap- made it possible to determine the effect of
pendix G. rutting on serviceability. A second objective
of the fourth session was to rate a small num-
ber of rigid pavements only for the purpose
ROAD TEST INDEXES of checking present serviceability indexes de-
rived from the first 49 sections. For these
The techniques previously described were reasons, flexible pavements from all four ses-
used in the derivation of present serviceability sions appear in Table 1—F; Table 2—F includes
indexes for the AASHO Road Test. This sec- only rigid pavement sections from the first three
tion includes tabulations of the actual data ob- sessions.
tained in the field rating sessions by the Road Present serviceability ratings shown in the
Test Rating Panel and data obtained from the third column of Tables 1 and 2 are mean
objective measurements of the pavements values for individual ratings given by the Road
rated. Relationships among the ratings and Test panel. In general, each mean represents
various measurements are shown graphically about ten individual ratings. For both pave-
and the results of the regression analyses in ment types, the PSR values range from about
which the serviceability indexes were derived 1.0 to 4.5 with nearly the same number of sec-
are given. tions in the poor, fair, good, and very good
The matter of precision required of an index categories (Fig. 1—F). The grand mean PSR
and precision attained in the Road Test indexes for all rated pavements was slightly less than
is discussed. Alternate measurement systems 3.0 for both pavement types.
are mentioned for the benefit of agencies not Over forty of the sections were revisited by
able to equip themselves with elaborate instru- the panel during the same rating session, and
ments. differences between first and second mean rat-
ings are shown in the fourth columns of Tables
Ratings for Selected Pavements 1 and 2. The replication differences ranged
After establishing concepts, ground rules, from 0 to 0.5; the mean difference was less
and rating forms for present serviceability than 0.2 for both flexible and rigid pavements.
ratings, the AASHO Road Test performance The fifth columns give the standard deviation
of individual PSR values for each section.
rating panel rated 19 pavement sections near These standard deviations are of the order
Ottawa, Ill, on April 15-18, 1958, 40 sections 0.5, an indication that only about two or three
near St. Paul-Minneapolis on August 14-16, individual ratings (out of ten) were farther
1958, 40 sections near Indianapolis on May than 0.5 rating points from the panel mean
21-23, 1959, and 39 sections on and near the PSR.
Road Test on January 20-22, 1960. Ten Illi- The mean ratings of the two truck drivers
nois sections, 20 Minnesota sections, 20 Indiana who rated certain Illinois sections are shown
sections and 24 sections on and near the Road in the sixth columns. The seventh columns
Test were flexible pavements; all remaining show mean ratings given to selected Illinois
sections were rigid pavements. Each section sections by a group of about 20 Canadian
was 1,200 ft long except those on the Road raters. The general agreement among the vari-
Test which averaged 215 ft. With the coopera- ous rating groups is apparent.
tion of the respective state highway depart- The eighth and ninth columns represent sum-
ments, sections were selected to represent 'a maries of the AASHO Panel response to the
wide range of pavement conditions. acceptability question (Fig. 1—F). The tables
Coincident with the rating session, Road give what fraction of the, panel decided the
Test crews and instruments were used to ob- present state of a particular pavement section
tain condition surveys and profile measure- to be acceptable and what fraction decided the
TABLE 1-F
DATA FOR 74 SELECTED FLEXIBLE PAVEMENTS
TABLE 2-F
DATA FOR 49 SELECTED RIGID PAVEMENTS
U,
w
>-
.8
.6
0 0
1.0 1.0
0
z
0' 8 0' .8
C
0
0
a,
a,
g .6 c
a 0
a
- - ----5 --.
0
I
.4
IIIIIIi.2II ___ 0
C
0
t .2
0
LL
.
1
0- = [ii
111111
0 .8 1.6 2.4
hrr
3.2 4.0 4.8
Present serviceability rating Present serviceability rating
Figure 4-F. Unacceptability vs present serviceability Figure 5-F. Unacceptability vs present serviceability
rating; 74 flexible pavements. rating; 49 rigid pavements.
300 THE AASHO ROAD TEST, REPORT 5
ing, and finally a measurement of the patched ible pavement sections. This gage is used to
area found in the section. determine the differential elevation between
The symbol SV is used for the summary the wheelpath and a line connecting two points
statistic of wheelpath roughness as measured each 2 ft away (transversely) from the center
by the Road Test longitudinal profilometer. For of the wheelpath. Rut depth measurements
each wheelpath the profilometer produces a were obtained at 20 ft intervals in both wheel-
continuous record of the pavement slope be- paths. Average rut depth values, RD, for the
tween points 9 in. apart. For a particular flexible sections are given in Table 1—F; the
wheelpath, the slopes are sampled, generally values range from 0 to nearly 1 in. Variances
at 1-ft intervals, over the length of the record. were calculated for the rut depths in each
A variance* is calculated for the sample slopes wheelpath, then the two wheelpath variances
in each wheelpath, then the two wheelpath were averaged to give the RDV values (Table
slope variances are averaged to give SV. 1—F). Figure 8—F shows the correlation be-
A Bureau of Public Roads roughness indica- tween SV and RDV for the 74 flexible sections.
tor, or roughometer, was adapted for use at the Profile information for rigid pavements in-
AASHO Road Test, but this development was cluded a measure of faulting in the wheelpaths.
not made until just before the Indiana rating These measurements are given in Table 2—F
session and still more developmental work was expressed in total inches of faulting (in wheel-
done on the AASHO roughometer after the paths only) per 1,000 ft of wheelpath.
Indiana session. The AASHO roughometer has The remaining measurements for flexible
a modified output and was operated at 10 mph, pavement sections are given in Table 1 in terms
so that roughometer values shown in Tables 1 of area affected by class 2 and class 3 crack-
and 2 are not the values that would be obtained ing, length of transverse and longitudinal
with the BPR roughometer at 20 mph. Never- cracks, and patched area, where areas and
theless, roughometer values in inches per mile lengths are expressed per 1,000 square feet of
are given; the roughometer values averaged for pavement area. Corresponding measurements
both wheelpaths, AR, are correlated with the for rigid pavements are shown in Table 2—F
corresponding mean slope variances. Figures in terms of length of class 2 and sealed cracks,
6—F and 7—F show the extent of this correlation spalled area, and patched area. Lengths for
for the last two rating sessions. rigid pavement cracks were determined by
One other instrument, a rut depth gage, was projecting the cracks both transversely and
used to obtain profile characteristics of the flex- longitudinally, choosing the larger projection,
then expressing the accumulated result in feet
* The variance of a set of N sample values, Y1, Y2, per 1,000 sq ft of pavement area. Only spalled
..., Y x is defined to be the sum of all N squared devia- areas having more than 3-in, diameters were
tions from the mean divided by N - 1. Thus the considered, and both spalling and patching are
variance of Y is Z (Y - Y) 2/(N - 1), where Y =
YIN is the sample mean. expressed in square feet per 1,000 sq ft of pave-
0
me
80
0
---
0) o 0
a. 40
0
'0
C
0
20
0
0 CPO 0
008
o 0
100
80
0
0
60
0
o
0 o
0 o 0 0 0
m 40
o
C 0
0
0
20 0
0 0 : 0
0
0 20 240 . 360 480 600 720 840
2)
Total cracking and patching (per 1000 ft
Figure 9—F. Mean slope variance vs cracking and patching; 74 flexible
pavements.
302 THE AASHO ROAD TEST, REPORT 5
II
60 - -r -
C
0
0 0
'0
0
0,
C
0 0 0
0 % 0
0
20
o
LO
40 80 120 160 200 240 280
2)
Total cracking and patching (per 1000 ft
Figure 10—F. Mean slope variance vs cracking and patching; 49 rigid
pavements.
deleting terms in Eq. 1—F will be to decrease 4.0, there is no way to infer what pavement
or increase the sum of squared residuals. The characteristics must be like in order to produce
change in residual sum of squares can be used a value of 1.0 or 5.0, except to extrapolate the
to deduce the significance of adding or drop- index on the assumption that linearity holds
ping terms from the index formula. over the full range of pavement characteristics.
The model for PSI is linear in that if all For these reasons it has been stated that
functions save one are given a numerical value, selected pavements should show all phenomena
then PSI versus the remaining function repre- of interest, the.. complete range of interest for
sents a straightline relationship. For this each phenomenon, and should be associated
reason it is desirable to choose functions R1 , with PSR values that span the full range of in-
R2,..., D1 , D 2,..., that have linear graphs terest. Therefore, pavement selection amounts
when plotted with PSR values. For example, to the assumption that all interesting phe-
logarithms and powers of the original meas- nomena and ranges have been encompassed by
urements may be used as linearizing transfor- the selections. Extrapolations of the index to
mations. measured values outside the range of those
A present serviceability index developed found in the selected pavements amounts to the
from observed ratings and measurements can assumption that the index formula remains
only reflect the characteristics that were linear in the region of extrapolation.
actually present in the observed pavements.
For any particular characteristic, the index can Choice of Functions for the Present Service-
only reflect the observed range of values for ability Index
the characteristic. For example, if the selected Measurements from the Illinois and Minne-
pavements had no potholes, there is no objec- sota sections were plotted in succession against
tive way to infer how potholing would affect corresponding PSR values to determine which
the present serviceability ratings, and the measurements were essentially uncorrelated
index cannot contain a function of potholing. with PSR and to deduce the need for lineariz-
As another example, if faulting in the selected ing transformations. It was indicated that the
pavements ranged from 0 to 10, there would be mean wheelpath slope variance, SV was highly
no way to infer the effect on PSR of pavements
whose faulting was in the range 50 to 100.' correlated with PSR, though curvilinearly. Fig-
This same argument applies to the present ures 11—F and 12—F show the nature of this
serviceability ratings themselves. If PSR's for correlation for all selected pavements. From
the selected pavements range only from 2.0 to several alternatives, the transformation
* It was for this reason that it was not possible to R, = log (1 + SV) (3—F)
determine the effect of rutting in flexible pavements
after the first three rating sessions which included was selected as the first function of profile
pavements with rutting ranging from 0 to only 0.37 in. roughness to appear in the PSI model for both
Thus the fourth rating session was necessary to deter- flexible and rigid pavements. The result of this
mine the effect of ruts in the range of 0.5 to 1.0 in.
deep. transformation is shown in Figures 13—F and
APPENDIX F 303
so
5.0
ccg
4.0
99,
0-
0
-
c2
00
0 0
00
0 00— 0
i
3.0
0
30
>
0,
0
0 0 0
-a-
0
- —%--
00
— -— — — 0 0 0
.-0
0
0
0 0
2.0 0
-
0
0
0 0
0
C
0,
20 --_.2_ _...2 0
CO 0
0 0
66 U,
0,
0
0 0
0
0
00
0 --
0 20 40 60 80 100
0
Mean slope variance (x 106) 0 20 40 60 80
Figure 11-F. Present serviceability rating vs slope Mean slope variance (x 106)
variance; 74 flexible pavements.
Figure 12-F. Present serviceability rating vs slope
variance; 49 rigid pavements.
14—F where PSR values are plotted against R,
for flexible and rigid pavements, respectively. The scatter diagram of PSR vs RD2 is shown
For the flexible pavements, mean wheelpath in Figure 15—F.
rut depth, RD, was included as a second pro- Although preliminary analyses considered
file measurement to appear in the PSI equation. the possibility of several functions, of surface
The selected function of rut depth was deterioration, for example, one function for
R2 = RD (4—F) each of the measured manifestations, it was ap-
--- I --
5.0 5.0
0og00 000
0
---0—-----
0
00
00
0
4.0 — - — 0 00000-- — — - — -
0
0
0 0
0 0
--- -- 0
0
0 - -
-- i
—.--
30 - - - -
0
— - - —0, ;- 0C 0
a
0, 00a, 3. 0 •
00
000 0—00 —0 - — -
—0
0, 00 0
00 0
> 0 0 0
00 0 0 0 —0000 --
0
a,
0 00 Q0, 0
" 20 2.0 0
.0
0
U,
0,
-- -- 0 000 o 0
0
-00
00 0
- - - a-
1.0 0 0
0
0,3
00
0
0 4 .8 12 16 2.0
00 .8 4 1.2 1.6 2.0
log (1+ mean slope variance) log (I + mean slope variance)
Figure 13-F. Present serviceability rating vs log Figure 14-F. Present serviceability rating vs log
(1 + mean slope variance); 74 flexible pavements. (1 + mean slope variance); 49' rigid pavements.
304 THE AASHO ROAD TEST, REPORT 5
parent that no loss would be incurredby lump- fair degree of correlation exists between these
ing all major cracking and patching into a variables for the rigid pavements. For both
single number to represent surface deteriora- flexible and rigid pavements the transforma-
tions. Values for C + P are not shown in tion
Tables 1—F and 2—F, but may be obtained from D— VC+P (5—F)
the cracking and patching measurements.
Scatter diagrams for the PSR versus C + P was selected as a linearizing transformation
are shown in Figures 15—F and 16—F. for C + P (Figs. 17—F and 18—F).
For whatever reasons, it is apparent that Thus the present serviceability index models
there is little correlation between PSR and to be used are
C + P for the flexible pavements, but that a
For flexible pavements:
PSI = A0 + A 1R1 + A 2R2 + B1D1 =
A0 +A1 log(1+SV) +A2RJY+B1VC+P
5.0
(6—F)
For rigid pavements:
a 0
0
4.0 So
00 PSI = A0 + A 1R1 + B1D1 = A0 + A 1 log
(1 + SV) + B1 VC + P
C
- s— -0
(7—F)
0
0
% 0
00 0
0,
0000
0
3.0
.0
8 °° 0 It is not expected that the coefficients A0, A1,
0
0, -00 0000 0
—a—
0
and B1 have the same values for both equations.
U
0 There are many other possibilities for Eqs.
2.0 0
6—F and 7—F--other instruments might be used
0 0 to detect deformation and deterioration, and
-0 2. -0 - - - summary values other than SV, C + P and
00 0 0
RD might be used. Moreover, different func-
°- 1.0 - 0
tions of SV, C+ P and RD could be chosen, or
more functions of pavement measurements
could be included.
0
One of the most important elements of pave-
0 .20 .40 .60 .80 1.00 ment serviceability is its longitudinal profile in
Mean rut depth squared (in 2 ) the wheelpaths. The profile of the road coupled
Figure 15—F. Present serviceability rating vs mean with the appropriate characteristics of the ve-
depth squared; 74 flexible pavements. hicle (mass, tires, springs, shock absorbers,
5.0
o
- 0
4.0 .8— -
00 00
0
- - - - —s.-.
0 0 0000
3 o__5
0 0 dD 0
0
0,
U
C,
C
2.0
C,
C,
a.
19
0
0 4 8 12 16 20 24 28
Square root of cracking and patching (per 1000 ft 2
Figure 16—F. Present serviceability rating vs square root cracking and
patching; 74 flexible pavements.
