Pavement Testing

MAY 2000
ISBN 99912 - 0 - 315 - X

Pavement Testing,
Analysis and
Interpretation of
Test Data
Ministry of Works, Transport & Communications,
Roads Department
Private Bag 0026
Gaborone, Botswana
Phone + 267 - 313511
Fax + 267 - 314278
MAY 2000
ISBN 99912 - 0 - 315 - X
Reproduction of extracts from this Guideline may be made subject to due acknowledgement of the source.
Although this Guideline is believed to be correct at the time of printing, Roads Department does not accept any
contractual, tortious or other form of liability for its contents or for any consequences arising from its use. Anyone
using the information contained in the Guideline should apply their own skill and judgement to the particular issue
under consideration.
2 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

ROADS DEPARTMENT
Under the policy direction of the Ministry of Works, Transport & Communications, Roads Department is
responsible for providing an adequate, safe, cost-effective and efficient road infrastructure within the borders of
Botswana as well as for facilitating cross-border road communications with neighbouring countries. Implied in
these far-ranging responsibilities is the obligation to:
1. ensure that existing roads are adequately maintained in order to provide appropriate level of service for
road users;
2. improve existing roads to required standards to enable them to carry prevailing levels of traffic with the
required degree of safety;
3. provide new roads to the required geometric, pavement design and safety standards.
The Department has been vested with the strategic responsibility for overall management of the Public Highway
Network (PHN) of some 18, 300 km of roads. This confers authority for setting of national specifications and
standards and shared responsibility with the District Councils and Department of Wildlife and National Parks for
the co-ordinated planning of the PHN.
Roads Department is also responsible for administering the relevant sections of the Public Roads Act, assisting
local road authorities on technical matters and providing assistance in the national effort to promote citizen
contractors in the road construction industry by giving technical advice wherever possible. This task is facilitated
by the publication of a series of Technical Guidelines dealing with standards, general procedures and best
practice on a variety of aspects of the planning, design, construction and maintenance of roads in Botswana that
take full account of local conditions.
In its endeavour to provide uniformity of practice in the provision of efficient and effective road infrastructure,
Roads Department has embarked on the preparation and publication of a number of Technical Guidelines. The
main objective of these Guidelines is to document best practice and to preserve local knowledge on a variety of
aspects of road planning, design, construction and maintenance that have evolved over many years in Botswana.
Guideline No. 1 The Design, Construction and Maintenance of Otta Seals (1999)
Guideline No. 2 Pavement Testing, Analysis and Interpretation of Test Data (2000)
Pavement Testing, Analysis and Interpretation of Test Data Roads Department 3

FOREWORD
Pavement testing and interpretation of test data is a fundamental phenomenon in the proper monitoring and
maintenance of road pavements in an economic way.
The first generation roads in Botswana have either reached, exceeded or are about to reach their design life and
would require major rehabilitation. The traffic has increased tremendously on the Road Network, both in terms of
number and axle loading, which means that the rehabilitation will require a considerable amount of funding. The
optimal use of such funds will depend on the timely intervention of maintenance and rehabilitation using the most
appropriate technology.
A significant proportion of the work of the Materials and Research Division of Roads Department involves both
pavement field testing and the analysis and interpretation of pavement test data. This information is used for the
determination of the pavement condition and maintenance requirements and to guide the road rehabilitation
plans.
Hence, it is in recognition of the importance of accurate data collection and its interpretation that guidelines on
the “Pavement Testing, Analysis and Interpretation of Test Data” had to be prepared in order to improve and
streamline in-house capacity for pavement assessment and reporting as well as ensure continuous use of compu-
ter tools and analysis techniques used in the process, some of which were developed in-house.
Gaborone 15th May 2000
Andrew Nkaro
Acting Director of Roads
Roads Department
Ministry of Works, Transport and Communications

ACKNOWLEDGEMENTS
This Guideline is one of a series of Guidelines that is being compiled under the Institutional Co-operation Agreement
that exists between the Roads Department and the Norwegian Public Roads Administration (NPRA). This
Agreement falls under a NORAD Technical Assistance Programme to Roads Department that is co-funded by
the Kingdom of Norway and the Government of the Republic of Botswana.
The production of the Guideline has been a joint project by Transportek-CSIR, Pretoria, Transport Research
Laboratory (TRL), United Kingdom and Stewart Scott International (SSI) Inc. who was the consultant for the
Working Group. The staff members involved were:
Dr Philip Paige-Green, CSIR
Dr Steve Kekwick, CSIR
Mr Colin Jones, TRL
Mr Harry Smith, TRL
Mr Philip Joubert, SSI
Mr Mihut Motrescu, SSI
The Working Group that guided the project and reviewed the Guideline, comprised of the following people:
Dr Bernard Obika, Roads Department
Mr Charles Overby, NPRA
Mr Barry Kemsley, Roads Department
Mr Baele Mmoloki, Roads Department
Mr Modise Segokgo, Roads Department
Mr Elliott Maswikiti, Roads Department
Mr Mogotsakgotla Kowa, Roads Department
Mr B Sharma, Roads Department
Photographs were provided by:

Charles Overby, NPRA
Transportek, CSIR

LIST OF ABBREVIATIONS
AASHTO American Association of State Highway and Transportation Officials
BI Bump Integrator
BLI Base Layer Index
BRMS Botswana Road Management System
CAM Crack Activity Meter
CBR California Bearing Ratio
CSIR Council for Scientific and Industrial Research, South Africa
DCP Dynamic Cone Penetrometer
DSN DCP Structural Number over depth of 800 mm
EMC Equilibrium Moisture Content
E80 Equivalent Standard Axle
FWD 800 Falling Weight Deflectometer
HDM Highway Design and Maintenance (World Bank series)
HGV Heavy Goods Vehicle
HSP High Speed Profilometer
IDMP Impuls Deflection Measurement Parameters
Im Thornthwaites moisture index
IRI International Roughness Index
LDI Linear Displacement Integrator
LHS Left Hand Side
LLI Lower Layer Index
LS Linear Shrinkage
MCCSSO Moisture, Colour, Consistency, Structure, Soil type, Origin, used in Soil Profiling
MGV Medium Goods Vehicle
MLI Middle Layer Index
MR Resilient Modulus
NAASR National Association of Australian State Road Authorities
NORAD Norwegian Agency for Development Co operation
NPRA Norwegian Public Roads Administration
OMC Optimum Moisture Content
PI Plasticity Index
PL Plastic Limit
PSI Present Serviceability Index
PSR Present Serviceability Rating
QI Quartercar Index
RHS Right Hand Side
RMS Road Management System
RTRRMS Response Type Road Roughness Measuring System
SAMDM South African Mechanistic Design Method
SATCC Southern Africa Transportation Communications Commission
SCRIM Sideways-force Coefficient Routine Investigation Machine
SN Structural Number
SNC Modified Structural Number
SSI Stewart Scott International
TMH Technical Methods for Highways
TRH Technical Recommendations for Highways
TRL Transport Research Laboratory, United Kingdom
UCS Unconfined Compressive Strength
VCI Visual Condition Index
VHGV Very Heavy Goods vehicle
VR Roughness Values

TABLE OF CONTENTS page
Roads Department ..................................................................................... 3
Foreword .................................................................................................... 4
Acknowledgements .................................................................................... 5
PART A GENERAL
1 INTRODUCTION
1.1 Background ....................................................................... 14
1.2 Purpose and scope of Guideline ........................................ 14
1.3 Structure of the Guideline .................................................. 15
2 ROAD PAVEMENTS
2.1 The road function .............................................................. 16
2.2 Principles of road maintenance economics ....................... 16
2.3 Typical Pavement Structures ........................................... 17
2.3.1 Surfacings..........................................................19
2.3.2 Pavement layers ................................................... 19
2.3.3 Subgrade............................................................19
2.3.4 Sealed/unsealed shoulders .................................... 19
2.4 The environment ............................................................... 20
2.5 Traffic loading ................................................................... 20
3 TYPES OF PAVEMENT DISTRESS
3.1 Causes and types of pavement distress ............................ 21
3.2 Roughness ......................................................................... 23
3.3 Deformation ...................................................................... 23
3.4 Cracking ............................................................................ 24
3.5 Surface problems and surface texture .............................. 25
3.6 Pumping............................................................................. 26
3.7 Potholes and patching ........................................................ 26
3.8 Edge break ........................................................................ 26
3.9 Moisture/drainage .............................................................. 26
3.10 Soluble salt damage ........................................................... 27
4 PLANNING AND SELECTION OF APPROPRIATE
TESTING
4.1 Purpose of pavement evaluation ....................................... 28
4.1.1 Structural evaluation ............................................. 28
4.1.2 Functional evaluation ............................................. 28
4.2 Existing information ........................................................... 28
4.2.1 As built data .......................................................... 29
4.2.2 Maintenance history .............................................. 29
4.2.3 Historical traffic .................................................... 29
4.2.4 Climatic records .................................................... 30
4.3 Evaluation framework and selection of tests .................... 30
4.3.1 Evaluation framework ........................................... 30
4.3.2 Initial assessment .................................................. 31
4.3.3 Detailed assessment ............................................. 31
4.4 Test frequency .................................................................. 31

PART B EVALUATION METHODS page
5 VISUAL EVALUATION
5.1 Methodology ...................................................................... 35
5.2 Data collection .................................................................. 35
5.3 Analysis and presentation .................................................. 37
6 RUT DEPTH MEASUREMENT