APPENDIX F 305
speed, etc.) produce the "ride" attained in that profile and its ordinate represents the rate of
vehicle over that road. The actual profile of change of displacement, or slope of the road at
the wheelpath as though taken with rod and any point. The second derivative of the dis-
level at very close spacing is called the displace- placement profile is the "acceleration" profile,
ment profile, p. The first derivative of the dis- p", and represents the rate of change of slope,
placement profile is the profile of the slope, p'. and the third derivative is the "jerk" profile,
A plot of the slope profile has the same abscissa p", the rate of change of acceleration. It has
(distance along the road) as the displacement been suggested that jerk may be more highly
5.0
'40
0
0
3.0
0)
o o
-C 2.0
0
0
a.
o
IN
H
0
- 0 4 8 12 16 20 24 28
Square root of cracking and patching (per 000 ti.2 )
5.0
4.0
— 3.0
U;
0 2.0
*
IME NEI
10
____—_
Nov. Dec. Jon. Feb. Mar. Apr. May June July Aug. Sept.
5.0
4.0
_—__--
- 3.0
U
ii
a: 2.0
10
1
z:::-
u•uuu
Nov. Dec. Jon. Feb. Mar. Apr. May June July Aug. Sept.
correlated with a rider's opinion of his ride 7—F) has only three undetermined coefficients,
than any of the other representations. Perhaps only three simultaneous equations need be
this is true if one is seeking to define "ride"— solved. Their solution gives the index
but the efforts at the Road Test were directed
towards a definition of the "smoothness of a PSI = 5.41 - 1.78 log (1 + ) -
road" independent of the vehicle that might use 0.O9VC+P
it. Considerable effort was spent in studying (10—F)
correlations of the variances of various profile
derivatives with the present serviceability rat- The multiple squared correlation coefficients
ings, but there was no evidence that elevation for these derivations are r2 = 0.844 for the
variance, acceleration variance, or jerk vari- flexible pavements, and r2 = 0.916 for the rigid
ance has higher correlation with PSR than the pavements.
slope variance. On the other hand, when a Therefore, the PSI formulas account for 84.4
number of the slope profiles were subjected to percent and 91.6 percent of the variation in
generalized harmonic analysis to determine PSR for flexible and rigid pavements, respec-
how variance was associated with the wave- tively. The respective root mean square resi-
length spectrum, there was some indication duals are about 0.38 and 0.32, respectively.
that slope variance in certain regions of the The last columns of Tables 1—F and 2—F
wavelength spectrum is more highly correlated show calculated values for the present service-
with PSR than is the total slope variance. ability indexes as well as for residuals. At the
Coefficients for the Present Serviceability bottom of the last column, the mean residual
Index was 0.30 for flexible pavements and 0.26 for
rigid pavements. In both cases, the mean resi-
Substitution of Eq. 6—F into Eq. 2—F gives dual is about twice the mean difference between
for flexible pavements replicate ratings given by the AASHO rating
panel.
PSR5 =A0 + A 1R1, + A2R21 + B1D1, + E, From the residual columns, six flexible and
(8—F) three rigid pavement residuals exceeded 0.5,
in which the largest replication difference given by the
panel. However, the index formulas span rat-
R11 = log (1 + SV,), R2, = RD,2 and D1, ings made more than a year apart whereas all
VC5 + P1 for the jth pavement. replicate ratings were made on successive days.
As stated before, it is quite possible that repli-
Least squares estimates for A0 , A1, A.2 and B1 cate PSR's would be more different when made
are found by minimizing the sum of squared over longer intervals of time.
residuals, E, through solving four simultane- When the 15 rigid pavement PSR values
ous equations for A0, A1, A2 and B1. The solu- from the fourth rating session were compared
tion of these equations gives the index with PSI values given by Eq. 10—F, the sum of
the algebraic deviations was practically 0
PSI = 5.03 - 1.91 log (1 + SV) whereas mean discrepancy was 0.3. Since only
1.38 RD2 -0.01 VC +P two of the deviations exceeded 0.5, it was in-•
(9—F) ferred that Eq. 10—F fitted the new PSR values
to about the same degree as it predicted those
Because the model for rigid pavement (Eq. from which it was derived.
Appendix G
A RATIONALE FOR ANALYSIS OF PAVEMENT PERFORMANCE*
The first objective for the AASHO Road Test The concepts and specifications to be described
is to find significant relationships between pave- in this area have evolved after consideration
ment performance and certain characteristics of many alternatives.
of pavement design and applied loads. To carry It is supposed that the present serviceability
out this objective detailed specifications are history of a pavement section plays a very use-
needed in three areas. First, pavement ful role in performance evaluation. At any
m st be defined so that erformance particular time the section's present service-
data can be obtained or everyjection in ability is a measure of its ability to serve high
the investittSnthere must be ex- speed, high volume traffic, and in Appendix F
perimental designs that give details for pave- a system for the development of present serv-
ment design and load characteristics of the iceability index formulas was described. Sepa-
sections. Finally, it is necessary to set out rate formulas were presented for flexible and
definite procedures that lead to the required for rigid pavements. When appropriate meas-
relationships. Several papers and talks have urements of surface deformation and deteriora-
described Road Test specifications in the first tion are made on day t, then substitution of the
two areas; and it is the main purpose of this measurements into the index formula gives an
paper to discuss specifications in the third area. index value Pt' for the index day. The complete
However, the three sets of specifications are in- serviceability history of a pavement section
terrelated in that analytical procedures are consists of index values for a series of index
determined to a large extent by the nature of days that begins when the section is first con-
the experimental designs and by the nature of structed and that, ends when serviceability loss
the performance data. For this reason pave- is such that major maintenance or replacement
ment performance and experimental designs is required. In both the illustration and the
are discussed before turning to a rationale for AASHO Road Test, serviceability index values
analysis. A numerical illustration that differs are obtained for every section on biweekly
from the AASHO Road Test pavement perfor- index days, and the serviceability history of a
mance studies in certain details but not in section is considered to be completed if and
principle is used. As a consequence, rationale when its index falls to 1.5 on a scale where
for the illustration is applicable to the Road maximum serviceability is 5.0. Although not
Test, and unless specific reference is made to all biweekly index values are plotted, Figure
the illustration, the following discussion per- 1—G, which shows the serviceability histories of
tains to the Road Test. two sections used in the illustration, indicates
It is evident that there are alternatives for a completed history for section 3212 after
virtually every specification that may be given about 17 index days. As in the case of the
in any of the three areas; thus there are many AASHO Road Test, it is assumed that the illus-
possibilities for the total set of specifications. trative road test is stopped after 55 index days
Because it may be supposed that a number of with the expectation that at least some sections
these possibilities are equally acceptable for will still have high serviceability at the end of
meeting the first objective of the Road Test, it the test. One such section is shown in Figure
cannot be claimed that the rationale to be de- 1—G where section 3222 has a serviceability
scribed represents the best, nor the only way to index of about 3.2 after 55 index days.
satisfy the objective, but it is assumed that The general continuous pattern of a service-
any other acceptable rationale would produce ability history is called a smoothed serviceabil-
essentially the same conclusions. ity history. Smoothed histories for the two
sections in Figure 1—G are indicated by the
solid lines. The smoothed history for a section
PAVEMENT PERFORMANCE DATA is defined by a moving average that includes
at least three (generally five) successive index
Inasmuch as a rationale for analysis is rather values and that uses the end values for the
meaningless unless the data that go into the history as end values for the smoothed history.
analysis are well defined, it is necessary to pin
down the specific nature of performance data. Smoothed serviceability history values on index
days will be denoted by Pt.
* An adaption of a paper given at the 40th Annual A second element of performance for a pave-
Meeting of the Highway Research Board. ment section is its history of load applications.
307
THE AASHO ROAD TEST, REPORT 5
5
very good section 3222
4
good
3 'biweekly index F
fair section 3212' values, Pt
2 history, p - - - - - - -
poor
—s-Pavement overlayed or replaced at this -
very poor level; history of orig. pvrnt. no longer observd
0
0 10 20 30 40 50 60
Index day, t
Figure 1—G. Present serviceability histories for two illustrative pavement sections.
Although theories* and procedures exist for ing function. Relative to a specified norm, or
dealing with mixtures of axle loads, reference base, it may be supposed that the conditions at
in this paper to any particular number of ap- any time or location are either normal, better
plications implies that each application repre- than normal, or worse than normal. It is con-
sents the same axle weight. For the illustra- sidered that the seasonal weighting function
tion, Figure 2-G gives both the number of axle reflects serviceability loss potential, and that
load applications between successive index days any particular section may or may not lose
and the accumulated number of applications serviceability during a period when the weight-
for any index day. The respective notation for ing function is high. No specific formula for
these two quantities is nt and N. If more than a weighting function will be given in this
one traffic lane is represented by flt and N, it paper, but it is supposed that such a formula
is assumed that lane to lane variation in nt is has been evolved to give values, q, for every
negligible and lii is averaged for all lanes be- index period (Fig. 3-G). This function pre-
fore the accumulation, N. Whenever it is nec- sumably depends in general on changes in
essary to evaluate accumulated applications moisture-temperature states, and has the value
between index days, linear interpolation is per- qt = 1.0 for normal conditions A value of zero
formed between successive values of N. is considered to be a lower bound at which no
Before specifications are given for perform- serviceability-loss potential exists for any pave-
ance data, ohe more history is discussed—a ment-load combination.
history that is associated with the general state The seasonal weighting function (Fig. 3-G)
of environmental conditions at any particular
time. This history is called a seasonal weight- averages about 1.0, so that environmental con-
ditions for the two years average normal even
* Scrivner, F. H., "A Theory for Transforming the though there is much seasonal variation. Rela-
AASHO Road Test Pavement Performance Equations tive to the selected location, this index might
to Equations Involving Mixed Traffic." HRB Special
Report 66 (1961). not average 1.0 at a second location, whether
1,200,000
1,000,000
800,000
Accumulated
Number of nt Axle Load
Axle Load 30,000 600,000
Applications Applications,
Between 20,000 400,000 Nt = Eflt
Successive
Index Days, 1 0,000 200,000
nt
10 20 30 40 50
Index Day,
Figure 2—C. Axle load application history for the illustration.
APPENDIX G 309
••u•••r&••uu
Wt = qt nt can be obtained by multiplication of
index day ordinates from Figures 2—G and 3—G.
Also, W is assumed to be the accumulation of
weighted axle load applications through any
01,11111
11111
ONOMMUMEMEME
index day. Graphs for both Wt and Wt are 0 10 20 30 40 50 60
shown in Figure 4—G. If the weighting func- Index day,
tion were taken to be 1.0 on every index day,
then the curve in Figure 3—G would be hori- Figure 3—G. Seasonal weighting function.
zontal at unit height, and Figure 4—G would
be identical with Figure 2—G. Thus, Nt is a Summarizing the definitions of the various
special case of W if q is always 1.0. In all the serviceability-time-applications relationships:
discussion that follows accumulated axle load
applications are represented by W but any dif- Serviceability history is the plot of observed
ference between W and N depends on the values of serviceability p,' on a time scale;
values prescribed for q. Smoothed serviceability history is the plot of
All of the variables have values that are ob- the 5-point moving average of the service-
served and computed at points in time. If ability history values on a time scale and
smoothed serviceability values for a pavement smoothed history values are designated by Pt;
section are plotted against accumulated axle Serviceability trend is the plot of smoothed
applications rather than against time, the re- serviceability history values p on an accumu-
sultant curve is called the section's serviceabil- lated axle application scale W where axle ap-
ity trend. Coordinates of points on the service- plications may be weighted or unweighted; and
ability trend are denoted by p and W, and the the
trend of p with W is defined to be the pave- Performance of a pavement is given by its
ment's performance. In other words, service-
ability trends are considered to be performance serviceability trend.
curves that show how pavements are affected The final step in the specification of perform-
by applied loads. ance data is to assume that for numerical
Trend plots for the two sections of Figure analysis a small number of pairs of coordinates
i—G are shown in Figure 5—G for the case when from any trend curve can be selected to repre-
applications are not weighted; that is, when sent satisfactorily the curve. In the Road Test
Vt = 1. Coordinates for the trend curves in rationale five pairs of coordinates were selected
Figure 5—G were obtained from ordinates of from every trend. If the trend was complete
Figures 1—G and 2—G on common index days. (i.e., p had fallen to 1.5) then the trend was
Similarly Figure 6—G shows trend curves for represented by five svalues that spanned the
the same sections when the seasonal weighting range of p. Specifically, W was noted when p
function of Figure 3—G is used to obtain W. was 3.5, 3.0, 2.5, 2.0, and 1.5. In the case of
That is, coordinates for Figure 6—G were ob- incomplete serviceability trends (p at the end
tained from ordinates of Figures 1—G and 4—G of the Road Test was greater than 1.5) the
on common index days. observations were spanned by noting pairs of
60,000 1,200,000
50,000 1,000,000
Number of
Weighted Accumulated
40,000 800,000
Axle Load Weighted
Applications 30,000 600,000
Axle Load
Between
Successive 20,000 400,000 Applications,
Index Days, Wt: Ewlt
wt v•t nt 10,000 200,000
0 10 20 30 40 50
Index Day,
Figure 4—G. Weighted axle load applications for the illustration (seasonal weighting
function from Fig. 3—G).
310 THE AASHO ROAD TEST, REPORT 5
5
section 3222
section 3222
a.
>'3
section 3212 sec'tion 3212
0 2 4 6 8 10 12
0 2 4 6 8 10 12
W (hundreds of thousands)
W (hundreds of thousands)
Figure 6-G. Performance curves for the two illustra-
Figure 5-G. Performance curves for the two illustra- tive pavement sections of Figure 1-G (seasonal
tive pavement sections of Figure 1-G (v, = 1). weighting function from Figure 3-G).
W and p at specific times (at 11, 22, 33, 44 and EXPERIMENTAL DESIGNS FOR PAVE-
55 index days). In both cases it is more con- MENTS AND LOADS
venient to record and use all W values in
logarithmic form so that recorded performance As details are given in the second area of
data appear in the form p, log W. Therefore, specification for the illustration, the reader who
if p = 2.5 when W = 200,000 applications, the is familiar with experimental designs at the
recorded performance data would be 2.5 and AASHO Road Test will recognize that the
5.30 for p and log W, respectively. illustration parallels in principle the main
In the example only three pairs of coordi- factorial experiments of the Road Test.
nates are used to represent serviceability It is assumed that the illustrative road test
trends. For the complete curves W is noted involves three rigid pavement tangents, 1, 2,
when p = 3.5, 2.5 and 1.5, and for incomplete and 3, each having two 12-ft traffic lanes, 1 and
trends W and p are noted at 15, 35 and 55 in- 2, on either side of its centerline. Axle load
dex days. For the two sections shown in specifications for the six traffic lanes are as
Figure 1-G, Table 1-G gives performance data follows: tangent 1, 4-kip single in lane 1, 8-kip
using both weighted and unweighted applica- single in lane 2; tangent 2, 16-kip single in
tions. lane 1, 30-kip tandem in lane 2; tangent 3, 24-
kip single in lane 1 and 36-kip tandem axle
vehicles in lane 2. Figure 2-G gives the illus-
TABLE 1-G trative specifications for frequency of axle load
applications over a 2-yr period.
PERFORMANCE DATA FOR Two ILLUSTRATIVE SECTIONS
OF FIGURE 1-G
The objective for the illustration is assumed
to imply that differences in pavement design
for test sections will be determined by only
Performance Data two factors, thickness of portland cement con-
Section t' For v, = 1 For v, from crete surfacing and thickness of a granular
Fig. 3-G subbase material. All other specifications for
p logW p logW basement soil, pavement materials, and con-
struction procedures are assumed identical for
32122 13.5 3.5 5.39 3.5 5.36 every test section.
16.0 2.5 5.49 2.5 5;54 Three fundamental principles of experi-
17.0 1.5 5.52 1.5 5.59 mental design are balance, replication, and
3222 15 4.2 5.45 4.2 5.48 randomization; these principles are to be used
35 4.0 5.89 4.0 5.81 in the design of the illustrative road test. The
55 3.2 6.08 3.2 6.05 principle of balance is used to rule out unde-
1
sired confusion among the effects of experi-
= Index day at which smoothed serviceability his- mental factors on performance. The effect of
tory equals p.
2
Complete history. a factor means a change in performance that
'Incomplete history. can be attributed to a change in the factor; for
APPENDIX G 311
example, a surface thickness effect is a change much the observations are influenced by resid-
in performance that is clearly attributable to ual variables that are uncontrolled. Replica-
a change in surface thickness. It is assumed tion can be performed in many categories; for
that balance should be maintained in each test example, the illustrative road test might be
tangent for surface thickness and for subbase repeated in toto at a different location, or at a
thickness, so that the analysis can determine different time, or both. At a selected location
whether performance differences are due to and time, any tangent might be completely
one or the other of these factors or possibly replicated by including a fourth tangent that
to their interacting effect. In the absence of has the same specifications as one of the tan-
prior knowledge about their interacting effect, gents in Table 2-G. An axle load might be
a sound experimental design for surface and replicated in both lanes of the same tangent,
subbase thickness is the complete factorial or serviceability index values might be repli-
experiment that includes all possible combina- cated for any index day. If there is sufficient
tions of levels selected for these two factors. replication in any category where conclusions
In each tangent each factor is assigned three are to be drawn about the effects of controlled
levels (that is, three values). Then the com- factors within the category, it becomes possible
plete factorial experiment in each tangent re- to discern between performance changes that
quires 3x3, or nine, different pavement designs. can be attributed to controlled effects and those
As indicated in Table 2-G, levels for subbase changes that must be attributed solely to un-
thickness are 3, 6, and 9 in. in each tangent, controlled or residual effects. For the latter,
but levels for surface thickness are selected so replication provides estimates needed to assess
that thicker pavements are used for heavier the reliability of controlled effects.
axle loads, there being one common surface In the illustration it is assumed that cost
thickness, 5.5 in., across all three tangents. considerations prohibit replication of the whole
Thus, although balance is maintained for sur- experiment, replication of tangents, and repli-
face and subbase thickness in each tangent and cation of lanes, but that replication will occur
loads are balanced with subbase thickness for certain pavement designs within each tan-
across all tangents, load and surface thickness gent. Table 2-G indicates that two different
levels are unbalanced so that extraneous sur- pavement designs are to be once replicated
face thickness-load combinations will not oc- within each tangent; thus there are to be
cur. However, the load effect can be observed eleven test sections in each of the six traffic
across the 5.5 in. surface thickness, and if there lanes or 66 test sections in all. More replica-
is no interaction between load and surface tion might be required if the illustration were
thickness effects, the load effect at 5.5-in, sur- an actual road test, as the number of replicates
face could serve as the general effect of axle should be sufficient to obtain reliable estimates
load on performance. of residual variation (within tangents).
Replication of observations for controlled The third principle, randomization, is closely
factor combinations provides a way to find how associated with the principles of balance and
TABLE 2-G
LEVELS FOR EXPERIMENTAL FACTORS IN ILLUSTRATION
replication. Balance is necessary to prevent For the AASHO Road Test, tables that cor-
confusion among controlled factor effects, but respond to Table 3-G involve five pairs of p and
it is also important that there be no confusion log W values for each of 284 flexible pavement
between controlled effects and residual effects sections in five tangents, and for each of 264
on performance. If for example, the sections rigid pavement sections in five other tangents.
in each tangent were constructed so that sur-
face thickness increased from thin to thick PROCEDURES FOR ANALYSIS
along the tangent and if an uncontrolled con-
struction variable that could affect pavement The analysis consists of procedures that pro-
performance (for example, humidity) also in- duce an empirical formula wherein perform-
creased as the tangent was paved from one end ance is associated with load and pavement de-
to the other, any conclusion about surface thick- sign variables. In order to use mathematical
ness effect would be confused to an unknown procedures it is necessary to assume some
degree with effects attributable to humidity algebraic form, or model, for the association.
during paving. It is well known that system- In addition to the experimental variables the
atic uncontrolled variables operate during model involves constants whose values are
almost any experimental investigation, so either to be specified or to be estimated from
randomization is necessary to minimize the the data. Thus, the analytical procedures are
risk that residual effects will be mistaken for for the estimation of constants whose values
controlled effects. As in any sampling situa- are unspecified in the model—constants that
tion, randomization is also necessary for ob- indicate the effects of design and load variables
taining proper estimates of residual variation. upon performance. The procedures also in-
For example, if each replicate were constructed clude methods for estimating the precision
adjacent to its companion section, it might be with which the data fit the assumed model.
expected that an underestimate of residual In essence the model is an equation for
variation in the tangent would be obtained. serviceability trends as illustrated by Figures
The eleven sections in each tangent are 5-G or 6-G. When pavement design and axle
assigned a random order of occurrence within load are specified for a pavement section, the
the tangent. As a result, conclusions about equation presumably predicts the section's
surface and subbase effects are not biased or serviceability after a given number of applica-
confused by the presence of systematic resid- tions. The equation should also be useful for
ual variation within any tangent. predicting the number of applications the sec-
The major performance studies in the tion will experience before reaching a pre-
AASHO Road Test have experimental designs assigned serviceability level.
that involve balance, replication, and random- There are many different mathematical
ization, in much the same way that has been forms that could be used as models for service-
described for the illustration. In addition, still ability trends, many of which may fit the data
other experimental designs appear in the Road with more or less the same precision. Only one
Test to provide for special studies whose ob- of the numerous models that have been investi-
jectives are somewhat different from the first gated at the Road Test will be used for illus-
Road Test objective. tration.
Specifications have been given for pavement If p denotes the initial serviceability trend
performance data and for experimental de- value for a particular test section, p, —p is the
signs within which the performance data are serviceability loss experienced by the section
obtained. Many alternatives were available for when its trend value is at p. It is assumed
nearly every specification; nevertheless, the net that Pb does not depend upon pavement design
result of the selected specifications for the variables in the range of interest, and that the
example is a set of performance data as given best estimate for Po is c0, the average of all
in Table 3-G. The performance data consist initial trend values for sections considered in
of three pairs of p and log W values for each the analysis. For the example, c0 will be 4.5.
of the 66 test sections when the weighting func- The assumption used for the nature of serv-
tion of Figure 3-G is used. Table 3-G includes iceability trends is that serviceability loss is a
data previously given in Table 2-G for sections power function of axle load applications,
3212 and 3222. Section 3212 appears in tan- co —p=KWfi (1-G)
gent 3, lane 2, at the first surface thickness
level and the second subbase thickness level. in which p is a positive power and K and/or
Similarly, section 3222 is in the second lane p may depend on load and design variables.
of the third tangent and has the second level If p = c1 is a serviceability level such that
of thickness for both surface and subbase. Thus whenever p for a section falls to c1 the section
section numbers are codes for factor levels. Any is "out of test" and no longer observed, the
section whose serviceability history was com- number of applications experienced by the sec-
plete has p values of 3.5, 2.5, and 1.5. All tion when p = c1 is called the experimental life
remaining sections had incomplete histories. of the section. For the example, as at the Road
314 THE AASHO ROAD TEST, REPORT 5
Test, c1 = 1.5. If p is the value of W in Eq. 1-G The left side of Eq. 5-G is called G; that is,
when p = c1 , c0 - c1 = Kp13, or K = (c0 - (Co - P
c1 ) / p, and -Eq. 1-G may be written either G = log (6-G)
Co - C1 /
;
C0 4.5 2.5
=0.5
o0 3.0
4.0 = 10,000,000 10,000,000
3.5
3.5
3.0 4.0 p
p
2.5 4.2
2.0 2.0 4.3
e =100,000
C1 =I.5
4.4
1 2 3 4 5 6 7 8 9 10 4.0 4.4 4.8 O.Z Mb 0.0 0 0.8 (.
I I I I 0
/
I
/
1.97
-.4
I log 5.61
1.09
-.8
log 6.44
Section 3212 -1.2
Section 3222
, OG
Section 1133
/ Section 1233
, -.4
log 2 7.41
/ log 7.16
-1.2
4.8
0. 046-
5.2 5.6 6.0 6.4 6.8
-
• 4.8 5.2 5.6 6.0 6.4 6.8
I
-1.6
10gw logW
Figure 9-G. Illustrative estimates for f and log p from section data.
and 4.3; therefore, little decrease in service- = nominal load axle weight, in kips
L1
ability has been observed during the experi- (i.e., load values as given in Table
ment. It is assumed that very little informa- 3-G);
tion about /3 and log p is given by data and L2 = 1 for single axle vehicles, 2 for tan-
graphs for sections whose serviceability loss is dem axle vehicles;
outside the realm .of measurement error. Spe- = the first pavement design factor, slab
cial rules are applied in order to obtain values thickness, in in.; and
of /9 and logp for such sections. After ex- = the second pavement design factor,
amining all # values for sections that experi- subbase thickness, in in.
enced an appreciable serviceability loss, a mini- The remaining symbols on the right side of
mum value is assumed for /3, and if the data Eq. 8-G are positive constants whose values
for any section give /3 to be less than the are either to be assumed (as is done for /3) or
assumed minimum, then the minimum /9 is estimated from the /3 - values. In general, Eq.
assigned to be the section's /3. For the example, 8-G implies that /9 will increase as axle load
minimum /3 is taken to be 1.0, and both sections increases and that /3 decreases as pavement
1133 and 1233 are given the value /3 = 1.0
After this assignment, log p is obtained by
fitting a line whose slope is 1.0 to the observed
design increases for a fixed loading. If there
were three pavement design factors, as at the /
Road Test, then the third factor, D3, would
points. Using this rule, log is 7.41 and 7.16, have been introduced in the combination
respectively, for sections 1133 and 1233 (Fig. a,D1 + a2D2 + a3D3 + a4. The constant term in
9-G). If all p-values for a section are equal to the design combination (a3 in Eq. 8-G) ap-
or greater than c0 = 4.5, as for section 1132 pears so that /3 is not necessarily infinite when
(Table 3-G), /9 is assumed to be 1.0, and log there is no thickness for D, and D2, and L2 has
is set at the median log for all sections that been added to L, so that /3 does not necessarily
differ only in subbase thickness from the sec- approach /3c, as L, approaches zero.
tion that has no G data. Table 4-G gives /3 and For the example, graphs and variance anal-
log values. ysis for /3 show little or no dependency of /3
After /3 has been determined for each sec- upon subbase thickness, so a2 is taken to be
tion, the /9-values are graphed against pave- zero. With only one variable, D1 , in the design
ment design and load variables, and an analysis combination, the effect of D1 can be relegated to
of variance is made to infer the nature and the exponent B1 by assigning values to b1 and
extent of any dependence of /9 upon design and b 3. For the illustration, a1 = a3 = 1.0. Because
load variables. Neither the graphs nor the /30 has already been assumed to be 1.0, Eq. 8-G
analysis of variance are given herein, but both is reduced to
proceed from the assumption that /9 is related B0 (L,,+ L2 )B2,
to design and load variables according to the /3 = 1.0 + (9-G)
model (D1 + 1)L2'1'
BO (L1 + L2)B2 in which only B0 B1 B2 and B3 remain to be
/3=f9o + (8-G) estimated from the /3 data. Logarithms of both
(a1D1 + a2.!)2 + a3 ) 8' L2 B, sides of Eq. 9-G give
in which
log (/3- 1.0) = log B0 + B. log (L1 + L 2 ) -
= a minimum value for [3; B3 log L2 - B1 log (D1 + 1) (10-G)
316 THE AASHO ROAD TEST, REPORT 5
For each lane, Eq. 10-G represents a straight The second phase of the analytical proce-
line when log (/3- 1.0) is graphed versus dures begins by using /3 values to obtain new
log (D1 + 1), and linear regressions of log estimates for log p from the data for each sec-
(/3.- 1.0) on log (D1 + 1) give lane estimates tion. The first estimates for log p were denoted
for the slopes, B1. Omitting lane 1 of tangent by log p and were obtained as log W intercepts
1, since the majority of /3 values in this lane (Fig. 9-G) for lines whose slopes were
were 1.0 by assignment rather than values ob- Using the same rules as for obtaining log ,
tained from performance data, the regression the new estimates, log, are obtained as log W
slopes are averaged to give 91 , the final esti- intercepts for lines whose slopes are /9. Table
mate for B1.The average slope for the remain- 4-G gives values for log' for each section in
ing lanes is B1 = 5.90. the illustration. In essence, the rationale as-
If B1 log (D1 + 1) is transposed to the left sumes that estimates for /3 from Eq. 12-G are
of Eq. 10-G, better than estimates based only on individual
section performance data, and therefore, log 'p-
log (/3 - 1.0) + B1 log (D1 + 1) = values represent better estimates for log p than
logB0 + B2 log (L1 + L 2 ) - B3 log L2 do the log p-values.