6.1 General .............................................................................. 38
6.2 Test procedures ................................................................. 38
6.3 Test frequency .................................................................. 38
7 ROUGHNESS MEASUREMENT
7.1 General .............................................................................. 40
7.2 Test procedures ................................................................. 41
7.3 Test frequency .................................................................. 42
8 PAVEMENT STRENGTH
8.1 General .............................................................................. 43
8.2 Deflection measurements .................................................. 44
8.2.1 Type of equipment ................................................ 44
8.2.2 Test procedures ..................................................... 45
8.2.3 Test frequency ...................................................... 45
8.2.4 Analysis and presentation .................................... 45
8.3 Dynamic Cone Penetrometer (DCP)................................ 46
8.3.1 General .................................................................. 46
8.3.2 Test procedures ..................................................... 46
8.3.3 Test frequency ...................................................... 46
8.3.4 Analysis and presentation .................................... 46
9 TEST PIT PROFILING AND SAMPLING

9.1 General .............................................................................. 49
9.2 Location............................................................................. 49
9.3 Size and depth ................................................................... 50
9.4 In situ measurements ........................................................ 50
9.5 Description and logging of the soil profile ......................... 51
PART C PAVEMENT DATA ANALYSIS

10 DATA ANALYSIS
10.1 Data analysis phases and methodologies .......................... 54
10.2 Chainage adjustment ......................................................... 54
10.3 Identification of uniform section ........................................ 56
10.4 Data evaluation ................................................................. 58
11 INTERPRETATION OF RESULTS
11.1 Simple methods of data evaluation .................................... 59
11.2 Parameters obtained from analyses .................................. 59

11.3 Estimation of remaining life ............................................... 60
11.4 Sensitivity analyses ............................................................ 61
PART D REFERENCES AND

APPENDICES page
REFERENCES.................................................................................63
APPENDICES
Appendix A Visual evaluation forms and description ............................ 65
Appendix B Visual Condition Index (VCI) ............................................ 67
Appendix C Roughness measurement .................................................. 68
Appendix D Pavement layer strength coefficients for
structural number (after Paterson, 1985)........................74
Appendix E Deflection measurement ................................................... 75
Appendix F Test pit profiles .................................................................. 83
Appendix G Equilibrium and predicted moisture content ....................... 86
Appendix H Cumulative sum method for identifying uniform sections....87
Appendix I Pavement condition rating form ........................................ 95
Appendix J Definition of terms ............................................................ 96
LIST OF TABLES page
Table 2.1 Composition of typical Botswana bituminous pavements......18

Table 2.2 Typical surfacing lives in Botswana (years)........................19
Table 3.1 Typical distress types associated with performance.............21
Table 3.2 Possible causes of traffic-associated distress......................22
Table 3.3 Possible causes of non traffic-associated distress................23
Table 4.1 Typical pavement evaluation data requirements...................30
Table 4.2 Suggested minimum testing and sampling frequencies.........33
Table 5.1 Degree and description of distress modes...........................36
Table 5.2 Description of extent of distress.........................................37
Table 7.1 Approximate comparison between IRI, QI and PSI
(after Paterson, 1987)....................................................... 41
Table 10.1 Chainage adjustment........................................................55
Table 11.1 Ranges of warning values for data interpretation............... 60
Table B1 Proposed weightings for VCI calculation
(modified after TRH 225)................................................ 64
Table E1 Repeatability of duplicate transient deflection tests............ 75

Table H1 Possible uniform sectioning for deflection data.................... 84
Table H2 Summary of deflection analysis........................................ 87
LIST OF FIGURES page
Figure 2.1
a-b Illustration of various design strategies on
maintenance and rehabilitation needs..................................17
Figure 2.2 Cross section terms and elements......................................17
Figure 2.3 Schematic illustration of various types of
bituminous surfacings.........................................................18
Figure 6.1 Example of relationship between deflection
and rut depth.....................................................................39
Figure 10.1Delineation of sections using as-built approach...................57
Figure C1 Operation of MERLIN......................................................65
Figure C2 Merlin data collection form.................................................67
Figure C3 Diagrammatical representation of the TRL
Integrator Unit fitted to a vehicle........................................69
Figure E1 Simplified pavements structure model.................................72
Figure E2 Diagrammatic representation of the Benkelman beam.........73
Figure E3 Diagrammatic representation of the FWD...........................77
Figure E4 Deflection bowl parameters...............................................79
Figure H1 Data and plots for Example 1.............................................84
Figure H2 Deflection data example.....................................................87
Figure H3 Outer wheel path data and 90th percentile
characteristics deflections...................................................89
Figure I1 Pavement condition rating form...........................................91

LAYOUT OF THE GUIDELINE
1 Introduction
Part A General 2 Road Pavements
3 Types of pavement
distress
4 Planning and selection

of appropriate testing
5 Visual evaluation
Part B Evaluation 6 Rut depth measurement

methods
7 Roughness measurement
8 Pavement strength
9 Test pit profiling and

sampling
Part C Pavement 10 Data analysis

data analysis
11 Interpretation of results
Part D References
Appendices

PART A GENERAL
1 Introduction
2 Road Pavements
3 Types of pavement distress
4 Planning and selection of
appropriate testing
1 Introduction
3 Types of pavement
distress

5 Visual evaluation

methods
8 Pavement strength

sampling

data analysis
Part D References
Appendices

1 INTRODUCTION
1.1 Background
Traffic is usually defined in terms of equivalent All roads are designed with a finite life, usually defined by the traffic that the
80 kN standard axles (E80). This is the number road can carry in terms of the cumulative number of equivalent standard
of 80 kN single axle loads that would cause the
same damage to a pavement as the actual axle axles (E80s). Once this design traffic has been carried, or as a result of
loads using the pavement. premature distress caused by some environmental influence, the road usually
needs to be rehabilitated.
The need for rehabilitation of a road is usually
identified from the Road Management System Prior to any rehabilitation design being carried out, it is necessary to fully
which shows increased deterioration of the
pavement in terms of its condition or maintenance
assess and evaluate the condition of the road pavement and to identify the
requirement. reasons for the distress. The time and resources required for these types of
investigation are generally limited and costly and thus it is essential that the
The consequences of incorrect decisions in
regard to pavement evaluations can signifi-
appropriate information be gathered and that it is presented in a systematic
cantly affect the ultimate rehabilitation design. and complete manner.
The primary reasons for undertaking pavement evaluations are to obtain
information:
n on the pavement condition;
n on the remaining life of the pavement;
n for research and development purposes;
n to assess newly built roads.
New road - one year after construction. Botswana is at the stage that many of the roads constructed during the late
1970s and early 1980s currently require, or will soon require, rehabilitation.
An increased need for pavement evaluation studies and rehabilitation designs
is thus envisaged.
1.2 Purpose and scope of guideline

The main purpose of this Guideline is to capture the knowledge and practice
that have been built up in Botswana (and elsewhere in the region where
Typical sign of pavement distress in the form of relevant) with respect to the evaluation of pavements. It provides practical
outer wheel path failures.
guidance on pavement evaluations and discusses the necessary techniques
and processes to carry them out correctly and cost-effectively and to present
the results in a standard and user-friendly format.
This Guideline should be used for the purpose intended, i.e. as a guideline,
and should not be used as a prescriptive approach, as the evaluation
requirements of each project are unique. The purpose of this guideline is to
assist the Botswana Roads Department staff in designing, carrying out and
presenting and interpreting the results of a pavement evaluation in a manner
that will provide the most useful results in a concise and cost-effective
manner.
A road section due to be rehabilitated after many
years in service. The Guideline is intended for use mainly by the staff of the Central Materi-
als Laboratory, Botswana Roads Department. It will also be important in the
preparation of Terms of Reference for pavement evaluations of sealed roads.
14 Roads Department Pavement Testing, Analysis and Interpreta of Test Data

1.3 Structure of the guideline
The guideline is structured in four main sections: Different evaluation techniques are necessary
for sealed and unsealed roads. This Guideline
PART A (Theoretical aspects) is specifically intended for use with sealed
The introduction that is given in this section is followed by a general overview roads.
of road pavements in Section 2 and types of distress in Section 3. The
planning of pavement evaluations including the selection of the appropriate
testing is discussed in Section 4.
PART B (Practical aspects)

Sections 5 to 9 summarise the test methods and techniques for the analysis
and presentation of the data. Each test technique, including:
n visual evaluation;
n rut depth measurement;
n roughness evaluation;
n pavement strength investigation;
n test pit profiling
is discussed individually.
PART C (Analytical aspects)

In Section 10, the data analysis and evaluation procedure is outlined and in
Section 11 simple preliminary methods for the interpretation of the data
collected are presented.
The process discussed in this guideline is summarised in the flow chart

below.
Analyse data
Uniform sections
Roughness Treatments
Visuals Rutting
RMS
Review
Recommendations
Identification Pavement Analysis and for optimal
Local reports interpretation
of distress evaluation upgrading
Other sources Strength Profiles
Deflections DCPs
PART D (References and Appendices)

Part D includes the References and Appendices. More detailed information
and standard data collection forms are also provided.