(11-G)
Log p-values are graphed against the design
in whichhe left sideof Eq. 11-G is estimated and load variables, and an analysis of variance
by log (/3 - 1.0) + B1 log (D1 + 1) for every is made to infer how and with what signifi-
section. For ny lane, average log (/3 - 1.0) cance the log '_values depend on design and
+ average B1 log (D1 + 1) is called an ad- load variables. The algebraic form for the a-
justed lane mean, and according to Eq. 11-G sociation of log p with design and load variables
the adjusted lane means depend linearly upon is assumed to be
log (L1 + L 2 ) and log L2. Figure 10-G shows
the six adjusted lane means for the example, 0 (a1D3 + a2 D2 + a3) A L31
p= A (13-G)
and includes lines that are obtained from a (L1 + L 2 )
linear regression analysis. The common slopef
single and tandem axle lines is an estimate, B2 , in which A0, A1, A2, and A 3 are positive con-
for B2. The intercept of the single axle line on stants and a1 , a2 and a3 are the same constants
the adjusted means axis is an estimate, log B0 , that appear in Eq. 8-G. Eq. 13-G implies that
for log 80, and the difference between inter- p increases with pavement design and decreases
cepts of the single axle and tandem axle lines with axle load. The constant a3 is included so
produces an estimate, B3, for B3. For the illus- that p is not necessarily zero in the absence of
trative data, log B0 = -0.66 or 90 = 0.22, surface and subbase, and is added to L3 in the
92 = 4.54, and 9, = 3.12. Substitution of these denominator so that p is not necessarily infinite
values in Eq. 9-G gives a new estimation when L1 is zero.
formula for /3, For the illustrative data, Table 4-G indicates
0.22 (L1 + L 3 ) 454 little or no association between log' and sub-
/3 = 1.0 + (12-G base thickness, and so a2 is taken to be zero
(D1 + 1) 5 .90 L2312 and both a1 and a3 are set at 1.0 as was done
For each section values for (estimates for in Eq. 8-G. In logarithmic form, Eq. 13-G
/3 from Eq. 12-G) are given in Table 4-G. therefore becomes
6.0
5.5 4K
4S
. ,3ØK'
5.0
4.5
4.0
U -.66+4.54 log (L1+L2 )
3.5 8KS -3.12 log L 2
4KS
3.0 MITT ED)
Estimates for log p given by Eq. 15-G are Eqs- 17-G and 18-G therefore represent the
shown for each section in Table 4-G. first goal of the analysis-to associate the 'per-
The results of the analysis can now be sum- formance data with design and load variables.
marized. If it is desired to estimate p when W In the example given there was no need to
is given, then Eqs. 3-G, 12-G and 15-G or discuss derivations for the pavement design
16-G combine to give coefficients, a1, a2, . . . since only one design
factor had significant effect on performance.
/ W \'3
= 4.5 - 3.0 1 -- ) ( 17-G) If, as was the case for flexible pavement ex-
\PI periments at the Road Test, more than one
3.0
4KS
2.2 (OMITTED)\
1.8
-
1.4
1.0
.6
'
\3OK T
.2
36KT
-.2
24K5\\
I I
-.6
1.0 1.2 1.4 1.6 1.8
log (L 1 +L 2 )
Figure 12-G. Adjusted mean log vs log (L, + L2).
TABLE 4-G
ESTIMATES OF 19 AND LOG p
4KS 2.5 3 0.54 1.39 1.20 6.51 6.63 6.60 163(5 4.0 3 0.70 4.48 7.30 5.31 5.29 5.31 24KS 5.5 3 0.81 9.58 8.72 5.33 5.33 5.35
6 0.54 1.66 1.20 6.52 6.74 6.60 6 0.70 38.83 7.30 5.31 5.34 5.31 6 081 9.58 8.72 5.35 5.35 5.35
6 rep. 0.54 2.51 1.20 6.40 6.91 6.60 6 rep. 0.70 3.16 7.30 5.40 5.36 5.31 6 rep. 0.81 2.66 8.72 5.38 5.32 5.35
9 0.54 1.39 1.20 6.64 6.78 6.60 9 0.70 15.78 7.30 5.31 5.33 5.31 9 0.81 11.92 8.72 5.33 5.34 5.35
4.0 3 0.70 1.25 1.02 6.99 7.22 7.65 5.5 3 0.81 3.05 2.34 6.06 6.11 6.09 7.0 3 0.90 2.66 3.27 6.00 5.98 5.96
6 0.70 1.00 1.02 7.05 7.03 7.65 6 0.81 2.08 2.34 6.24 6.19 6.09 6 0.90 1.43 3.27 6.01 5.93 5.96
6 rep. 0.70 1.00 1.02 7.53 7.49 7.65 6 rep. 0.81 2.91 2.34 6.11 6.16 6.09 6 rep. 0.90 1.00 3.27 6.08 5.87 5.96
9 0.70 1.00 1.02 7.16 7.12 7.65 9 0.81 4:36 2.34 5.86 5.90 6.09 9 0.90 3.23 3.27 6.03 6.03 5.96
5.5 3 0.81 1.00 1.00 7.41 7.40 8.42 7.0 3 0.90 2.91 1.39 6.32 6.74 6.70 8.5 3 0.98 1.99 1.82 6.40 6.44 6.47
6 0.81 1.00 1.00 7.41 7.40' 8.42 6 0.90 1.05 1.39 6.90 6.62 6.70 6 0.98 1.48 1.82 6.48 6.35 6.47
9 0.81 1.00 1.00 7.41 7.40 8.42 9 0.90 1.00 1.39 6.93 6.65 6.70 9 0.98 2.51 1.82 6.40 6.58 6.47
8KS 2.5 3 0.54 9.73 3.88 5.42 5.45 5.48 30KT 4.0 3 0.70 11.93 13.73 5.02 5.02 5.06 36KT 5.5 3 0.81 23.93 6.91 5.49 5.51 5.50
6 0.54 1.64 3.88 5.48 5.40 5.48 6 0.70 23.85 13.73 5.04 5.05 5.06 6 0.81 1.97 6.91 5.61 5.53 5.50
6 rep. 0.54 2.52 3.88 5.36 5.33 5.48 6 rep. 0.70 6.68 13.73 5.15 5.13 5.06 6 rep. 0.81 5.97 6.91 5.58 5.58 5.50
9 0.54 1.79 3.88 5.62 5.56 5.48 9 0.70 5.74 13.73 5.04 5.01 5.06 9 0.81 15.79 6.91 5.47 5.49 5.50
4.0 3 0.70 1.25 1.35 6.61 6.56 6.53 5.5 3 0.81 2.39 3.71 5.89 5.86 5.83 7.0 3 0.90 15.83 2.74 6.03 6.10 6.12
6 0.70 1.15 1.35 6.86 6.70 6.53 6 0.81 11.94 3.71 5.82 5.86 5.83 6 0.90 1.09 2.74 6.44 6.04 6.12
6 rep. 0.70 2.27 1.35 6.33 6.60 6.53 6 rep. 0.81 1.19 3.71 6.05 5.87 5.83 6 rep. 0.90 1.00 2.74 6.21 5.94 6.12
9 0.70 1.99 1.35 6.40 6.62 6.53 9 0.81 3.68 3.71 5.85 5.85 5.83 9 0.90 5.90 2.74 6.06 6.21 6.12
5.5 3 0.81 1.00 1.08 7.26 7.16 7.30 7.0 3 0.90 1.78 1.80 6.37 6.37 6.44 8.5 3 0.98 1.25 1.63 6.75 6.56 6.62
6 0.81 1.00 1.08 7.53 7.42 7.30 6 0.90 1.99 1.80 6.40 6.45 6.44 6 0.98 1.25 1.63 6.99 6.74 6.62
9 0.81 1.00 1.08 7.16 7.06 7.30 9 0.90 1.99 1.80 6.40 6.45 6.44 9 0.98 1.99 1.63 6.40 6.50 6.62
TABLE 5-G
ESTIMATED PERFORMANCE DATA (ILLUSTRATION) FOR WEIGHTED APPLICATIONS
2.5-In. Slab 4.0-In. Slab 5.5-In. Slab 4.0-In. Slab 5.5-In. Slab 7.0-In. Slab 5.5-In. Slab 7.0-In. Slab 8.0-In. Slab
Load Subbase Subbase Load Subbase --
p (logW) p (log'W) Load Thickness (log"W) p (logW) p (lojW) Thickness p (log'W) p (log W) p (log"W)
(kips) Thickness P (log"W)
or (kips) T (in.)
p
or or or (ki (in.) or or or
(in.) or or
() log W (p) log 14' ())) log W () log ii' ) log W () log W () log W () log W (3) log W
4KS 3 (4.5) 5.48 (4.5) 5.48 (4.5) 5.48 16KS 3 3.5 (5.25) (4.5) 5.48 (4.5) 5.48 24KS 3 3.5 (5.29) 3.5 (5.82) (4.5) 5.48
(4.2) 5.81 (4.5) 5.81 (4.5) 5.81 2.5 (5.29) (4.1) 5.81 (4.4) 5.81 2.5 (5.33) 2.5 (5.91) (4.4) 5.81
(3.8) 6.05 (4.4) 6.05 (4.5) 6.05 1.5 (5.31) (2.1) 6.05 (4.1) 6.05 1.5 (5.35) 1.5 (5.96) 4.0) 6.05
6 (4.5) 5.48 (4.5) 5.48 (4.5) 5.48 6 3.5 (5.25) (4.5) 5.48 (4.5) 5.48 6 3.5 (5.29) 3.5 (5.82) 4.5) 5.48
(4.2) 5.81 (4.5) 5.81 (4.5) 5.81 2.5 (529) (4.1) 5.81 (4.4) 5.81 2.5 (5.33) 2.5 (5.91) 4.4) 5.81
(3.8) 6.05 (4.4) 6.05 (4.5) 6.05 1.5 (5.31) (2.1) 6.05 (4.1) 6.05 1.5 (5.35) 1.5 (5.96) (4.0) 6.05
6 rep. (4.5) 5.48 (4.5) 5.48 6 rep. 3.5 (5.25) (4.5) 5.48 6 rep. 3.5 (5.29) (4.5) 5.48
(4.2) 5.81 (4.5) 5.81 2.5 (5.29) (4.1) 5.81, 2.5 (5.33) (4.1) 5.81
(3.8) 6.05 (4.4) 6.05 1.5 (5.31) (2.1) 6.05 1.5 (5.35) Neg. 6.05
9 (4.5) 5.48 (4.5) 5.48 (4.5) 5.48 9 3.5 (5.25) 3.5 (5.88) (4.5) 5.48 9 3.5 (5.29) 3.5 (5.82) (4.5) 5.48
(4.2) 5.81 (4.5) 5.81 (4.5) 5.81 2.5 (5.29) 2.5 (6.01) (4.4) 5.81 2.5 (5.33) 2.5 (5.91) (4.4) 5.81
(3.8) 6.05 (4.4) 6.05 (4.5) 6.05 1.5 (5.31) 1.5 (6.09) (4.1) 6.05 1.5 (5.35) 1.5 (5.96) (4.0) 6.05
8KS 3 3.5 (5.35) (4.5) 5.48 (4.5) 5.48 30KT 3 3.5 (5.02) 3.5 (5.70) (4.5) 5.48 36KT 3 3.5 (5.43) 3.5 (5.94) (4.5) 5.48
2.5 (5.43) (4.3) 5.81 (4.4) 5.81 2.5 (5.05) 2.5 (5.78) (4.4) 5.81 2.5 (5.48) 2.5 (6.05) (4.4) 5.81
1.5 (5.48) (3.8) 6.05 (4.4) 6.05 1.5 (5.06) 1.5 (5.83) (3.9) 6.05 1.5 (5.50) 1.5 (6.12) (4.2) 6.05
6 3.5 (5.35) (4.5) 5.48 (4.5) 5.48 6 3.5 (5.02) 3.5 (5.70) (4.5) 5.48 6 3.5 (5.43) (4.5) 5.48 (4.5) 5.48
2.5 (5.43) (4.3) 5.81 (4.4) 5.81 2.5 (5.05) 2.5 (5.78) (4.4) 5.81 2.5 (5.48) (4.3) 5.81 (4.4) 5.81
1.5 (5.48) (3.8) 6.05 (4.4) 6.05 1.5 (5.06) 1.5 (5.83) (3.9) 6.05 1.5 (5.50) (2.6) 6.05 (4.2) 6.05
6 rep. 3.5 (5.35) (4.5) 5.48 6 rep. 3.5 (5.02) (4.5) 5.48 6 rep. 3.5 (5.43) (4.5) 5.48
2.5 (5.43) (4.3) 5.81 2.5 (5.05) (3.5) 5.81 2.5 (5.48) (4.3) 5.81
1.5 (5.48) (3.8) 6.05 1.5 (5.06) Neg.' 6.05 1.5 (5.50) (2.6) 6.05
9 3.5 (5.35) (4.5) 5.48 (4.5) 5.48 9 3.5 (5.02) 3.5 (5.70) (4.5) 5.48 9 3.5 (5.43) (4.5) 5.48 (4.5) 5.48
2.5 (5.43) (4.3) 5.81 (4.4) 5.81 2.5 (5.05) 2.5 (5.78) (4.4) 5.81 2.5 (5.48) (4.3) 5.81 (4.4) 5.81
1.5 (5.48) (3.8) 6.05 (4.4) 6.05 1.5 (5.06) 1.5 (5.83) (3.9) 6.05 1.5 (5.50) (2.6) 6.05 (4.2) 6.05
Numbers in parentheses for "out of test" sections are estimates of log W when p is given. Numbers in parentheses for "in test" sections are estimates of p when log W is given
'Negative estimate.
APPENDIX G
319
10
d8
A
10 4 10 5 106 1o7
Applications, W
TABLE 6-G
SUMMARY OF RESIDUALS
Residual
No. Sect. 9 6 5 3 3 5
Average t == 15
35
0.00
0.01
0.01
0.01
0.02
0.03
-0.01
0.09
0.02
0.08
-0.03
0.07
0.00
0.04
( - t = 55 0.06 -0.05 -0.08 0.09 0.05 0.03 0.01
All 0.02 -0.01 -0.01 0.06 0.05 0.03 0.02
In
• No. Sect. 0 3 4 6 6 4
Average p == 2.5
3.5 -
- 0.05
-0.03
0.02
0.04
0.00
0.01
0.01
0.00
-0.01
0.00
0.01
0.01
Cd
-
(logW log W) p = 1.5 - -0.01 0.06 0.00 -0.03 0.01 0.01
w All - 0.01 0.04 0.00 0.00 0.00 0.01
Average p == 3.5 -
- 0.06
0.05
0.07
0.06
0.04
0.02
0.04
0.03
0.04
0.03
0.05
0.04
2.5
-
I log'W log W I p = 1.5 - 0.08 0.06 0.03 0.04 0.05 0.05
All - 0.06 0.06 0.03 0.03 0.04 0.04
APPENDIX G 321
case that p = 2.5). For a particular load, the include about 90 percent of all individual resi-
plotted curve shows the number of applications duals. If the central curves of Figure 14-G are
expected for any surface thickness at the time used to estimate the slab thickness required to
when the serviseabi1ity level has dropped to keep serviceability above 2.5 for a specified
2.5. Applications beyond 106 represent extrap- number of applications, W, the two mean resi-
olations for all curves, and each load curve has dual bands correspond to approximately ± 4
been extended beyond the lower and upper sur- percent of the slab thickness estimated from
face thickness level used in the experiment. the performance equation curve. Thus the error
Corresponding curves, of course, could be bands for an estimated 6-in, slab requirement
graphed for other values of p. would represent ± 1/4 in.
To show how close the observations are to Another form of Figure 13-G is shown in
the curves of Figure 13-G, Figure 14-G re- Figure 15-G which shows the associated sur-
peats three of the curves from Figure 13-G face thickness and load for selected applica-
and includes appropriate data that were given tions expected when p = 2.5. Figure 15-G
as observations in Table 3-G. Data for only brings out the effect of load upon surface thick-
those sections whose p reached 2.5 can be ness requirements.
shown in Figure 14-G. The dotted curves repre- When bands for residual variation are added
sent the band formed by ±0.10; that is, plus to the graphs, either or both of Figures 13-G
or minus twice the mean absolute residual in and 15-G constitute a summarization of the
log W. If residual deviations have a normal data given in Table 3-G, and presumably satisfy
frequency distribution about the curves, plus (for p = 2.5) the objective for the investiga-
or minus two mean absolute deviations should tion.
observed extrapolated
4
region region 4
..
•
••
EE
10 1O5 1 106
I,
12
axle appl.
(extrapolate d)
I I
WE
4 8 .12 16 20 24 28
Single Axle Load (kips)
I I I
0 8 16 24 32 40 48
Tandem Axle Load (kips)
I 10 0
SPRING, 1960 VALUES
SUMMER, 1960 VALUES
0 0
8 6
S
0 0 00
0 0
0 S 0
0 o 0 0
• • S
0- 0
0 0
S •
o. 0OS
S
SPRING, 1960 VALUES
SUMMER, 1960 VALUES
0
104 108 112 116 12 16 20
DRY DENSITY PCF MOISTURE CONTENT, %
S
0
o •
w 0
0
-J
4 0
> 0
.S
CO 4 o
C.) 0
S
SPRING, 1960 VALUES •0S S
SUMMER, 1960 VALUES , o
•
0
60 70 80 90
PERCENT SATURATION
Figure 1—H. Embankment soil test data from trenches, main loop sections.
323
". 250 250
z 0
Cr
VALUES LU VALUES
a-
SPRING 1959 S SPRING 1959
m
SUMMER 1959 0 SUMMER 1959
-j 200 - 0 200 0
U
0
U)
U) 0 0
0 -J 0
0 0
150 S 150
0. ________________
.8 o co
S
o •• S •. 0
- S • S
S .
Li
106 110 114 118 8 12 16 20
DRY DENSITY P C F MOISTURE CONTENT, %
z 0
Li - VALUES
0
SPRING 1960
co — 0 SUMMER 1960
-i 200
U
5 0
U) 0
0
-J
• °
150
0
0
0
.:.
°
•
C')
4 -
_j Inn
Li60
70 80
PERCENT SATURATION
Figure 2—H. Embankment soil test data from trenches, main loop sections.
S
0
0
0
S
0 C-,
- S -
I
U) U)
4 4
H I nn IOO I I
108 112 116 10 14 18
DRY DENSITY PCF MOISTURE CONTENT, %
8
0'
LOOP AVERAGES 250
SPRING LOOP AVERAGES
cr
0 SUMMER LU SPRING
a-
0 SUMMER
0
0
Ia
0
U)
D S
-J 0
S • 150
0
o
S
S
70 80 90 70 75 80 85
PERCENT SATURATION PERCENT SATURATION
Figure 3—H. Embankment soil test data from trenches, main loop sections, loop
averages.
8 8
0 0
0
0 0 0
0 0 0 0
° 0
00
o8
00
0000
0C 0 0 010 0 01 000
0
#01 00 0 0 8,400948 00
0 00 80o 00 8 e
01
00 80; o 0 01 0
00 O 0 0000
e %Oo 0
00000 0
00
0-
104 108 112 116 10 14 18 22
DRY DENSITY P C F MOISTURE CONTENT, %
Si
—J 0
4
CPO 0
> 4 0 0
0 01 00 0 410
00o
Co
o 80 0 0 o 0
1)
0 0
0 0
00
0 e0:g% 0 0
0 0
0 I
65 75 85
PERCENT SATURATION
IRE IRE
0
0
:000o
8 0 0
o
o 00
0 P*
0 0
04:€ o
0 0 0 : %
O l N ,Ot0 000
9,00 0 0
0
0o 0: 0
80 00
OOQ % 90
00
0
000 ojo 0
0
0 :00o0oJo0
0
010
NC
60 70 80 90
PERCENT SATURATION
This appendix lists all Road Test data sys- "printout" are given. A folder includes such
tems except those connected with bridge re- material as field data sheets, tabular and
search. Bridge data systems are listed and graphical summaries, partial or complete
described in Report 4. analysis, material, from related studies, and
A 4-digit number and a code letter, indicat- maps. The one or more folders associated with
ing which Road Test branch was mainly re- a data system are stored in branch files. Be-
sponsible, identify each data system. Branch cause of the wide variety of material contained
code letters are as. follows: in these folders and the impracticality of de-
Code Branch scribing this material, requests therefor should
be preceded either by personal study or by de-
B Bridge tailed correspondence with the Highway Re-
D Data Analysis search Board. When it is determined what
F Flexible Pavement folder material is of interest, appropriate
reproduction can be made to satisfy the request.
I Instrumentation In certain data systems, basic data appear
L Materials Laboratory as analog oscillograph chart records (tapes).
M Maintenance Requests for tapes or tape reproduction should
also be preceded by personal study or detailed
0 Operations correspondence.
R Rigid Pavement Much of the information in data systems has
S Special Assignments been put on IBM cards. Either cards and/or
printouts with a' descriptive cover sheet are
The data system title, given after the identifi- available from the Highway Research Board.
cation code, is followed by a short descriptive Requests should be addressed to the Director,
paragraph except where the title is considered Highway Research Board, 2101 Constitution
self-explanatory. Avenue, N. W., Washington 25, D. C.
At the end of the descriptive paragraph one Prices will be based on the cost of repro-
or more of the words "folder," "tape," or duction.
1001—D Data Systems Card File 1310—L Background for Selection of Pavement
File of 5- x 8-in, cards, one card for Designs
each data system. Card includes Historical material on design of
identification, brief description, rigid and flexible pavement sections,
items and quantities, and specific 1951-1958. (Folder)
location of items.
1312—L Preliminary Information on Embank-
1020—D IBM Card Layout File ment Soil, 1956 (Folder)
Punched card layout, column by
column, for each data system hav- 1313—L Miscellaneous Notes on Embankment
ing punched card data. Layout in- Construction (Folder)
cludes explanatory notes, diagrams, 1316—L Locations of Test Holes in Embank-
etc., when needed. (Folder) ment, Spring 1957 (Folder)
1110—F Purdue Base Materials Study 1317—L Miscellaneous Moisture, Density, and
A Purdue University thesis that re- Field CBR Tests of Embankment
ports the results of a laboratoryin- Under Mulch Cover, July 1957
vestigation of the behavior of (Folder)
gravel and crushed stone mixtures
subjected to repeated loading. Ma- 1318—L Moisture Profiles of Embankment
terials were obtained on the site of Taken Immediately Before Placing
the AASHO Road Test. (Folder) Mulch, 1957 (Folder)
326
APPENDIX I 327
1319—L Effect of Thickness of Mulch Cover on 1342—L Daily Reports, Concrete Pavement
Embankment Moisture Content, Fall Construction (Folder)
1957 (Folder) 1343—b Aggregate Inspection Reports, Con-
1321—b Miscellaneous Tests for Moisture and crete Pavement Construction (Folder)
Density of Embankment, Spring 1958 1344—b Miscellaneous Inspection Reports
(Folder) Darex, water, steel. (Folder)
1322—b Moisture Profiles of Embankment, 1345—b Concrete Control for Bridge 'Construc-
Spring 1958 (Folder) tion
1323—b Moisture Tests on Shoulders of Rigid Miscellaneous information. (Folder)
Pavements, August 1958 (Folder)
1346—b Field Test Results on Bituminous Con-
1324—b Summary and Study of Moisture Con- crete Used on Shoulders (Folder)
tent of Embankment Through Begin-
fling of Traffic, 1958 (Folder) 1347—b Extraction Test on Material Removed
from Pavement, Bituminous Concrete
1325—b Information from' 1956 Production of Construction
Mulch (Folder) Tests on surface, course material re-
1326—b Inter-Laboratory Tests, 1956 Produc- placed at beginning of construction,
tion of Mulch (Folder) boop 3. (Folder)
1327—L Subbase Compaction Data, Shoulders, 1348—L Aggregate Inspection Reports, Bitu-
(Folder) minous Concrete Construction
1328—b Subbase Compaction Data, Turn- Tests on material at aggregate pro-
arounds (Folder) ducer plants. (Folder)
1329—b Summaries of Various Laboratory 1349—b Moisture Content of Concrete Pave-
Studies of Methods for Stabilizing ment Samples
Original Subbase Material (Folder) Tests on samples from boop 1 to
determine moisture content of con-
1330—b Summaries of Various Field Studies crete. (Folder)
of Methods for Stabilizing Original
Subbase Material (Folder) 1350—L Investigation of Surface 'Cracks in
Concrete Pavement
1331—b Pilot Section Work on Compaction of bimited investigation to determine
Subbase (Folder) cause of surface cracks in turn-
1332—L Nominal Compaction of Mulch around section of boop 5. (Folder)
(Folder) 1351—b A Correlation Between Plastic Limit
1333—L Tests on Pit-Run Gravel Used in Sub- and Optimum Moisture Content
base Material (Folder) An investigation using data from
1334—b Laboratory Tests on Original Unstabi- Data System 2330. (Folder)
lized Subbase (Folder) 1500-1515-0 Operations Branch Records
1335—b Investigation and Selection of Binder Data not generally available except
Soil Used to Stabilize Subbase through special release agreements.
(Folder) Includes such items as tire recap
1336—b Pilot Section Construction Test Data, inventories, vehicle failure reports,
Base Course (Folder) fuel and lubrication data, accident
data, driving schedules, loading dia-
1337—b Maximum Dry Density of Type A grams and calculations. (Folders)
Base Used on Shoulders (Folder)
1600—D AASHO Calendar Code
1338—b. Compaction Test Data for Type A Month-by-month calendar , that
Base Used on Shoulders (Folder) shows both calendar day and
1339—b Type A Base, Plant Gradations AASHO code date, where July 1,
(Folder) 1956 = code day' 1, July 2, 1956 =
day 2, etc. (Folder)
1340—b Miscellaneous Compaction Data from
Pilot Section and Turnarounds (Soil, 1800—I Instrumentation Branch Files
Base and Subbase) (Folder) Designs and graphs pertinent to
certain electronic instrumentation
134.1—b Report of Routine Modulus of Rupture at the AASHO Road Test. (Folder)
Tests, Concrete Pavement Construc-
tion 2000—b Construction Plans ,
Tests used to determine length of Blueprints for initial layouts of the
curing period. (Folder) AASHO Test Road. (Folder)
328 THE AASHO ROAD TEST, REPORT 5
2110—L Field CBR Tests, Flexible Pavements 2129—L CBR Tests, Stone Base, Flexible Pave-
Tests on embankment soil after ments
compaction before subgrading sub- Routine tests, construction period.
base. (Folder, 1-page printout) (Folder)
2111—L Moisture Content of Flexible Pave- 2130—L Thickness of Subbase, Flexible Pave-
ment, Embankment and Subbase ments
Tests made through core holes prior Thickness and moisture content
to test traffic. (Folder) after subgrading. (Folder, 18-page
printout)
2119—L Embankment Density, Flexible Pave- 2131—L Construction Test Data, Subbase,
ments Flexible Pavements
Tests made after subgrading. Moisture, density, and compaction
(Folder) data, quality control charts. (Fold-
2120—L Thickness of Base Course, Flexible ers, 20-page printout)
Pavements 2132—L Density of Subbase After Subgrading,
Tests made after subgrading. Flexible Pavements
(Folder, 19-page printout) Moisture, density, and compaction
data, after subgrading. (Folder, 9-
2121—L Gravel Base Construction Data, Flexi- page printout)
ble Pavement Base Type Studies
2140—L Marshall Test Results, Bituminous
Construction moisture and density Concrete Construction
tests, laboratory CBR tests, grada- Laboratory control tests, quality
tion, liquid and plastic limits.
(Folders) control charts. (Folders, 15-page
printout)
2122—L Cement-Treated Base Construction 2141—L Extrction Test Results, Bituminous
Data, Flexible Pavement Base Type Concrete Construction
Studies Laboratory test results, quality
Laboratory and field moisture and control charts. (Folder, 8-page
density tests, compressive strength printout)
tests, gradation tests, mixture de-
sign information. (Folder) 2142—L Field Density Tests Results, Bitumi-
nous Concrete Construction
2123—L Stone Base Construction Data, Flex- Laboratory tests on field density
ible Pavements samples. (Folders, 47-page print-
Field density and compaction test out)
data. (Folder, 25-page printout) 2143—L Hot-Bin Gradations, Bituminous Con-
2124—L Tests of Stone Base After Subgrading, crete Construction
Flexible Pavements Test data, quality control charts.
(Folder)
Field moisture and density tests,
thickness measurements. (Folder, 2144—L Illinois Highway Department Labora-
6-page printout) tory Test Results, Bituminous Con-
crete Construction
2125—L Stone Base Plant Samples, Flexible Illinois Highway Department test
Pavements reports on bituminous concrete. Ex-
Gradation and moisture content, traction, Marshall, high-pressure
plant inspector's daily reports. air meter tests. (Folder)
(Folder) 2145—L Illinois Highway Department Labora-
2126—L Stone Base Road Samples, Flexible tory Test Results on Asphalt, Bitumi-
Pavements nous Concrete Construction
Gradation. (Folder) Illinois Highway Department daily
report sheets. Tests at refinery and
2127—L Shoulder (Type A) Base Plant Sam- on Road Test. Penetration, spot
ples, Flexible Pavements test, etc. (Folder)
- Gradation, inspector's daily reports. 2146—L Construction Test Data, Bituminous
(Folder) Concrete Base, Flexible Pavement
Special Base Experiment
2128—L Shoulder (Type A) Base Road Sam- Mix design information. Results of
ples, Flexible Pavements extraction, Marshall and field
Gradation. (Folder) density tests. (Folder)
APPENDIX 1 329
2147—L Test Data, Special Stability Mixes on 2170—b Effect of Compaction Procedure on
Turnarounds, Bituminous Concrete Maximum Density of Base
Construction Prelithinary studies prior to con-
Extraction, Marshall, and field den- struction. Includes some field test
sity test results. Quality control data. (Folders)
charts. (Folders) 2210—b Field CBR Tests, Rigid Pavements
21487b Tests on Recovered Asphalt, Bitumi- Tests on embankment soil after
nous Concrete Construction compaction prior to subgrading
Illinois Highway Department Tests. subbase. (Folder, 1-page printout)
Penetration, ductility, etc. (Folder) 2211—b Moisture Content of Rigid Pavement,
2149—L Bituminous Concrete Mixture Designs, Embankment and Subbase
Flexible Pavements Tests through core holes before test
Mixture designs by Road Test traffic. (Folder)
Materials baboratory, Illinois High- 2219—b Embankment Density, Rigid Pave-
way Department, and The Asphalt ments
Institute. (Folder) Testsafter subgrading. (Folder)
2150—F Surfacing Thickness, Bituminous Con- 2220—b Thickness of Subbase, Concrete Pave-
crete Construction ments
Asphaltic concrete thicknesses Thickness and moisture content
measured as the pavements were after subgrading. (Folders, 16-page
laid. Data at six transverse points printout)
for 25-ft longitudinal stations.