2 ROAD PAVEMENTS
2.1 The road function

Roads are constructed to perform a service to the road user. In this context,
the road needs to provide good trafficability/passability under all weather
Functional and structural performance are two conditions with maximum comfort and at minimum cost and with the highest
different aspects of road performance, which
are not necessarily related.
possible degree of safety. This is defined as the functional performance of
the road. The functional performance includes aspects such as the riding
quality, the skid resistance and the effective drainage of water from the pave-
ment surface.
The engineer, whilst recognising these functional requirements, regards the
pavement as a load bearing structure that is required to perform under the
prevailing traffic and environmental conditions with minimum maintenance.
These aspects are considered as the structural requirements of the road.
Roads are designed to provide a specific service in terms of their carrying
capacity under the expected pavement moisture regime and maintenance
programme. Any overstressing of the pavement by traffic or reduction in the
load-bearing capacity as a result of excessive ingress of water into the pave-
ment layers and subgrade can result in structural failure of the pavement.
Adequate drainage of the pavement layers and sub-
grade is essential for good pavement performance.
2.2 Principles of road maintenance

economics
The economics of road maintenance revolve around optimising the total
cost of construction, maintenance and rehabilitation for a particular road
over a specific period. The concepts of analysis period and design period
require careful consideration during the pavement design phase. The analysis
period is a planning period during which full reconstruction of the pave-
ment is undesirable and is a function of the road category but is also frequently
related to the roads’ geometric life. Where the traffic situation on a road is
likely to change considerably in the short term, a short analysis period will
be used.
The structural design period is a period during which it is predicted with a
high degree of certainty that no structural failure will occur and no structural
maintenance will be required.
If no rehabilitation is planned within the analysis period, the analysis and
design period are the same. There will, however, always be maintenance
Reseals using sprayed bituminous surface requirements, mostly resealing, within the structural design period, which,
treatments are not regarded as structural
maintenance.
if not carried out timeously or to the required quality, could result in premature
structural failure of the pavement.
These aspects are illustrated in Figure 2.1. The total life cycle cost of the
road over the analysis period is clearly a function of carrying out the correct
The total life cycle costs of these different maintenance at the correct time. Figure 2.1 (A) shows a design strategy that
options could differ significantly. Accurate cost allows for a resurfacing and structural rehabilitation, which if the resurfacing
comparisons using vehicle operating costs and
construction. Rehabilitation costs with pre- is neglected or delayed will result in premature failure. The strategy illustrated
vailing discount rates for each project are in Figure 2.1 (B) allows the riding quality of the road to deteriorate to a
therefore essential during the design stage. greater degree with time and relies on three resurfacings to ensure that
premature structural failure through water ingress does not occur.

(a) (b)
Figure 2.1 a + b. Illustration of various design strategies on maintenance and rehabilitation needs.
2.3 Typical pavement structures

A typical pavement cross-section is illustrated in Figure 2.2 where terms and
elements are described. The terminology used in this Guideline is defined in
Appendix J.
Centre
line
Cut back slope Roadway

Shoulder breakpoint
Cross section terms Shoulder Carriageway Shoulder
Open side drain Lane Lane Embankment side slope
Original ground level

Finished road level
Cutting Pavement layers Formation level

Cross section elements Improved
subgrade layers
Fill
In-situ subgrade
Roadbed
Subgrade
Figure 2.2 Cross section terms and elements.

The common bituminous pavement structures used in Botswana consist of
a combination of layers composed of the various materials (Table 2.1). It is,
however, clear that the pavement structure can comprise a wide number of
permutations. It should be noted that various modes of distress are particularly
associated with specific pavement structures, e.g. block cracking with
stabilised bases, shrinkage cracking of asphalts, etc.
Surfacing Asphalt, bituminous surface treatment
Crushed stone, natural gravel, stabilised

Base
aggregate or gravel
Crushed stone, natural gravel, stabilised aggregate or

Subbase
gravel
Selected layers Natural or stabilised gravel
Note: Stabilised material binders include pozzolans (lime, cement etc.) and bitumens
(penetration grades, emulsions, cutbacks etc.)
Table 2.1 Composition of typical Botswana bituminous pavements.
2.3.1 Surfacings
Various bituminous surfacings are available for use on roads (Figure 2.3).
The choice of the most appropriate surfacing type is usually based on the
life-cycle costs of the different surfacings, i.e. it is a function of their initial
construction cost and their effective service lives. However, aspects such as
the local maintenance capability, grade of the road, stiffness of the pave-
ment structure and traffic need to be considered during selection of the
surfacing type. A guide to typical surfacings and their expected lives is shown
in Table 2.2
SAND SEAL SINGLE OTTA

1 Prime 3 2 SEAL
1 2
2 Binder No Prime
3 Sand 1 Binder 1
2 Graded
aggregate
SINGLE CHIP
SEAL 3
1 Prime 2
2 Binder 1
3 Stone
DOUBLE OTTA
SEAL 2
DOUBLE CHIP 2.layer 1
No Prime
SEAL 1 Binder
1 Prime 5 1.layer 2
2 Graded 1
2 Binder 4 aggregate
3 Large stone 3
2
4 Binder 1
5 Small stone
ASPHALT
CAPE SEAL CONCRETE
1 Prime 4 1 Prime
2 Binder 3 2 Asphalt 2
3 Stone 2 Premix
1
4 Slurry 1
Fig 2.3 Schematic illustration of various types of bituminous surfacings.

Surfacing Poor conditions Good conditions
Asphalt (hot mix) 10 - 14 15 - 20
Double seal 6-8 9 - 13
Single seal 4-6 5-9
Double Otta seal 7 - 10 12 - 15
Single Otta seal +

7-9 9 - 11
sand seal
Double sand seal 5-7 7 - 11

Table 2.2 Typical surfacing lives in Botswana (years).
2.3.2 Pavement layers

The pavement layers provide the load-spreading capability of the pavement
and are constructed from selected processed, natural or stabilised materials.
The base, subbase and selected layers comprise the pavement layers and are
generally specified in terms of a strength requirement at a specified moi-
sture content and compaction effort. If all the layers are unstabilised, the
stiffness/strength will decrease systematically from the base to the lower
selected layer. If the subbase is stabilised, it is usually stiffer and stronger
than the base. Stabilised subbases are usually used to provide a sound working
platform when natural materials are poor (low CBR), and to ensure that the
compaction required for high quality bases is obtained. The problem of
Sealing and careful maintenance of stabili-
reflection cracking in the surfacing associated with stabilised bases has sation (reflection) cracks can negate the
resulted in a reduction in their use, but specific circumstances (usually the problem of water ingress into the pavement
lack of suitable base materials) may necessitate their use. and allow the use of stabilised bases.
2.3.3 Subgrade
The subgrade materials are those in situ. They are usually improved by rip-
ping and recompacting to provide a uniform layer, and reduce material
variation in strength, which is undesirable and may lead to local pavement
distress. Subgrades are normally classified in terms of CBR, and the overlying
pavement structure is designed to limit stresses (and deformations) in the
subgrade.
2.3.4 Sealed/unsealed shoulders

From both a safety and a structural point of view, the road shoulders adjacent
to the carriageway are extremely important. Gravel/unsealed shoulders are
usually less costly to construct but require material of a specific quality for Road with sealed shoulders.
optimal performance and must be maintained regularly. Poor material and
maintenance results in unsafe conditions, particularly when the drop-off
between the seal and the gravel becomes large. Unsealed shoulders, if poorly
maintained, can have a major influence on the pavement drainage as a res-
ult of ponding of water adjacent to the seal and subsequent seepage into the
pavement base. Sealed shoulders are initially more costly to construct but
have the advantage of reducing water access into the structural layers of the
pavement and providing safer conditions for the motorist. A frequent pro-
blem with sealed shoulders is that they are used as an extra lane by slow
moving vehicles and, if they are structurally inadequate (i.e. not built to the Road with unsealed shoulders.

same standard as the rest of the pavement), will require substantial
maintenance.
2.4 The environment

Environmental influences such as the climate, including rainfall and
temperature, in situ material (i.e. subgrade) thickness, nature and condition
and regional moisture and drainage conditions play an important part in the
performance of the road structure. The subgrade material must be brought
to a standard where it is uniform and has adequate strength to support the
expected traffic loading (see section 2.5). This is usually achieved by
mechanical improvement of the in situ material (ripping and recompaction)
or importation of a better quality material.
Grass and bushes should not be allowed to grow
close to the surfaced area, as this will have a
negative influence on the pavement performance. Climate has a significant effect on the road structure, particularly in terms
of the precipitation. High rainfall or periods of concentrated rainfall can
Perched water tables are caused by any low result in changes in the moisture content of the pavement and subgrade
permeability material (e.g. calcrete, silcrete, materials, which under poor drainage conditions can detrimentally affect
ferricrete, clay, etc.), beneath the natural ground
surface, which impedes drainage and results in a the pavement structure. Particular note needs to be taken of perched water
build-up of moisture in the material adjacent to the tables where moisture often moves beneath the road and leads to structural
road. Specific problems related to perched water deterioration of the pavement under traffic loading.
tables have been recorded on the Nata-Gweta
and the Tutume-Maitengwe roads.
2.5 Traffic loading

A pavement is designed to carry a pre-determined number of standard loaded
axles before structural rehabilitation is necessary. An increase in the traffic
as a result of higher growth rates than predicted or excessive diversion of
traffic from other routes shortens the life of the road. Of greater consequence,
however, is the effect of overloaded vehicles. Damage caused to pavements
increases exponentially with axle load. A 10 per cent axle overload can
therefore cause some 50 per cent more damage than the legal load, while a
20 per cent axle overload can more than double the damage that the legal
Overloaded vehicles are a major cause of pre- axle load would cause. Other factors such as the tyre configurations and
mature distress due to the exponential effect of pressures also have a major impact on the stress distribution on the pave-
load on pavement damage. Hence the pavement
life can be significantly reduced. ment and can result in localised distress.
Pavement distress, shown as excessive rutting.

3 TYPES OF PAVEMENT
DISTRESS
3.1 Causes and types of pavement
distress
The structural failure of a pavement is usually manifested as rutting, cracking,
ravelling and/or shear failure in the pavement. The process is often self-
perpetuating in that the development of distress allows the ingress of water
into the pavement exacerbating the conditions that may have initially led to
the cracking.
The need for a pavement evaluation is usually determined by the structural
performance of the pavement. Functional characteristics are generally
controlled by ongoing maintenance activities whereas structural defects
indicate potential structural problems and require more costly intervention.
The typical distress types associated with functional and structural
performance are summarised in Table 3.1.
granular
Performance Distress type granular
Functional Riding quality

Skid resistance riding quality
Surface drainage
Structural Deformation rut depth

Cracking
Surface distintegration cracking
Table 3.1 Typical distress types associated with performance. time and traffic
The distress or expected mode of failure of pavements is a function of the

type of pavement. This is primarily related to the subbase, base and surfacing granular
type. Stabilised materials behave in a different manner to granular materi- cemented
als, and asphaltic concrete behaves differently to thin bituminous seals.
pre-cracked post-cracked phase influence
Typical distress modes discussed are: phase of water
n granular layers: shear failure and deformation/rut; curvature
deflection
n stabilised layers: wider-spaced cracking (> 1 m), surface crushing;
deformation
n asphalt concrete: closer-spaced cracking (< 1 m), rutting.
traffic
cemented subbase indicators
Granular materials usually fail in shear because of their shear strength being
exceeded by the applied load. This is generally the result of excessive water The consequences of deformation occuring
in the pavement, which reduces the shear strength of the material. Granular are more severe than cracking, since the
useful life of the pavement can be consider-
materials are also prone to plastic deformation under load. This results in ably prolonged if cracks are attended to in
the gradual development of ruts, either due to compaction shown up by time and kept sealed.
traffic compaction or else by non-reversible plastic strain under load.
Pavements where only the base course is
Stabilised materials are initially far more rigid than granular materials and cemented with granular subbase are sensitive
fail in tension if overloaded with inadequate support. In time, stabilised to high axle loading and loss of strength by
ingress of water through surface cracks.
material cracks under loading or due to shrinkage and therefore behaves
more like a granular material. This need not necessarily result in any distress It is particularly important that the first scheduled
of the pavement. The shrinkage cracking associated with strongly stabilised reseal is not missed or deferred due to the
early development of block cracks in this
pavement type.

materials has led to a significant reduction in the use of stabilised materials
in the base course. Stabilised subbases are, however, very common as are
weakly stabilised selected layers to improve material quality in the lower
layers of pavements.
Thin bituminous seals are not considered to be structural layers within the
pavement but only waterproof the base and protect it from traffic action.
Distress of the surfacing usually reflects the performance of the underlying
road structure, particularly the top of the base to which the seal is bonded.
bituminous mix Asphalt concrete on the other hand, is considered as a structural pavement
granular or cemented
layer. It is a relatively stiff material, particularly under low temperature
conditions, and its performance is thus related to the deflection of the pave-
riding quality
ment. High deflections will result in fatigue failure and extensive cracking.
Environmental influences such as drying and ultra-violet radiation harden
rut depth the binder with time and thereby increase its stiffness, making it more
susceptible to cracking under traffic loading. High surface temperatures,
cracking
wheel loads and wheel contact pressures may also result in permanent
time and traffic deformation (rutting) of the asphalt mix.
Roads with high traffic speed require a levelling It is possible to obtain a preliminary indication of the possible causes of
layer made of bituminous mix in order to archieve distress in a pavement from a simple visual evaluation. A distinction must
good riding quality.
be made between traffic and non-traffic associated distress. Traffic associa-
ted distress is usually restricted to the wheel paths while non-traffic-asso-
cemented
ciated distress tends to occur over the full width of the road or occurs at the
cemented (preferred)
pavement edges and adjacent to culverts. Table 3.2 indicates possible causes
of traffic-associated distress.
riding quality
Distress type Possible cause
cracking Cracking :
Longitudinal Subgrade problems, poor construction.
rut depth
Crocodile Fatigue failure of surfacing or base.
time and traffic Poor drainage/high moisture content.
Poor bond under bituminous surfacing.
The use of stabilised layers in a pavement Insufficient compaction of subgrade.

Rutting
should be identified in any pavement evaluation Insufficient compaction of pavement layers.
as different behaviour can be expected. Insufficient pavement strength.
Shear failure of bituminous surfacing or granular
b a se .
Bleeding/flushing Excessive application of bitumen.

Soft base resulting in punching of aggregate.
Pumping Moisture/drainage problems with significant

movement in upper pavement layers.
Ravelling of surface Insufficient or dry binder.

Poor aggregate adhesion.
Surface potholes Poor bonding to base.

Disintegration of weak aggregate.
Salt damage.
Vehicular damage.
Spalling around cracks.
Unstable premix causes shoving. Deep potholes (into base) Insufficient shear strength of base or subbase.
Excessive moisture.
Table 3.2 Possible causes of traffic-associated distress.
Typical examples of non traffic-associated distress and their potential causes

are summarised in Table 3.3.

Distress type Possible cause
Longitudinal cracking Shrinkage through ageing.

Reflection of stabilisation cracks from base or
subbase.
Shrinkage of natural gravels through drying or self
cementation.
Volumetric movement beneath shoulder (expansive
clays).
Settlement of fill.
Transverse cracking
Shrinkage of asphalt through ageing.
subbase.
Volumetric shrinkage associated with leaking
culvert.
Shrinkage of natural gravels through drying or self
cementation.
Tearing by paver or steel wheeled rollers (asphalt).
Settlement at culvert or stuctures.
Block cracking Shrinkage through ageing.

subbase.
Map cracking Shrinkage through ageing of surfacing.
Soluble salt blistering.

Star cracking Water vapour blisters.
Chemical reactions.
Surface irregularities Collapsible sand.

Expansive soil.
Subsoil mole or insect activity.
Deep potholes (into base) Insufficient shear strength of base or subbase.

Excessive moisture.
Table 3.3 Possible causes of non traffic-associated distress.
3.2 Roughness
The roughness of a road is the longitudinal irregularity of the wheel paths of Significant increases in roughness are normally
the road that affects the riding quality and the dynamics of moving vehicles. associated with structural deterioration and
indicate the need for a more comprehensive
The road roughness also has a significant impact on the operating costs of evaluation.
vehicles making use of the road. Surface roughness, as well as rutting, can
also have a major influence on the drainage of water from the pavement
surface. Although primarily a functional characteristic of roads, the surface
roughness is often indicative of structural deficiencies.
3.3 Deformation
Rutting is the most common manifestation of deformation and indicates The shape of the rut gives a useful first indi-
plastic deformation of the pavement layers and/or subgrade under repeated cation of the pavement layers that are likely to
have caused the rutting. A wide, evenshaped rut
traffic loading. Rutting is restricted to wheel paths. is typically indicative of problems in the lower
layers and narrow more sharply defined ruts
Other typical forms of deformation include depressions and mounds, are usually caused by problems in the upper
displacements, corrugations and undulations, all manifested as deviations pavement layers.
of the road surface from a uniform flat condition and having a detrimental
effect on the riding quality. Deformation often results from subgrade and

construction deficiencies and is frequently associated with culverts and
embankments. Shoving of asphaltic concrete surfacings under traffic is also
a common cause of surface deformation.
3.4 Cracking
The onset of cracking is a primary cause of subsequent rapid deterioration
of a pavement, principally due to ingress of moisture and weakening of
pavement layers, which is why its identification and treatment is a key
component of any pavement evaluation. Different forms of cracking can be
Crocodile cracking and associated shoving. due to different fundamental causes, so it is vital in the visual evaluation to
identify the types of cracking. For practical purposes, four different forms
of cracking are defined, as indicated in the following sections.
I. Crocodile and Map cracking

Traffic-associated crocodile cracking (interconnected polygons less than 300
mm diameter) normally starts in the wheel paths as short longitudinal cracks.
With thin surfacings, this progresses through star and map cracking
Typical map cracking. (interconnected polygons more than 300 mm diameter) to crocodile cracking.
With thicker surfacings or asphalt bases, the cracks generally propagate
further along the wheel path before secondary cracks form, resulting in the
familiar map pattern.
II. Block cracking

Block cracking is usually caused by the shrinkage of stabilised layers. These
cracks are not confined to the wheel paths. The distress initially appears in
the form of longitudinal or transverse cracks and develops into block
Block cracks resulting from stabilised base. They cracking. The spacing between these block cracks largely depends on the
have been well sealed to prevent water ingress.
type of material, the thickness of the pavement layer and the shrinkage
characteristics of the stabilizing agent. Although traffic loading does not
initiate this distress, it tends to cause secondary cracking, which may
eventually become crocodile cracking.
III. Longitudinal cracking

Poor construction techniques often cause longitudinal cracking. These
include poor construction joints and segregation of materials during
Swelling of subgrades or settlement of embank-
construction. These problems are visually easily detectable and should be
ments may also cause longitudinal cracking. identified during the detailed visual inspection.
Such cracks are not confined to the wheel paths
and are often found near the shoulders of the
pavement.
IV. Transverse cracking
Transverse cracking is usually caused by temperature-associated distress.
This cracking is thus not usually directly associated with structural problems,
but it can initiate further deterioration if the ingress of surface water occurs.
Shrinkage cracking of the asphalt surfacing is indicative of asphalt that does
not respond easily to changes in temperature, i.e. mostly dry and stiff asp-
halt. Older pavements are usually more susceptible to this form of cracking.
The closing of these cracks in the wheel paths through the effect of trafficking
Transverse cracking is very common and is usually is an indication that the problem may be limited to the surface layer only.
caused by temperature-associated distress.