(Folder, printout) 2221—b Subbase Construction Data, Concrete
Pavements
2151—L Tests on Asphalt by Outside Agencies Construction test data. Moisture,
Special tests on asphalt by Stand- density and compaction data. Qual-
ard Oil Company of Indiana, Stand- ity control charts. (Folders, 15-
ard Oil Company of Ohio. Complete page printout)
certified analysis by Standard Oil
Company of Indiana. (Folder) 2222—b Density of Subbase After Subgrading,
Concrete Pavements
2152—b Reports on Bituminous Concrete by
Other Agencies Moisture and density test data.
(Folder, 10-page printout)
Hubbard-Field tests by Florida.
Observations and reports of tests 2230—L Routine Construction Test Data, Con-
by Bureau of Public Roads. crete Pavement Construction
(Folder) 14-day strength tests, slump, air
2153—b Tests on Pilot Sections, Bituminous content, etc. Quality control charts.
Concrete Construction (Folders, 19-page printout)
Extraction, Marshall, and field den- 2231—b Static and Dynamic Properties of
sity test data. Quality control Laboratory Cured Concrete Beams
charts. (Folders) and Cylinders
2154—b Tests on Bituminous Concrete Shoul- Tests made on specimens molded
during construction of concrete
ders, Bituminous Concrete Construc- pavements. Moduli of elasticity,
tion Poisson's ratio, fiexural and corn-
Extraction, Marshall, and field pressive strengths. Summarized in
density tests. Quality control charts, field office report 26. (Folders, 1-
(Folders) page printout)
2155—b Illinois High-Pressure Air Meter Tests 2232—b Proportioning Plant Reports, Con-
on Bituminous Concrete crete Pavement Construction
Reports of tests by Illinois High- Daily reports of proportioning
way Department Laboratory. Air engineer and paving engineer. Tests
content of compacted density test by Illinois Task Force engineers.
samples. (Folder) (Folders)
2160—b Effect of Test Temperature on Mar- 2233—b Tests on Aggregate at Proportioning
shall Stability Plant, Concrete Pavement Construe-
Results of laboratory study. Sam- tion
pies of binder and surface course Daily report sheets of proportion-
mixtures tested over a temperature ing engineer. Gradation, specific
range from 401 to 160° F. (Folder) gravity. (Folder)
330 THE AASHO ROAD TEST, REPORT 5
2331—L Subbase Plant Samples, Construction 2345—L Analysis of Plant Production, Original
Inspector's daily report sheets. Sand-Gravel Mulch Placed 'on Turn-
Gradation tests. (Folder) arounds in 1956 (Folder)
2332—L Subbase Road Samples, Construction 2351—F Plate Bearing Tests and CBR's, 1956-
58
Laboratory reports. Gradation tests, Data sheets for plate bearing, CBR,
CBR, PT. Quality control charts. and moisture content tests on em-
(Folder) bankment soil. (Folders, printout)
2333—L Tests on Borrow Pit Soils by Outside 2520-0. Vehicle Axle Spacings and Axle
Agencies Weights
Reports by State of Michigan, Bu- Detailed dimensions and weights
reau of Public Roads, University of for the original 70 vehicles used in
Minnesota, The Asphalt Institute. test traffic. (3-page printout)
Routine tests, triaxial shear, X-ray
diffraction. (Folder) 2521-0 Axle Loads and Tire Pressures
2334—L Soil Core Samples from Centerline of Detailed loads and tire pressures
Roadway for the test vehicles. (3-page print-
out)
Report on original survey by Illi-
nois State Highway Department. 2522-0 Contact Areas and Tire Prints
Tests on core samples taken along Dimensions and prints for the con-
centerline of AASHO Road Test. tact area of test vehicle tires.
Routine classification test data. (Folder, 10-page printout)
(Folder)
2523-0 Axle Spacings and Axle Weights
2335—L Tests on Soil by University of Cali- Detailed weights and dimensions
fornia for all 127 test vehicles. (Folders,
"A Study of the Deformation Char- 5-page printout)
acteristics of the AASHO Road 2130—L Moisture Cells
Test Subgrade I Under Repeated
Loading," C. K. Chan, H. B. Seed. Miscellaneous information on mois-
Tests on AASHO Road Test soil ture .cells used on rigid pavements.
compared to silty clay from Vicks- (Folder)
burg, Miss., and clayey silt from 2140—S Frost Depth Data, Flexible Pavements
WASHO Road Test. (Folder) Description of instrumentation,
2336—L Tests on Soil, Subbase, and Base Ma- field layout, and results from frost
terial by the California Division of depth indicators. (Folders, 57-page
Highways printout)
Routine tests. Resilience test. 2141—F Description of Thermocouple Program
Samples obtained from roadway. Field layout of thermocouples used
(Folder) in connection with frost depth indi-
2337—L Mineral and Chemical Composition of cators and, for determining iso-
Road Test Aggregates thermal lines. (Folder)
Properties of parent materials. 2143—F Loop 1 Thermocouple Program
(Folder) Field layout and temperature print-
2340—L Subgrade Construction Data Analysis outs for thermocouples installed in
Procedures Loop 1 flexible pavement. (Folder)
Various analyses and other infor- 2147—F Thermocouple Program for Deflection
mation used during construction of Tests
earth embankment. Includes sta- Field layout and mat temperatures
tistical acceptance procedures. recorded in connection with routine
(Folder) flexible pavement deflections.
2341—L One-Point Procedure for Obtaining (Folder) •
Maximum Density 2148—F Thermocouple Program for Determin-
Various charts and graphs pertain- ing Isothermal Lines
ing to the development of maximum Field layout and 24-hr temperature
density-optimum moisture content data for thermocouples installed in
curves used during construction of Loop 1 flexible pavement to obtain
embankment. (Folder) temperature isothermals. (Folder)
332 THE AASHO ROAD TEST, REPORT 5
3200—R Rigid Pavement Adversity Factor 4102—F Initial Condition Survey of Cement-
Study Treated Base Sections
Initial efforts to establish rigid Maps of cement-treated base sec-
pavement adversity factor based on tions made just after construction
rainfall and frost. (Folder, 1-page and before surfacing. (Folder)
printout) 4111—F Correlation Between Slope Variance
3201—R Rigid Pavement Adversity Factor and Rut Depth
Study Slope variance plotted against rut
Further data and formulas for rigid depth by loops and lanes. Also rut
pavement adversity factor based on depth variance against rut depth.
frost. (Folder, 1-page printout) (Folder)
3240—S Rigid Pavement Frost Depth Data 4120—F Transverse Profilometer Data, Flexi-
ble Pavements
Field layout, procedures, and re- Routine data from the transverse
sults from frost depth indicators in- profilometer. (Folder, printouts)
stalled in rigid pavements. (Folder,
52-page printout) 4121—F Rut Depths from Transverse Profil-
ometer
3300—S Weather Station Data Rut depths computed from nine
Description of instruments, install- points of the transverse profilo-
ations, and climatological data ob- meter records. (Folder, 87-page
tained during the Road Test. (36- printout)
page printout)
4123—F Routine Rut Depths, Flexible Pave-
3301—S Loop G Weather Data ments
Daily maximum and minimum tem- All rut depth data obtained using
peratures at the top, middle and the manual gage. (Folders, 395-
bottom of rigid and flexible pave- page printout)
ments in specially constructed
sections near the Road Test admin- 4124—F Special Transverse Rut Depth Data,
istration building. Ground tempera- Flexible Pavements
ture 6 in. below pavement. (Folder, Special rut depth data obtained
36-page printout) with the transverse profilometer.
(Folders)
3303—S Pavement and Air Temperatures
Measured in Loop G 4126—F Rut Depth Studies, Flexible Pave-
ments
Special project to determine pave-
ment and air temperatures hourly Rut depths taken at 5-ft intervals
over the 2-year period in rigid and in all paved shoulder and special
base experiments. (Folder, 138-
flexible pavements. (Folder, 397- page printout)
page printout)
4127—F Rut Depth Analyses
3304—S Area Weather Summation Special analyses of rut depth for
Daily precipitation; maximum, factorial flexible pavement sections.
minimum, and mean air tempera- (Folder)
ture; and mean frost depth at the
Road Test and in neighboring areas. 4128—F Comparison of Rut Depth Measure-
(Folders, 30-page printout) ments
Correlations of rut depth by manual
3310—S Ground Water Table gage with those by transverse pro-
Daily ground water table at vari- filometer. (Folder)
ous installations in the teat road 4129—F Final Rut Depth Measurements, Flex-
tangents. (Folder, 53-page print- ible Pavements
out) Field data sheets for rut depths
4100—F Routine Condition Surveys, Flexible measured at end of traffic for all
Pavements sections still in test. Measurements
Field survey maps and summary taken at 5-ft intervals in each sec-
tion. (Folder)
data obtained from weekly surveys
during the traffic phase. Amount 4130—F Subsurface Layer Thickness Changes
and types of cracking and patching. Data from settlement rod measure-
(Folders, 1,480-page printout) ments. (Folder, 129-page printout)
APPENDIX I 333
4199—F Historical Records of Flexible Pave- 4251—L Effect of Weather and Sawing on Con-
ment Test Sections crete Beams
One chart for each flexible pave- Report of tests on beams from
ment section and subsection in the Loop 1. Strength and dynamic test
Road Test, showing serviceability properties. (Folders)
history, cracking and patching 4261—b Test in Failed Areas, Rigid Pavements
history, longitudinal profile history,
rut depth history, deflection history, Moisture and density tests. (Fold-
and history of any overlay studied ers, 7-page printout)
after failure of the original section. 4263—b Trench Studies in Failed Areas, Rigid
(Folders) Pavements
4200—R Routine Condition Surveys, Rigid Profile measurements, moisture and
Pavements density, gradation, CBR, and plate
Weekly measurements of cracking, load tests. (Folders)
faulting, patching, spalling, etc. 4290—R Prior-to-Maintenance Rigid Pavement
Field maps, special data in failed Surveys
areas. (Folders, 680-page printout) Roughometer, cracking and patch-
4201—R Section Summary of Condition Sur- ing data, serviceability values for
veys, Rigid Pavements each section. (Folders)
Summary by section and index day 4292—R History Plots for Rigid Pavement
of measurements taken in Data Test Sections
System 4200. (340-page printout) One chart for each rigid pavement
4202—R Condition Survey, Section and Panel section, in the Road Test, showing
Average serviceability, cracking and patch-
ing, cracking index, edge strains,
Same as Data System 4201, but in- static deflections, pumping index.
cludes cracking index. (Folder, 340- (Folders)
page printout)
4300—D Routine Longitudinal Profiles
4220—R Transverse Profiles with Profilometer Analog traces of wheelpath slope in
Transverse profile at nine points each wheelpath of each lane. Bi-
taken with transverse profilometer, weekly runs throughout period of
seven times on selected sections Road Test. (Tapes)
over a two-year period. (Folder)
4305—R Roughometer Data
4222—R Transverse Profile, Failed Rigid Pave- Roughometer data taken with me-
ment Sections chanical counter.. Data considered
Contours of subbase taken with rod impractical for the short Road Test
and level in failed sections. sections. (Folders)
(Folder)
4306—R Michigan Profilograph Data
4230—R. Consolidation of Rigid Pavement Sub- Comparison of Road Test profilo-
base meter data with Michigan profilo-
Measurements intended to detect graph data taken on Road Test sec-
consolidation of subbase near panel tions. (Folder)
corners. Faulty instrumentation 4307—R Roughness by Roughometer
yielded no practical data. (Folders,
126-page printout) Results of 15 determinations on all
factorial sections from June 1959
4240; 4241, 4242, 4243—R Rigid Pavement through November 1960. Rougho-
Pumping Surveys meter equipped with electronic
counter. (Folder, 308-page print-
Basic information and pumping out)
scores for each rigid pavement test
section after each rain. (Folders, 4308—R Roughometer Correlation Data
292-page printout) Correlation study made in 1959 be-
tween Road Test roughometer and
4250—b Routine Subsurface Studies, Loop 1, BPR-102 roughometer having elec-
Rigid Pavements tronic counter. (Folder)
Moisture and density tests, CBR,
and plate load test. (Folders, 6-page 4309—R Routine Roughometer Data
printout) Continuation of Data System 4307.
APPENDIX I 335
5226—R Analysis of Curl Data 5255—R Deflection Basin and Strain Influence
Coefficients for orthogonal poly- Study
nomials; smoothed data. (Folder) Special study to determine distance
5240—R Internal Temperature of Subgrade of vehicle from gage point at
and 9.5- and 12.5-Inch Slabs in Loop 1 threshold values of dynamic edge
Temperatures at several points in strain, dynamic corner deflections,
and under seven slabs over same static rebound edge deflections, and
period of time, usually 24 to 36 hr. static rebound corner deflections.
(Folder, 157-page printout) (Folder)
5241—R Internal Slab Temperatures of Sub- 5256—R Dynamic Strain and Deflection vs
grade and 2.5- and 5.0-Inch Slabs in Speed
Loop 1 Special study in Loops 4 and 6 at
Same as Data System 5240 except two different times to determine
for different slab thickness. (Folder, interrelationships of strains, deflec-
188-page printout) tions, load, speed, and temperature.
(Folders)
5242—R Curing Temperatures in Plastic Con-
crete, Loop 1 5257—R Special Studies of Dynamic Strains
Special studies of temperatures and Studies to determine the point at
beneath two 12.5-in, slabs in Loop 1 which maximum strain occurs for
during plastic and hardening phases reinforced and non-reinforced pave-
of freshly poured concrete. (Folder, ment, transverse and longitudinal
12-page printout) distribution of strain. (Folder)
5243—R Temperatures for 24-Hour Study of 5258—R Static and Dynamic Load-Strain,
Strains and Deflections Load-Deflection Studies
Temperatures in selected slabs in Studies to support the theory that
Loop 1 during special 24-hr studies strain or deflection is directly pro-
on traffic ioops. (8-page printout) portional to load when other factors
are constant. (Folder)
5250—R Routine Dynamic Strains and Deflec-
tions in Traffic Loops 5260—R Maximum Vertical Movements
Dynamic edge strains and corner A study to show the relationship
deflections under regular test traffic between pavement design and verti-
in routine dynamic measurements cal movements due to traffic.
program. (Folder, 92-page print- (Folder)
out) 5270—R Concrete and Soil Temperatures in
5251—R Effect of Vehicle Placement on Dy- LoopG
namic Strain and Deflections Special study to determine tempera-
Effect on dynamic strain and deflec- ture stress in top and bottom of a
tion of lateral placements on either concrete pavement specially built
side of desired load placement. near the Road Test administration
(Folder) building. (Folder)
5252—R Uniformity of Dynamic Strain and 5280—R Routine Static Rebound Deflections
Deflection in Loop 3 Static rebound corner deflections of
Uniformity of dynamic strains and PCC pavement under normal loop-
deflections in sections with constant load thickness combinations. Ben-
slab thickness and constant load. kelman beam probe at edge of pave-
(Folder) ments, one round every two weeks.