In some cases, it may be necessary to quantify the extent of the cracking but
this is not usually necessary for routine pavement evaluations. Quantification The visual assessment should classify observed
is most conveniently carried out by expressing the cracking as a percentage cracks in the four categories, crocodile/map,
of the surface area over a defined length of road or lane. Crack intensity is a block, longitudinal, transverse.
measure of the total length of cracking per unit area (m/m2). A method
currently in use is to quantify the cracked area as the sum of rectangular
areas surrounding individual cracking networks in square metres as a per-
centage of the subsection area. For linear cracks, the area is defined by a 0.5 In the BRMS cracking is classified as “all cracks”
and “wide cracks”. For pavement evaluation
metre wide strip extending the length of the crack1.. It is also possible to purposes, however, it is more important to
quantify movement associated with cracks using the Crack Activity Meter record the type of cracking with a classification
(CAM) but this is considered to be more of a research tool. of its severity and extent as discussed
previously.
3.5 Surfacing problems and surface

texture
Pavements may be structurally adequate but may have problems with the
surfacing. The visual survey must therefore identify surfacing problems,
and the later deflection and DCP data processing will clarify whether there
are underlying structural problems. Main surface defects to be identified are:
n disintegration or ravelling;
Loss of stone from single seal.
n bleeding or polishing.
Ravelling is manifested as loss of adhesion between the stone in the seal
and the binder. If this occurs because of water, it is defined as stripping.
Inspection of the loose material generated during ravelling will indicate which
process occurred. Stripped aggregate is clean while ravelled aggregate is
coated with binder. Disintegration of aggregate (as a result of inadequate
hardness or durability) is often manifested as ravelling.
Bleeding is a manifestation of excessive bitumen, either as a result of too
much being applied during construction or a change in the effective aggregate
gradation resulting from disintegration of the aggregate or embedment in
the base under traffic. Polishing of the aggregate is usually associated with Loss of surface texture as a result of bleeding (degree 5).
specific aggregate types that wear evenly under traffic to form a smooth
surface. The leads to a reduction in the skid-resistance of the pavement. For certain types of surfacing and traffic
These aspects primarily affect the functional behaviour of the road. conditions, the presence of bleeding may not
necessarily be a problem. In general, surfacings
The edges of cracks, specifically those associated with stabilised and self- that bleed will last longer than dry surfacings.
cementing materials, frequently spall under wheel loads. If left unchecked,
It should be noted that unmaintained ravelling
this spalling develops into potholes. and spalling would usually lead to water pene-
trating the base and the formation of potholes.
The texture of road surfacings primarily affects the functional performance A functional defect could thus lead to structural
of roads in terms of their skid resistance, road noise and wheel-spray. distress.
Cracking of the seal may result from structural characteristics e.g. high
deflections, or may be caused by environmental effects such as ageing
attributable to ultra-violet damage or drying out (these are usually exacerbated
by lack of maintenance). It is important to differentiate between these two
aspects and this is best done by determining whether the distress is restricted
to the wheel paths (structural problems) or whether it covers the entire road
surface (environmental or maintenance problems).
The texture is usually described as part of the visual evaluation (Appendix
A) but in certain instances, it may be necessary to quantify the surface friction.
Apparatus such as the Portable Skid-resistance Tester (sometimes called a Pendulum tester to determine skid resistance.

Skid resistance determination can be carried Pendulum Tester) can be used to quantify the pavement friction but this
out using commercial equipment e.g. SCRIM or
Grip-tester, but this is seldom cost-effective or
equipment only measures a very small area of road and is more appropriate
necessary for a limited pavement evaluation or to localised accident investigations and laboratory work.
where a resurfacing type remedial action is in
any case identified. The surface texture can be measured using the sand-patch method. This
type of testing is considered to be too specialised for routine pavement
evaluations and would only be carried out on a project basis where it has
been identified as a requirement.
3.6 Pumping
Pumping occurs when water pressures generated in the pavement by traffic
loading cause water containing fine material to be ejected through cracks. It
is usually manifested as discolouration of the surfacing associated with
The presence of pumping is a sign of serious defici-
encies in the base layer and normally signals the cracking and is indicative of a loss of fine material from the pavement.
onset of rapid deterioration. Pumping is most readily apparent just after rains, because the freshly pumped
fines lie along the cracks and have not been dispersed by traffic. At other
times, tell-tale discolouration around cracks can be a good indication of
previous pumping.
3.7 Potholes and patching

Potholes are the result of loss of material from the base course once the
surfacing has failed. Potholes are typically the result of pumping and loss of
surfacing attributable to lack of maintenance. Potholes are distinguished
from ravelling by being more than 150 mm in diameter and greater than 25
mm in depth. Patches indicate the location of potholes that have been
maintained and are not necessarily defects, but give an indication of previous
distress. Both potholes and patches are indicative of either material or
drainage problems in or beneath the pavement and are a strong warning of
Large patch but poorly executed resulting in impending structural deterioration.
failure.
3.8 Edge break

Edge breaks caused by the breaking away of the edge of the surfacing adjacent
to the shoulder under traffic is mostly the result of poor shoulder maintenance.
Edge break is more a functional problem than a structural problem but should
be included in routine visual evaluations. Localised incidences are often
associated with popular taxi pick-up/drop-off points. Extensive edge break
is, however, a good indicator that the pavement is too narrow for the traffic
being carried.
Bad edge break.
3.9 Moisture/drainage
The effect of excessive moisture in the pavement structure has been discussed
(Sec 2.4) but bears repeating as it is probably the most important contributor
to pavement failures. Excessive moisture in the pavement structure is gene-
rally the result of inadequate side-drainage, poor shoulders, unmaintained
surface seals or a combination of these. High moisture contents in the pave-

ment materials result in a decrease in the material strength and stiffness as
well as the development of excess pore-water pressures under load. Both of
these conditions can result in shear failure of the material at stresses lower
than those designed for.
It is generally accepted that the moisture content within a sealed pavement
structure reaches an equilibrium value some time after construction2. This
can, however, be influenced by the drainage and maintenance. Irrespective
of this, unless there is a well-maintained sealed shoulder, the moisture con-
Inadequate maintenance of side drains.
tent beneath the outer wheel path can be expected to fluctuate significantly
through the year. During any pavement evaluation, the expected pavement
moisture regime in relation to the season should be recorded, particularly
when DCP or deflection surveys are carried out.
Similarly, the rating of the drainage and its influence on the pavement moi-
sture regime must be carefully evaluated and recorded. Suggestions for this
are included in Appendix A.
During the rating of drainage particular attention should be paid to:
n condition of surface or sub-surface drains (if any); Water ponding at the edge of the surfacing will
n road camber; weaken the pavement structure.
n topography and geology of the surrounding area; Drainage inadequacy is often the cause of local-
ised areas of distress and the visual evaluation
n vegetation in the immediate area. should identify such inadequacies and possible
These factors will help establish specific remedial needs for the section in causes.
the subsequent data processing and section analysis.
3.10 Soluble salt damage

Problems caused by soluble salts in semi-arid and arid areas are such that
they deserve particular discussion. Soluble salt damage is manifested as
blistering of the bituminous seal, star cracking, separation of the seal from
the base course and eventually disintegration of the bituminous surfacing.
The problems arise from the solution of soluble salts in the pavement layers
or subgrade, migration through the pavement and crystallization beneath
the surfacing when the moisture carrying them evaporates at the surface.
Salt damage to road surfacing.
Close up of salt blisters.

4 PLANNING AND SELEC-
TION OF APPROPRIATE
TESTING
This chapter highlights the minimum requirements
for pavement evaluations.
4.1 Purpose of pavement evaluation

Prior to embarking on any pavement evaluation it is essential to understand
the objective of the evaluation and to make use of evaluation techniques that
will best achieve this objective. The primary objective of most evaluations
that will be carried out will be to provide the input for rehabilitation designs
(structural evaluation). Other objectives may include:
n evaluation of experimental road sections;
n ad hoc failure investigations, and
n post-construction pavement audits.
Planning of pavement evaluations must be aimed at satisfying the requirements
of the specific evaluation. In addition to this the needs for a structural
evaluation differ from those for a functional evaluation.
4.1.1 Structural evaluation

Predictions of pavement life are highly dependent Structural evaluations are carried out to determine whether the pavement
on accurate past traffic loadings. Wherever will carry the traffic it has been designed for and can be carried out at any
possible, therefore, the Roads Department should
maintain a reliable record of regular traffic counts time in the pavement’s life. The remaining structural capacity can be
and, ideally, vehicle axle loads for its road network. determined and compared with the traffic that the pavement has carried, or
is expected to carry over the remainder of its life.
Structural evaluations require an investigation of the strengths (and preferably
stiffnesses) and thicknesses of the individual pavement layers as well as the
overall interaction of the layers within the pavement structure. These
evaluations thus require sophisticated testing of parameters such as
deflections, deflection bowl parameters and detailed testing such as Dynamic
Cone Penetrometer (DCP) profiles or test pits. Visual evaluations of cracking,
disintegration and potholing are necessary inputs for a structural evaluation.
Visual inspection is of vital importance in addres-

sing the pavement condition. 4.1.2 Functional evaluation
Functional evaluations identify the capability of the pavement structure to
provide a comfortable and safe service to the road user. The primary parame-
ters determined in functional evaluations are the riding quality, skid resistance
and a visual evaluation of aspects such as potholes and edge break.
4.2 Existing information

Considerable amounts of information regarding any road project usually exist
within the Roads Department. This includes:
n the original design;
n the as-built records or completion data;

n traffic information;
n maintenance records;
n ad hoc reports, and of course
n data in the Road Management System.
Other information such as the climatic records over the life of the road can
be obtained from the relevant source if necessary. Evaluation of these data
is essential in order to plan the most effective evaluation procedure. Other
documents in this series of Guidelines could also provide useful informa-
tion.
4.2.1 As built data

The as-built and/or completion records should be compared with the origi-
nal pavement and material design to ensure that the pavement was built to
the original specification. Particular care should be paid to the layer
thicknesses, material properties and in situ densities or compaction. Traffic
information should be evaluated to investigate whether the traffic carried
on the road was according to the design assumptions or whether there has
been a significant change, especially in heavy or overloaded vehicles. A
comparison of the condition of the pavement in the two directions of the
road will often indicate whether loading has played a part in the pavement Different degrees of distress in different directions
condition, as one lane may carry more loaded vehicles than the return lane. of traffic.
4.2.2 Maintenance history

The type and frequency of maintenance carried out on the pavement is also
a good first indication of the possible causes of distress on a road. Highly
localised problems are often the result of poor pavement drainage or localised
material/construction deficiencies, while large areas of rutting or cracking
indicate excessive traffic and/or design or construction deficiencies. Similarly,
ad-hoc reports would normally be produced because of some form of pro-
blem. These would usually identify one or more probable causes of the pro-
blem. It is also important to talk to the local staff (regional engineers,
technicians and section heads), as valuable information may be available Localised structural problem.
which has not been reported or documented.
The need for a pavement evaluation will frequently be identified during Visual examination of the surfacing will usually
allow the identification of past resealing activities.
routine analysis of the road network using the Road Management System.
Typical maintenance actions in Botswana would
This information is invaluable for determining the primary modes of distress be a fog spray after 2 to 3 years, a surface
and possible causes of failure. treatment after 5 to 7 years and then a slurry
seal after 3 years.
4.2.3 Historical traffic

It is essential to evaluate all historical traffic records. These should include
not only the number of vehicles, but also a break down into various categories
(cars, BUSES and medium (MGV), heavy (HGV) and very heavy goods
vehicles (VHGV)) and, if possible, details on the actual axle loads and
distributions and growth rates. For studies of premature failures particularly,
this is essential in order to determine whether the road has been overloaded
either in terms of the cumulative equivalent standard axles or by occasional
excessively heavy axles.

4.2.4 Climatic records
Climatic records can be obtained from the The importance of moisture on the performance of pavement structures has
Meteorological Institute, Gaborone. been highlighted in this guideline. An evaluation of the rainfall records over
the life of the road will indicate whether there have been significant deviations
from the average quantity of precipitation in the area as well as information
on the distribution of the rainfall. Although the rainfall may have been similar
to the long-term annual average rainfall, the records may show that it was
concentrated over a short period resulting in abnormal moisture conditions
in the pavement.Records of precipitation that occurred during construction
can also provide insight into potential pavement problems. Periods of high
rainfall could have resulted in disruption of the construction process as well
as attempts to compact layer materials at excessive moisture contents.
4.3 Evaluation framework and

selection of tests
4.3.1 Evaluation framework
Pavement evaluation can be expensive and the A wide range of techniques and tools is available for the evaluation of
correct approaches and techniques must be pavements. Not all of these are always available at a reasonable cost and
carefully selected to optimise cost-effectiveness.
many are unnecessary for certain evaluations.
Table 4.1 below summarises the evaluation need for typical pavement
evaluation projects as well as the appropriate tools and their availability.
Many of the evaluation techniques require tools available within the
department but the availability of some of the sophisticated testing equipment
is limited.
Evaluation technique When required Tools Availability
As-built data ) - )
Maintenance history ) - )
Historical traffic ) All evaluations - ) In house
Visuals ) - )
Rutting ) Straight edge )
Deflection/curvature All structural evaluations* Benkelman Beam, In house/limited

Deflectograph, FWD
DCP All structural evaluations DCP In house
Roughness All evaluations (usually Merlin, LDI, BI, HSP In house/limited

available from the BRMS)
Surface texture Where functional aspects SCRIM, Griptester, In house/limited

are evaluated Pendulum tester
Test pits and material Where the mode of distress Field test equipment,
In house
testing needs to be confirmed nuclear density meter
RMS data is obtained on a network level and is * Judicious interpretation of deflection/curvature results is required for pavements with
not normally sufficiently detailed for project stabilised layers
level evaluations. Table 4.1 Typical pavement evaluation data requirements.
As-built, historical, visual, rutting and roughness data fulfil the minimum
need. These are usually the data types in a road management system (RMS)
that first identify a possible need for a more comprehensive evaluation. For
post-construction audits carried out prior to the development of any distress,

the investigation would need to be targeted at the specific reason for the
audit. More details on the individual tools are provided in the following
section.
A pavement evaluation will normally consist of an initial assessment followed
by a more detailed evaluation. The process is iterative and requires more
sophisticated techniques as the required degree of knowledge increases.
4.3.2 Initial assessment

The initial assessment will usually be based on an evaluation of significant Structural deterioration will not always be signal-
changes in routine RMS survey data (e.g. visual distress, riding quality and led immediately from RMS survey data that is
primarily related to functional performance.
rut depth) indicating onset of pavement deterioration. This should be Significant changes in functional performance
complemented by a study of the as-built records and discussions with the do, however, indicate possible structural
local engineer. A drive-over visual survey is essential at this point, but for deterioration.
logistical and cost reasons, could be carried out by local staff. The informa-
tion available at this stage should be sufficient to identify whether a more
detailed evaluation is warranted.
4.3.3 Detailed assessment

The more detailed assessment will invariably include a comprehensive visual
survey, rut depth measurements, a deflection survey, a Dynamic Cone Pene-
trometer (DCP) survey and test pitting and sampling. Laboratory testing of
the samples collected will then be necessary.
4.4 Test frequency

Testing and sample collection is the most costly component of the field
evaluation. It is essential to develop the pavement evaluation programme to
obtain all the information necessary to carry out the rehabilitation design (or
any other purpose for which the evaluation is being carried out) with the
Field testing and sampling is costly and should be
minimum amount of data collection and field and laboratory testing. A carefully planned and executed.
balance needs to be struck between the cost of the evaluation and the amount/ As the number of tests increases, the cost of
quality of information required. testing increases but the incremental improve-
ment in the quantity and quality of information
The pavement evaluation should include a minimum number of samples/ decreases rapidly (concept of diminishing
readings per length of road as well as a minimum number per uniform section. returns).
The variability of the pavement structure and condition as well as the length
of the project will normally dictate the number of sample/readings - short
projects will necessarily have more observations per kilometre on average
than long projects.
Numerous investigations have been carried out into the suggested mini-
mum number of samples or observations required in pavement evaluations.
This guideline includes recommendations for the various measurement
techniques.
The most important consideration in undertaking field evaluations is theref-
ore to optimise the survey needs to limit the number of separate site visits:
ideally all field data should be gathered at the same time, e.g:
n deflections;
n visual data;
DCP testing, a useful tool in pavement evaluation.

n rut depth measurements;
n DCP data, and
n any specific test pitting and materials sampling.
This assumes that any roughness data has been collected separately, usually
as part of the pavement management process.
An experienced engineer/ senior technician should lead the survey team
and be on site during the fieldwork to ensure that all the important data are
recorded, and to instigate changes in the field sampling frequency if site
conditions require.
The team leader should be experienced in pave- For example, it may be evident directly from the site visit that certain sections
ment evaluations to enable important decisions are reasonably uniform. Since the primary objective of the field evaluation
to be made on site.
is to identify different uniform sections, and obtain sufficient data to classify
them, the team leader may therefore judge that sampling frequency may be
decreased where there is little to be gained from the additional sampling.
Similarly, there may be sections that require more comprehensive testing
for proper classification. This flexibility can ensure that the site evaluation
is properly optimised to obtain the necessary data in the most cost-effective
way.
The exact method of obtaining field data will depend on the field team
components. One practical approach is for:
n the DCP team to also take rut measurements and record visual data;
Measurements of roughness using a MERLIN
apparatus. n the deflection team (if a Benkelman beam is used) can later under
take test pitting and material sampling. This is because the deflection
survey team will normally progress much faster than the DCP team.
n the team leader should generally stay with the DCP team, taking
responsibility for the visual evaluation, and occasionally checking the
progress of the deflection team.
Whenever possible the team leader should Experience has shown, however, that the visual and rutting surveys are best
have any construction or as-built data on site, done first as an independent exercise. The deflections can be done at a
especially to identify where sections change,
and to assist in the task of optimising the similar time and then, with this information the DCP and test pits can be
survey needs. located. This optimises the time spent on DCP and test pits. This is obviously
the «ideal» situation, and the former method may be more practicable.
Critical to every survey is the need to tie-in the data with fixed physical
features. Milestones (kilometre posts) are essential to note, as these will
usually provide the main means of reference when later processing all the
data. Other significant features will be:
n town boundaries;
n junctions;
n rivers/bridges, and if possible;
Milestones are useful features as the main means
of reference. n culverts.
Sections in cut or on fill should also be demarcated, as these can have an
influence on the pavement behaviour and condition.
Nominal frequency of testing for the various main field data parameters is
suggested in Table 4.2 below. This assumes that the survey length is at least
ten kilometres but, in line with the foregoing, the actual frequency selected
should be adapted to suit the survey and may be varied in the field as deemed

necessary. Specific aspects related to individual evaluation techniques are
discussed in subsequent sections where relevant.
Measurement type Frequency Comment
Deflection (Benkelman 100 metres This is normally sufficient. For areas of

beam, Falling Weight localised distress, more frequent read-
Deflectometer) ings would be appropriate to better
delineate the section. The Deflectograph
gives a far more comprehensive survey
automatically (one reading every 6
metres).
Dynamic Cone 500 metres More frequent sampling may be needed