5253—R Dynamic Strains and Deflections vs (Folder, 291-page printout)
Temperature 5282—R Static Rebound Deflection-Placement
Results of three 24-hr studies to Studies
determine effect of daily variations Study to determine effect of load
in air temperature on strains and placement on deflection of PCC
deflections. (Folder, 30-page print- pavements. Probe at edge of pave-
out) ment. (Folder)
5254—R Routine Dynamic Strain and Deflec- 5283—R Uniformity of Static Rebound Deflec-
tion Studies tions
Same as Data System 5250 except Study to compare deflection of all
data are recorded on punched tape PCC pavements under a uniform
in van. (Folder) 12-kip single axle load on Loops 3,
APPENDIX I 339
3.0, 2.5, 2.0, and 1.5. (19-page print- ability index formulas used for
out) rigid and flexible pavement.
7244—R Performance Data for Rigid Pave- (Folder)
ments, Loop 1, Design 1 7620—D Special Index Study
Smoothed present serviceability on Correlation between different index
index days 11, 22, 33, 44, and 55 for formulas and different profile statis-
each wheelpath of Design 1 sections tics for large sample of test road
in Loop 1, lane 1. (1-page print- sections. (Folder, printout)
out) 7664-7674—D Development of Flexible Pave-
7246—R Rigid Pavement Performance Equa- ment Performance Equations
tion Estimates Graphs, tables, and regression
Performance •data estimates and analyses, and formulas for, various
residuals from rigid pavement per- mathematical models leading to
formance equation given in Report flexible pavement performance
5. (9-page printout) equations given in Report 5. (Fold-
7305-7309—D Present Serviceability Ratings ers)
of Selected Pavements 7679—F Comparison of WASHO and AASHO
Individual ratings of selected pave- Pavement Performance
ments by members of the AASHO A comparison of serviceability
Road Test performance rating losses in WASHO test pavements
panel. Ratings from five sessions with those predicted by preliminary
that included Illinois, Minnesota, AASHO Road Test performance
Indiana, and Road Test sections. equations. (Folder)
Mean ratings used to develop pres- 7681-7694—D Development of Rigid Pavement
ent serviceability index formulas. Performance Equations
(118-page printout) Graphs, tables, regression analyses,
7315-7319—D Measurements for Rated Pave- and formulas for various mathe-
ments matical models leading to rigid
Longitudinal and transverse profile pavement performance equations
summaries; cracking, patching, and given in Report 5. (Folders)
other measures of surface phenom- 8100—F Initial Plans for Flexible Pavement
ena for all sections rated by the Instrumentation
performance rating panel. Measure-
ments used to develop present serv- Information on installation of set-
iceability index formulas. (Folders, tlement rods, curvature strips, pres-
12-page printout) sure cells, deflection gages, frost
depth indicators, and thermocouples
7322—D Serviceability Index Data and Asso- as proposed during planning stages
ciated Measurements for All Test of the Road Test. (Folder)
Road Sections
Primary data system for test sec- 8150—F Curvature Strip Calibration
tion performance. Biweekly values Gage factors determined during
for each test section and subsection calibration of curvature strips.
in every flexible and rigid pavement (Folder)
experiment. Includes cracking, 8151—F Curvature Strip Installation
patching, rut depth, slope variance,
roughometer data, and present serv- Fabrication and calibration proce-
iceability index for each wheelpath dures for curvature strips in flexible
of each section. (1,056—page print- pavements. (Folder)
out) 8170—L Results of Experiments on Use. of
7350—D Seasonal Weighting Factor Data for AASHO Nuclear Density Gage on
Flexible Pavement Rough Stone Base Surface
Biweekly deflection values for selec- Used to establish procedures for
ted Loop 1 flexible pavement sec- trench studies. (Folder)
tions; seasonal factor values and 8171—L Limited Tests on Ability of Nuclear
weighted applications for each Gage to Determine Density of Asphalt
index day. (1-page printout) Pavement (Folder)
7611—D Development of Present Serviceability 8172—L Investigation of Methods for Obtain-
Index Formulas ing Density of Bituminous Specimens
Correlations and regression analy- Parafin and ' vacuum saturation
ses to determine present service- methods. (Folder)
342 THE AASHO ROAD TEST, REPORT 5
8173—L Studies of Extraction Test Procedure 8370—L Development of AASHO Model Nu-
Studies of accuracy of reflex-type clear Density Gage
extractor. (Folder) Results of studies leading to final
model used during construction of
8174—L Calibration of Marshall Compactor base course. (Folder)
Tests to establish proper test pro- 8371—L Evaluation of Nuclear Moisture and
cedure for mechanical compactor.
(Folder) Density Gages
Results of experiments reported in
8250—R Calibration of Vans 3 and 4 published paper. (Folders)
Detailed calibration procedure for 8372—L Study of Operator Variability in
instrumentation in vans 3 and 4, Determination of Liquid and Plastic
rigid pavement instrumentation. Limits of Soils.
(Folder)
Data for report published in HRB
8290—L Investigation of Concrete Testing Pro- Abstracts, October 1961. (Folder)
cedures 8373—L Methods for Obtaining In-Place Den-
Studies of cylinder molds, capping sity of Soils and Granular Bases
compound, testing machine, curing Results of several investigations
pit. (Folder) made primarily to establish pro-
8291—L Comparison Between Concrete Test- cedures for use during construction.
ing Crews (Folder)
Studies of two different testing 8376—L Study of Variations and Control of
crews used during construction of Gradation of Granular Bases
concrete pavements. (Folder) Inter-laboratory studies; quality
8300—R Slope Wheel Calibration control chart analyses of plant pro-
duction. (Folder)
Calibration records for profilometer
slope wheels. (Folder) 8377—L Calibration of Infrared Heat Lamp
Oven
8301—R Slope Wheel Calibration, Horizontal
Reference Oven used for mass production dry-
ing of soil samples during construc-
Data System 8300 when horizontal tion of Road Test embankment.
reference is used. (Folder) (Folder)
8302—R Longitudinal Profile of Special Test 8378—L Sonic Apparatus
Course Information on sonic apparatus
Profilometer check runs on special built for obtaining fundamental
calibration course, for control chart transverse and torsional frequencies
analysis of profilometer system. of concrete specimens. (Folder)
(Folder)
8379—L One-Point Liquid Limit Procedure
8303—R Profilometer Calibration Check Test data for "Rapid Determination
Continuation of Data System 8302. of Liquid Limit of Soils by Flow In-
(Folder) dex Method," by H. Y. Fang, HRB
8304—R Calibration of Longitudinal Profilo- Bulletin 254, 1960. (Folder)
meter 9100—S Flexible Condition Survey Summary,
Slope wheel calibration using auto- Loop 2, Post-Traffic Studies
matic chart reader. (Folder) Measurement by wheelpath of
cracking, sealed areas, patches,
8305—R Roughometer Correlation Study overlays. (Folders, 34-page print-
Background information on all out)
roughometers brought to the
AASHO Road Test for correlation 9101—S Accumulated Applications and Dates
with the AASHO roughometer and for Sections Out of Test, Loop 2, Post-
profilometer. (Folder) Traffic Studies
Includes initial present serviceabil-
8307—R Road Test Profilometer Calibration ity index, type of tire design (con-
Constants ventional or low-pressure low
Listing of all calibration constants silhouette), day the section went
used in converting field readings to out of test, accumulated applica-
slope variance. (Folders, 6-page tions for the section. (Folders, 2-
printout) page printout)
APPENDIX I 343
9121—S Beam Deflections, Loop 2, Post-Traffic 9154—S Flexible Pavement Dynamic Deflec-
Studies tions, Loop 6, Military Vehicle Post-
Benkelman beam deflections using Traffic Studies
AASHO normal procedure with 3- Total embankment and subbase de-
kip wheel load and 32-kip axle load, flections for all units except the
measured at two locations in each HETAG and GOER. (Folders, 4-
wheelpath. Individual data, range, page printout)
mean, and mat temperatures. (20- 9 155—S Flexible Pavement Dynamic Deflec-
page printout) tions, Loop 6, Military Vehicle Post-
9123—S Flexible Pavement Rut Depth Meas- Traffic Studies
urements, Loop 2, Post-Traffic Studies Total, embankment, and ,subbase de-
Mean rut depth data for each flections for the HETAG and GOER
wheelpath of each section. (Fold- units. (Folders, 3-page printout)
ers, 7-page printout) 9156—S Flexible Pavement Drop Test, Loop 4,
9126—S Flexible Pavement Transverse Pro- Breaking, Impact and Acceleration
files, Loop 2, Post-Traffic Studies Post-Traffic Studies
Rod and level measurements of Total deflections and embankment
transverse profiles for flexible pave- pressures for combinations of axle
ments, at 12 locations at 1-ft inter- load, vehicle speed, and drop dis-
vals. (Folders, 6-page printout) tance. (Folders, tapes, 10-page
printout)
9150—S Flexible Pavement Dynamic Deflec-
tions, Loop 6, Tire Pressure-Tire De- 9165—S Flexible Pavement Embankment Pres-
sign, Post-Traffic Studies sure, Loop 4, Military Vehicle Post-
Total, embankment, and subbase Traffic Studies
deflections for each combination of Embankment pressures for combin-
axle load tire pressure, tire size, ations of axle load, vehicle speed,
and cord type. Each deflection and longitudinal and transverse
value is mean of at least four read- placements, HETAG and GOER
ings at controlled transverse place- units. (Folders, 38-page printout)
ment. (Folders, tapes, 14-page 9166—S Flexible Pavement Embankment Pres-
printout) sure, Loop 4, Tire Pressure-Tire De-
9151—S Flexible Pavement Dynamic Deflection sign Post-Traffic Studies
Studies, Loop 2, Post-Traffic Studies Embankment pressures for combi-
Total, embankment, and subbase nations of tire pressure, tire size,
deflections for various combinations cord type, vehicle speed, and place-
of axle load, tire pressure, and tire ment. (Folders, 27-page printout)
size. Each value is mean of at least 9167—S Flexible Pavement Embankment Pres-
four readings at controlled trans- sures, Loop 4, Military Vehicle Post-
verse placement. (Folders, 1-page Traffic Studies -
printout) Embankment pressures for all units
9152—S Flexible Pavement Dynamic Deflec- except the HETAG and GOER,
tions, Loop 6, Commercial Construc- combinations of axle load, vehicle
tion Equipment, Post-Traffic Studies speed, longitudinal and transverse
Total, embankment, and subbase de- placement. (Folders, 10-page print-
flections for various combinations out)
of axle load, tire pressure, and ve- 9168—S Flexible Pavement Embankment Pres-
hicle speed. Each value is mean of sure, Loop 4, Commercial Construc-
six field readings at controlled tion Equipment, Post-Traffic Studies
transverse placement. (Folders, 2- Embankment pressures for com-
page printout) binations of axle load, tire pressure,
9153—S Flexible Pavement Dynamic Deflec- vehicle speed, and, placement.
tions, Loop 6, Special Suspension (Folders, 25-page printout)
Post-Traffic Studies 9169-5 Flexible Pavement Embankment Pres-
Total, embankment, and subbase de- sure, Loop 4, Special Suspension Sys-
flections for combinations of axle tems, Post-Traffic Studies
load and vehicle speed. Each value Embankment pressures for combi-
is mean of six readings at con- nations of axle load, vehicle speed,
trolled transverse placement. (Fold- and placement. (Folder, 12-page
ers, 6-page printout) printout)
344 THE AASHO ROAD TEST, REPORT 5
9170—S Stiffness Measurements, Shell Oil Vi- 9250—S Rigid Pavement Strains and Deflec-
brator Post-Traffic Studies tions, Loop 6, Tire Pressure-Tire
Measurements of force ratio, dy- Design Post-Traffic Studies
namic stiffness, and phase angle on Pavement edge and slab corner
pavement by Shell Oil vibrator. strains and deflections for combina-
(Folder, 38-page printout) tions of axle load, tire pressure, tire
size, cord type and vehicle speed.