Penetrometer if there is reason to suspect that cond-
itions vary significantly.
Rut depth 50 metres More frequent sampling may be needed

(straight edge) if there is reason to suspect that condit-
ions vary significantly. Rut depths should
be taken at least at every point where
deflections are made.
Visual evaluation Effectively Data forms allow comprehensive rec-

continuous ording of distress type, degree and
extent. Photographs should be used
sparingly but effectively to supplement
the visual survey.
Test pits and material 2 or 3 per 10 Location and need should be identified
sampling kilometres during the field survey. Dependent on
number of uniform sections.
Table 4.2 Suggested minimum testing and sampling frequencies.

PART B EVALUATION
METHODS
5 Visual evaluation
6 Rut depth measurement
7 Roughness
8 Pavement strength
9 Test pit profiling and sampling
1 Introduction
3 Types of pavement
distress

5 Visual evaluation

methods
8 Pavement strength

sampling

data analysis
Part D References
Appendices

5 VISUAL EVALUATION
5.1 Methodology
The visual evaluation is one of the most important aspects of the pavement
evaluation process. This usually involves a comprehensive survey of the
road while a variety of parameters is visually evaluated in detail. The
conventional means of collecting visual evaluation data for road manage-
ment systems is from a moving vehicle (usually not exceeding 20 km/h)
with periodic stops to get information that is more detailed. Visual assess-
ment for pavement evaluation purposes, however, will normally be
determined at walking pace. If budget and time constraints exist a method Visual inspection of the pavement is one of the
of representative sampling may be used. This will require a prior subdivision most important aspects.
of the road into sections of similar visual condition.
Visual evaluation generally requires a rating of degree and extent of the
various distress parameters by an experienced person or preferably a team
of two or more. It is essential that all raters use a standard rating system. The
raters must also be calibrated against each other periodically for consistency.
Typical pavement conditions visually evaluated include:
n Cracking;
n Ravelling;
n Failures/potholing;
By carrying out visual inspection immediately
n Patching; after rainfall, pavement defects such as rutting/
depressions clearly show up.
n Pumping;
n Deformation;
n Bleeding (flushing);
n Surfacing condition;
n Drainage, and
n Edge-break.
In most cases a nominal rating system can be used although, in certain The visual assessment provides the primary
instances, quantification of the various types of distress should be carried means of interpreting other data parameters and
identifying possible mechanisms of distress.
out.
It is also important that during the visual evaluation, variations in the width of Methods of rating roads have been described
the road are noted. This is necessary for costing any proposed improvements in a number of publications of which TMH 93
appears to be the most suitable for use in
as well as being useful indicators of problems arising from inadequate width, Botswana.
i.e. edge breaking, shoulder problems, outer wheel path failures, etc.
5.2 Data collection

Many techniques and methods are used for visual evaluations, each being
developed and adapted to the needs and methods of the evaluator and the
project. In this guideline, two techniques that have been used successfully in
Botswana and in other areas of southern Africa are discussed.
Excessive shoulder drop off.

1. Degree and extent rating technique: In this technique the exact
location within the assessment section (usually 0.5 or 1.0 km) is not
recorded, only the extent of the occurrence of the distress is summari-
sed in an extent rating together with the representative degree
(severity) rating.
This technique is usually applied to the initial assessment, which is
done from a slow moving vehicle. The results are used in the initial
definition of uniform sections.
2. Degree rating on short segments: In this technique the severity re-

presentative of the distress is recorded over a shorter length (100m or
less).
The second technique is usually applied to the detailed assessment, which is

done by walking. The assessment provides the actual position of distress
items within an accuracy that is related to the length of the segments.
However, the first technique is usually sufficient for typical pavement
evaluations.
Where appropriate, consideration should also be given to the possible use
of the data in HDM-3 or 4 project evaluations. The data format should then
be compatible with format as described in the BRMS assessment manual.
Methods of rating roads have been described in The rating of degree and extent of the various distress parameters must be
a number of publications including TMH 9, TRH 6, carried out by an experienced evaluator following a standard procedure.
TRH 2, Botswana Road Management System
Manual, and Overseas Road Note 18 3,4,5,6,7. This is usually carried out per one km (initial assessment) or 100-metre
section (detailed assessment) but specific localised conditions should
be accurately located and measured (usually paced out or measured by
measuring wheel).
It is recommended that the degree and extent of the distress modes be rated
using five point scales as set out in Table 5.1.
Degree Description
1 Distress difficult to discern visually (sound)

.
2 Easily discernible distress but of little consequence
Degree 3 cracking. 3 Distress is notable with respect to possible consequences (warning)
4 Distress important with respect to possible consequences
5 Distress extreme with respect to possible consequences (servere)
Table 5.1 Degree and description of distress modes.
If no distress is present or discernible, no rating should be given.

Analysis of the pavement evaluation data for rehabilitation purposes should
make use of only degrees 1, 3 and 5 (sound, warning and severe respectively).
Degree 5 cracking. In these cases both degrees 1 and 2 would be considered as sound, degrees
3 and 4 as warning and degree 5 as severe.
Degrees 2 and 4 should only be used when The extent of the distress must also be rated. This is frequently done on a
detailed pavement failure studies are being five-point scale of areal extent (20 per cent intervals) but problems are often
carried out.
encountered when various degrees of distress are observed within a section,

the higher degrees being significantly less in extent than the general degree If necessary, more than one combination of
of distress. It is thus recommended that the following rating system be used descriptors can be provided, e.g. degree 5/
extent 1 and degree 3/extent 5.
for extent (Table 5.2):
Reference should be made to TMH 93 or the
Extent Description BRMS6 Raters Manual for additional detail
regarding visual condition rating.
1 Isolated occurrence, not representative of the section being evaluated
2 Intermittent (scattered) occurrence, over parts of the section length
Intermittent occurrence over most of the section being evaluated or

3
extensive occurrence over a limited portion of the section
4 Frequent occurrence over a major portion of the section
5 Extensive occurrence
Table 5.2 Description of extent of distress. Degree 3 pumping.

The combination of the extent and the degree should best describe the ge-
neral condition of the pavement. For general rehabilitation studies, a three-
point scale should be employed using isolated, scattered and extensive.
Field data collection should be carried out using standard forms to ensure
that all the necessary information is collected. Examples of form that can be
utilised are included in Appendix A. A summary of the descriptive definitions
for each visual parameter is also included in Appendix A. A copy of this
should be attached to the clipboard for field use.
Degree 5 pumping resulting in potholes.
5.3 Analysis and presentation

The visual evaluations can be used alone, or in combination with other
evaluation parameters such as riding quality and rut depths. As the data
consist of various parameters and include severity and extent, they are most
easily evaluated in terms of a single parameter. A Visual or Pavement Con-
dition Index can be developed from the data. This expresses the condition
of the pavement at a point or over a designated length of road as a single
value. Many methods of doing this exist and most Pavement Management
Systems include algorithms for calculating an index. Many of these are highly
complex involving the degrees and extent of a number of distress parame-
ters with various weightings of their importance. The proposed approach
that is widely used in southern Africa can be calculated in the visual condit- The method described in TRH 225 and used
ion data spreadsheet using the models provided in Appendix B. previously in Botswana is recommended.