9171—S Velocity Measurements, Shell Oil Vi- (Folders, tapes, 8-page printout)
brator Post-Traffic Studies
Wave length and velocity measure- 9251—S Rigid Pavement Routine Strain and
ments with Shell Oil vibrator. Deflection Studies, Loop 2, Post-
(Folders, 35-page printout) Traffic Studies
Strains and deflections at pavement
9180—S Flexible Pavement Trench Measure- edges and corners for combinations
ments, Before and After Post-Traffic of axle load, tire pressure, tire size,
Studies, Loop 2 and vehicle speed. (Folders, 1-page
Moisture content, density, satura- printout)
tion, CBR, plate load tests, and 9252—S Rigid Pavement Strains and Deflec-
gradations on base, subbase and tions, Loop 6, Commercial Construc-
embankment soil. (Folders, 3-page tion Equipment Post-Traffic Studies
printout)
Strains and deflections at pavement
9181—S Flexible Pavement Trench Measure- edge and corner for combinations of
ments in Failed Areas, Loop 2, Post- axle load, tire pressure, and vehicle
Traffic Studies speed. (Folders, 4-page printout)
Moisture content, density, satura- 9253—S Rigid Pavement Strains and Deflec-
tion, CBR, and plate load test on tions, Loop 6, Special Suspension Sys-
base, subbase and embankment soil tems Post-Traffic Studies
in Loop 2 failed areas during post- Strains and deflections at pavement
traffic period. (Folders, 1-page edges and slab corners. (Folders,
printout) 6-page printout)
9182—S Flexible Pavement Edge Sampling 9254—S Rigid Pavement Strains and Deflec-
Measurements, Loop 2, Post-Traffic tions, Loop 6, Military Vehicle Post-
Studies Traffic Studies
Moisture content of base and sub- Pavement edge and slab corner
base; moisture content, density, and strains and deflections for all units
saturation on embankment soil. All except the HETAG and GOER.
measurements along edge of pave- (Folders, 8-page printout)
ment. (Folders, 4-page printout)
9255—S Rigid Pavement Strains and Deflec-
9200—S Rigid Pavement Condition Survey, tions, Loop 6, Military Vehicle Post-
Loop 2, Post-Traffic Studies Traffic Studies
Wheelpath measurements of crack- Pavement edge and corner strains
ing, patching, faults, corner breaks, and deflections for HETAG and
and faults at joints. (Folders, 29- GOER units. (Folders, 6-page
page printout) - printout)
9201—S Rigid, Pavement Condition Survey 9256—S Rigid Pavement Drop Test, Loop 6,
History, Loop 2, Post-Traffic Studies Braking, Impact and Acceleration
Section measurements of cracking, Post-Traffic Studies
spalling, patching, faults and Compression strain and deflection,
cracks; corner breaks and faults at tension strain for combinations of
joints. (Folders, 15-page printout) axle load, vehicle speed and drop
distance. '(Folders, tapes,' 5-page
9242—S Rigid Pavement Pumping Survey, printout)
Loop 2, Post-Traffic Studies
Pumping scores for Loop 2 test sec- 9263—S Rigid Pavement Trench Measure-
tions. (4-page printout) ments, Loop 2, Before and After Post-
Traffic Studies
9243—S Rigid Pavement Pumping Survey Moisture content, density, satura-
Summary, Loop 2, Post-Traffic Studies tion, CBR, plate load tests, and
Summation of pumping score for gradations on subbase and embank-
each section after each rain. (4- ment soil. (Folders, 3-page print-
page printout) out)
APPENDIX I 345
9280—S Rigid Pavement Static Rebound De- 9410—S Post-Traffic Bridge Studies, Special
flections, Loop 2, Post-Traffic Studies Suspension System with Concentric
Corner and edge deflections with 6- Loading, Series 0
and 32-kip axle loads at two loca- Stresses computed from measured
tions on each section. (8-page print- strains using the modulus of elas-
out) ticity for steel. (Folders, tapes, 26-
page printout)
9282—S Rigid Pavement Edge Sampling Meas
urements, Loop 2, Post-Traffic Studies 9411—S Post-Traffic Bridge Studies, Commer-
Moisture content of subbase; mois- cial Construction Equipment with
ture content, density, and satura- Concentric Loading, Series 1
tion of embankment soil. All meas- Continuation of Data System 9410.
urements taken along pavement (26-page printout)
edge. (Folders, 2-page printout)
9412—S Post-Traffic Bridge Studies, Military
9298—S Rigid Pavement Overlay Condition Vehicle with Special Suspension Sys-
Survey, Loop 2, Post-Traffic Studies tem and Concentric Loading, Series 2
Wheelpath measurements of crack- Continuation of Data System 9410.
ing, sealed areas, patching and (26-page printout)
overlay. (1-page printout)
9413—S Post-Traffic Bridge Studies, Military
9299—S Rigid Pavement Overlay Rut Depths, Vehicle with Concentric Loading,
Loop 2, Post-Traffic Studies Series 3
Rut depth data for each wheelpath Continuation of Data System 9410.
of each section. (Folders, 2-page (26-page printout)
printout)
9414—S Post-Traffic Bridge Studies, HETAG
9301—S Tire Pressure Studies Vehicle with Concentric Loading,
Calibration curves and measure- Series 4
ment procedures associated with Continuation of Data System 9410.
tire pressure equipment. For use (26-page printout)
with computing differential load
from tire pressure records. (Fold- 9420—S Post-Traffic Bridge Studies, Commer-
ers) cial Construction Equipment, Dy-
namic Load Effects on Slab, Series 0
9302—S Tire Pressure Studies
Records and procedures for detr- Stresses computed from measured
mining relationships between load strains. (Folders, tapes, 6-page
changes and tire pressure changes. printout)
(Folders, tapes) 9421—S Post-Traffic Bridge Studies, Military
9322—S Rigid and Flexible Pavement Per- Vehicle, Dynamic Load Effect on Slab,
formance, Loop 2, Post-Traffic Studies Series 1
Cracking, patching, slope variance, Continuation of Data System 9420.
rut depth data, and serviceability (6-page printout)
index values for each test section. 9422—S Post-Traffic Bridge Studies, Commer-
(34-page printout) cial Construction Equipment, Dy-
9400—S Post-Traffic Bridge Study, Tire Pres- namic Load Effects on Slab, Series 2
sure-Tire Design Index, Series 0 Continuation of Data System 9420.
Stresses computed from measured (6-page printout)
strains using modulus of elasticity 9430—S Post-Traffic Bridge Study, Drop Tests,
for steel. (Folders, tapes, 20-page Series 0
printout) -
Measurements by drop tests of
9401—S Post-Traffic Bridge Studies, Tire Pres- beam stresses and deflections at two
sure-Tire Design Effects, Series 1 locations, off and on bridge. (Fold-
Continuation of Data System 9400. ers, tapes, 2-page printout)
(21-page printout) 9440—S Post-Traffic Bridge Studies, Brake
9402—S Post-Traffic Bridge Studies, Tire Pres- Tests, Series 1
sure-Tire Design Effects, Summary Measurements by break test of
Mean values of stresses and deflec- bridge beam stresses. Brakes ap
tions by vehicle speeds and vehicle plied when vehicle on test span.
numbers. (10-page printout) (Folder, tapes, 1-page printout)
Appendix J
Region 1
F. M. Auer, Planning and Economics Engineer, C. D. Jensen, Director of Research and Testing,
New Hampshire Department of Public Pennsylvania Department of Highways
Works and Highways W. McAlpin, Assistant Deputy Chief Engi-
B. Bly, Engineering Assistant to Commis- neer (Research), New York State Depart-
sioner, Vermont Department of Highways ment of Public Works
T. V. Bohner, Special Assistant, Engineering J. F. McGovern, Structures Maintenance Engi-
neer, Massachusetts Department of Public
Department, D. C. Department of High- Works
ways and Traffic
L. W. Novinger, Contract and Design Engineer,
W. M. Creamer, Chief, Highway Staff Services, Delaware State Highway Department
Connecticut State Highway Department A. Savage, Engineer of Primary Highways,
W. Hauck, Supervising Civil Engineer Maine State Highway Commission
(Road Designing), Rhode Island Depart- Van Breemen, Research Engineer, New
ment of Public Works Jersey State Highway Department
Region 2
Shelburne, Chairman, Director of Re- A. 0. Neiser, Assistant State Highway Engi-
search, Virginia Department of Highways neer, Kentucky Department of Highways
Abercrombie, Engineer of Materials and T. W. Parish, Assistant Chief Engineer (Con-
Tests, Georgia State Highway Department struction), Louisiana Department of High-
ways
L. Bransford, Engineer of Research and In-
Service Training, Florida State Road De- R. S. Patton, Engineer of Surveys and Designs,
partment Tennessee Department of Highways and
Public Works
L. D. Hicks, Chief Soils Engineer, North Caro- Angel (2) Silva, Director, Puerto Rico Depart-
lina State Highway and Public Works ment of Public Works
Commission
0. Thompson, Testing Engineer, Mississippi
W. McAlpin, Director, Program Office, and State Highway Department
Assistant Chief Engineer, West Virginia J. F. Tribble, Materials and Research Engineer,
State Road Commission Alabama State Highway Department
J. D. McMahan, Construction Engineer, South E. L. Wales, Engineer of Materials and Tests,
Carolina State Highway Department Arkansas State Highway Commission
346
APPENDIX J 347
Region 3
E. Chastain, Sr., Chairman, Engineer of H. E. Marshall, Research Engineer, Ohio De-
Physical Research, Illinois Division of partment of Highways
Highways R. L. Peyton, Assistant State Highway Engi-
J. G. Butter, Consultant, Iowa State Highway neer, State Highway Commission of
Commission Kansas
E. A. Finney, Director, Research Laboratory, J. S. Piltz, Engineer of Desigi, Wisconsin State
Michigan State Highway Department Highway Commission
C. K. Preus, Materials and Research Engineer,
R. A. Helmer, Research Engineer, Oklahoma Minnesota Department of Highways
State Highway Department F. V. Reagel, Engineer of Special Assignments,
J. W. Hossack, State Engineer, Nebraska De- Missouri State Highway Commission
partment of Roads W. T. Spencer, Soils Engineer, Indiana State
C. P. Jorgensen, Manager, Research and Plan- Highway Department
ning, South Dakota State Highway Com- W. A. Wise, Director, Field Division, North
mission Dakota State Highway Department
Region 4
R. E. Livingston, Chairman, Planning and Re- C. W. Johnson, Materials and Testing Engi-
search Engineer, Colorado Department of neer, New Mexico State Highway Commis-
Highways sion
J. R. Bromley, Superintendent and Chief Engi- D. F. Larsen, Chief Materials Engineer, Utah
fleer, Wyoming State Highway Department State Road Commission
L. F. Erickson, Assistant Construction Engi- C. E. Minor, Materials and Research Engineer,
fleer, Idaho Department of Highways Washington Department of Highways
L. B. Fox, Construction Engineer, Montana W. G. O'Harra, Materials Engineer, Arizona
State Highway Commission Highway Department
T. S. Huff, Chief Engineer of Highway Design, W. M. Wachter, Highway Engineer, Hawaii
Texas State Highway Department Division of Highways
F. N. Hveem, Materials and Research Engi- W. 0.. Wright, State Highway Engineer, Neva-
fleer, California Division of Highways da Department of Highways
ADVISORY PANELS
These panels were among those appointed by the Highway Research
Board to advise and assist the Board and its project staff on matters
related to the pavement research. A complete listing of panels is given in
AASHO Road Test Report 1 (Special Report 61A).
Instrumentation
This panel advised on available means for measuring physical phe-
nomena and reviewed the work of the instrumentation systems contractors.
Maintenance
This panel advised on pavement maintenance techniques and formulated
criteria for specific types of maintenance.
Performance Rating
This panel advised and aided the staff in the development of a system
for rating the performance of the test sections.
The following served as members of this panel during the years indi-
cated:
R. C. Boyd, Maintenance Engineer, Iowa State J. M. Griffith, Engineer of Research, The As-
Highway Commission (Resigned 1959) phalt Institute (1956- Resigned March 31,
A. A. Anderson, Chief Highway Consultant, 1961)
Portland Cement Association (1956-1960)
APPENDIX J 349
Statistical
This panel advised on matters relating to experiment design and data
analysis.
Data Analysis
This panel advised on matters relating to the analyses of the AASHO
Road Test data and presentation of findings.
PROJECT PERSONNEL
Staff During Research Phase
W. B. McKendrick, Jr., Project Director W. R. HudsOn, Assistant Rigid Pavement Re-
W. N. Carey, Jr., Chief Engineer for Research search Engineer
Peter TalOvich, Business Administrator R. J. Little, Assistant Rigid Pavement Re-
L. A. Ptak, Accountant search Engineer
R. S. Semple, Purchasing Assistant M. Viest, Bridge Research Engineer
J. W. Fisher, Assistant Bridge Research
A. C. Tosetti, Assistant to the Project Director Engineer
W. R. Milligan, Assistant Operations Man- P. E. Trick, Chief, Data Processing and Analy-
ager sis
D. L. Thorp 1 Shop Superintendent R. C. Ham, Assistant Chief, Data Processing
A. C. Benkelman, Flexible Pavement Research and Analysis
Engineer F. Shook, Materials Engineer
L. E. Dixon, Assistant Flexible Pavement D. R. Schwartz, Engineer of Reports
Research Engineer H. R. Hubbell, Assistant Engineer of Reports
H. M. Schmitt, Assistant Flexible Pavement H. H. Boswell, Maintenance Engineer
Research Engineer James Gardner, Maintenance Superintendent
F. H. Scrivner, Rigid Pavement Research Engi- R. C. Leathers, Engineer of Special Assign-
neer ments
H. C. Huckins, Instrumentation Supervisor
H. H. Cole (1958) W. J. Schmidt, Chief, Public Information
Other Engineering Personnel
0. B. Andersland T. W. DeVries H. Y. Fang J. F. Reynolds
E. L. Skok, Jr. R. K. Williamson
R. I. Kingham, Canadian Good Roads Associa- S.' M. King, American Trucking Associations
tion (1957-1961)
G. D. Campbell, Canadian Good Roads Associa- R. A. Liii, American Trucking Associations
tion (1956-1957) (1955-1957)
B. E. Coiley, Portland Cement Association E. Teske, Portian :1 Cement Association
(1956-1960) W. (1957-1961)
G. A. Wrong, Provinc of Ontario, Canada
F. N. Finn, The Asphalt Institute (1956-1960) (1958-1959)
Temporary Personnel
The following engineers were assigned to the project by the Bureau of
Public Roads for periods of about six months each. They were assigned to
the various branches, where they served in important engineering and
technical capacities.
Materials Construction Flexible Pavement Bridge Research
Branch Branch Research Branch Branch
G. Yemington W. A. Eager R. H. Hogrefe J. W. Schmidt
J. P. Clark J. C. Becker E. E. Biggs G. N. Lind
G. R. Brooks T. J. Chipera C. C. Berge V. Buchele
W. S. Dunbar R. H. Jones 0. M. Stump D. C. Briggs
H. Marshall D. B. Lewis W. K. Perry G. C. Hoxie
A. R. Cowan Donald Jacobsen P. M. Jorgensen R. A. Richter
Daniel Dake Claude Manaton D. K. Phillips G. W. Million
G. L. Green V. W. Segelke R. G. Shutt ' L. R. Cayes
R. L. Lacy C. H. Snow D. C. McConnon J. L. Budwig
J. R. Bishop Rothe Davis H. Kusumoto C. F. Galambos
W. H. Bray G. K. Hossner Dallas Vestal N. C. Mueller
G. D. Gibson G. S. Katayama W. T. Medley
R. D. Gingrich D. J. Philbrick
R. C. Kay K. D. Jaeger
J. H. Hatton
R. E. Stanford
N. W. Loeffler
.1 Rigid Pavement Data Analysis Public Information
Research Branch Branch Branch
R. W. Hayman R. A. Lawrie R. A. Van de
T. E. Difloe R. E. McGuire Meulebrocke
C. W. Friesen R. E. Gish D. F. Berwick
Stewart Spelman Robert Talley , H. H. Ridgeway
L. L. Humphrey R. H. Gausman T. H. Lavender
R. D. Morgan R. B. Puckett P. E. Cunningham
T. 0. Willett J. S. Bowers
W. Briggs D. C. Lewis Operations
L. P. Lamm N. L. Arthur Branch
W. G. White N. J. Van Ness K. B. Casey
D. G. Ross J. S. Wesley
W. S. Mendenhall, Jr. Maintenance R. L. Diffenderfer
A. R. Montgomery Branch
G. E. Price C. A. Ballinger
D. E. Carison
352 THE AASHO ROAD TEST, REPORT 5
Photography
Still photography was done by Frank W. Bazzoni of the staff. His work
was augmented by and all motion picture photography and the production
of motion pictures relating to the project were done by Bureau of Public
Roads photographers Roy B. Dame, T. Welby Kines, George Crum, and
Charles Ritter.
,,
(I