6 RUT DEPTH
MEASUREMENT
6.1 General
Over the years, various lengths of straight The measurement of rut depths is best carried out using a standard straight
edges as well as string stretched across the edge and a calibrated wedge. On the basis of its general use and its ease of
road have been used for rut depth measure-
ment. The rut depths obtained using the handling and transportation, a 2-metre straight edge is the most practical.
different techniques can vary considerably
depending on the mode of rutting, wheel path
Automated rut depth measurement devices are also available, usually towed
width, carriageway width and general shape behind or a component of some other pavement evaluation device e.g. high-
of the road. This aspect requires standard- speed profilometers. These usually equate with a two-metre straight edge
isation if results are to be comparable.
or, in the case, of high-speed profilometers, can usually be set to equate with
any desired straight edge length. It is important to specify the requirement
to the service provider.
6.2 Test procedures

It is recommended that a standard 2-metre straight edge with a calibrated
wedge be used for rut depth measurements. Where no 2-metre straight edge
HSP showing rut measurement equipment at rear. is available, it is possible that a 2-metre length of string be substituted. The
string technique is not generally recommended, however, as it requires con-
The RMS would normally include rut depth
siderable practice to avoid over-reading.
measurements of the road network determined
every 2 or 3 years using a profilometer. This Normally it is sufficient to measure the maximum rut in the outside wheel
data should not be neglected during pavement
evaluations. paths, which in most cases have the greater rut. On most roads, it is usually
found that the heaviest trafficking occurs in one direction but ruts should
nevertheless be recorded for both directions. For single carriageway two
lane roads, the remedial action will be governed by the worst case, while for
dual carriageways it may be possible to have different remedial actions for
each carriageway.
Generally, for the delineation of uniform sections where absolute values are
not essential, other methods of rut depth measurement may be used, even
though there may be slight differences in actual recorded values due to the
method adopted. In terms of recognised terminal conditions, the 2-metre
straight edge should, however, be employed.
A rut measurement team in action. Note the flag
man for safety reasons.
6.3 Test frequency

It is recommended that a spare 2-metre straight
edge is available during rut depth measure-
ments in case of accidental damage to the one
in use by vehicles.
It is generally recommended that a minimum testing frequency for rut depths
is every 50 metres although it may often be necessary to carry out rut depth
measurements at intervals as low as 20 metres depending on the variability
and the length of the expected uniform sections. It must be ensured that the
number of measurements is adequate to provide a statistically acceptable
value per uniform section, i.e. at least 20 readings per wheel path per uni-
form section.

No specific further data processing is necessary for the rut depth mea-
surements at this stage. However, the data from different wheel paths must
be kept separate and no attempt made to combine or statistically simplify
them. The cumulative sum («Cusum») approach can be used to identify
preliminary uniform sections using the rut depth data (see Appendix H).
Where mechanical surveys have been carried out, the 95th percentile value
of the rut depth per 100 metre section can be evaluated using the “Cusum”
method. Simple relationships between rut depth and other parameters, e.g.
deflections, can be investigated at this stage. The relationship between the Shoving due to unstable base course layer.
rut depth and deflection can also be used to estimate the maximum deflection
that will result in a specified terminal rut depth under the past accumulated
traffic (Figure 6.1) as well as overstressed subgrades. Trends where the rut
depth increases with deflection (e.g. Figure 6.1) indicate that overstressing
of the subgrade was probable8. Ideally, specific relationships should be
developed for generic pavement structures in Botswana.
40
30
The presence of significant rutting with the rut
Rut depth (mm)
depth increasing with increased deflection is

95% Confidence limit
20
indicative that the pavement structure does not
protect the subgrade.
Terminal rut depth (15 mm)
Regression line
10
1,32 mm
0
0,5 1,0 1,5 2,0
Deflection (mm)
Figure 6.1 Example of relationship between deflection and rut depth.

7 ROUGHNESS
MEASUREMENT
7.1 General
Roughness measurements can be carried out Most pavement defects contribute in some way to increasing the roughness
using various types of equipment and apparatus, of the road pavement, although in its early stages cracking may cause little
each having its own advantages and disadvan-
tages. or no change. However, without proper maintenance, the cracked surfacing
deteriorates and the resulting potholes and subsequent patching will cause
an increase in roughness. The roughness of roads with similar pavement
construction is therefore a good measure of their relative pavement condit-
ion, but does not identify the nature of the failures or their causes.
Roughness measurements may make use of various principles.
Absolute profile – profile elevation relative to a true horizontal datum e.g.,
rod and level, dipstick, straight edge.
Moving datum profile – measures deviations of the profile relative to a da-
tum moved along the road e.g. rolling straight edge, profilographs, MER-
LIN. The use of a rolling straight edge would not normally be recommended
for pavement evaluation purposes, as the results can usually not be interpreted
in terms of a standard roughness scale such as the IRI.
Vehicle motion instruments – either measure displacement between axle
and body of a vehicle (e.g. Bump Integrator and Linear Displacement
Integrator (LDI)) or measure accelerations of axle or vehicle body using an
accelerometer e.g. Automatic Road Analyser. Details of this type of equipment
are provided in Appendix C.
HSP showing laser based riding quality Dynamic profile instruments – these instruments measure elevations
measurement apparatus. electronically relative to an artificial horizontal datum e.g. high-speed
profilometers.
Roughness is obviously speed dependent and each of these techniques the-
refore has unique differences. The International Roughness Index (IRI), the
most widely used parameter for roughness measurement, is a mathematical
statistic of the longitudinal profile obtained from the individual outputs of
the measurement techniques. Approximate comparisons of the IRI values
with QI, BI and PSI are shown in Table 7.1.

IRI QI BI PSI
IRI QI BI IRI
0 0 0 5 (m/km IRI) (count/km) (mm/km)
PSI
(m/km IRI)
0 0 0 0
0 5
1000 4,0
2 18 1369 3,5 20
2 2
40 2000 3,0
4 46 2976 2,4 4
60 4
4000 2,0
80 1,5
6 74 4687 1,7 6 6
100 6000
1,0
8 102 6468 1,2 8
8000
8
0,5
140 10
10
10 130 8305 0,8 10 000
12 12
180
12 158 10186 0,6 12 000
14 14
14 000
14 186 12106
16 220
16
16 000
16 214 14059
18 18
18 242 16042
Comparison of roughness scales.
20 270 18051
Table 7.1 Approximate comparisons between IRI, QI, BI and PSI (after
Paterson, 19871).
Use has been made in Botswana of various straight edges but a MERLIN9
and a number of Rolling Straight Edges10 have been used more frequently.
High-speed profilometers are used for the RMS.
7.2 Test procedures

The most common moving vehicle roughness units used in Botswana are
the International Roughness Index (IRI), the Quartercar Index (QI), Bump
Integrator and the Present Serviceability Index (PSI). It is recommended
that the IRI be utilised routinely as this is the measure employed in the
BRMS.
The method of roughness measurement is also not usually critical if the data
values can be reliably converted to IRI (now preferred as an international
standard) or QI (tending to be superseded by IRI). The most important
factor is for the data to be tied in with fixed physical features, especially
milestones (kilometre posts), as there are likely to be differences between
vehicle distance measurement readings and project road distances. Such
differences will be accommodated during the first part of data processing to
ensure all field data can be directly correlated based on the fixed field features.
The MERLIN should be used for routine evaluations of short pavement
sections, probably less than 10 km. The time and effort required for sections
longer than this probably make vehicle-mounted measurement devices more
cost-effective and practical. Additional details on the use of the MERLIN
are included in Appendix C. The MERLIN apparatus, in operation.

7.3 Test frequency
The test frequency is a function of the method used. Most roughness
measurement devices have fixed measurement intervals. The measurement
interval for the MERLIN is based on a whole number of revolutions of the
wheel such that 200 measurements are made within the test section. BI and
LDI usually average the readings collected every 100, 500 or 1000 metres
and this is fixed for that equipment. Essentially an averaged continuous
reading is obtained for roughness measurements.

Depending on the method of measuring the roughness, raw data may be in
Careful and regular calibration of any response the spreadsheets. These data will need to be converted to IRI units using the
type measuring system is essential to ensure
that the IRI values calculated are repeatable
latest calibration models for the specific vehicle or method used to obtain
and accurate. the data. The data can be used to determine excessive vehicle operating
costs (VOC) due to high roughness, e.g. in the HDM-3 or 4 models. As for
the rut depth evaluation, relationships between the road roughness and rut
depth can be used as a simple indication of overstressing of the subgrade8.
Additional information on the analysis of the roughness measurements is
provided in Chapter 10.

8 PAVEMENT STRENGTH
8.1 General
The relative strength of a given pavement is inversely proportional to its
deflection. Knowledge of the peak or maximum deflection in itself is
insufficient, however, to determine the pavement strength, as this would
depend on the specific pavement structure. In situ measurements of the actual
strengths of the individual layers are necessary to quantify the pavement
structure fully.
Traditional in situ strength testing has mostly been carried out with the in situ
CBR test or the plate-loading test. Both of these tests are destructive in that
the layer to be tested needs to be exposed. A plunger with the same dimensions
as the laboratory CBR test is forced into the layer for the in situ CBR test
and a circular steel plate is loaded for the plate-loading test. Both tests require
vehicles or large masses as reactions (kentledge) for the applied loads and are
time consuming but provide results that can be related to the design process.
The CBR result is closely comparable with the laboratory CBR test result
even though the grading is likely to be different (the maximum particle size in
the laboratory test is 19 mm) while the plate load test provides an estimate of
the elastic modulus of the layer material. It should also be noted that, unless
otherwise treated, the material is tested at an in situ moisture condition that
may differ considerably from the laboratory soaked CBR normally used for
the pavement design. The soaked condition can dramatically influence the The DCP apparatus, in operation.
CBR result.
The Dynamic Cone Penetrometer (DCP) can be used to estimate an Like the in situ CBR and plate load tests, the
strength/modulus determined with the DCP is
approximate CBR strength of the individual layers in a pavement structure that at in situ moisture and density conditions
rapidly with minimal disturbance to the pavement. In this way, many results and is thus not directly comparable with the
can be obtained quickly and the statistical evaluation of these will generally soaked laboratory results that are determined
provide a more representative measurement of the pavement strength than a at a prescribed density.
limited number of in situ CBR or plate loading tests.

The nominal relative strength of a pavement can also be calculated in terms
of the AASHTO structural number11. This method uses the sum of the sum-
med products of the thickness and strength coefficient for each layer to prov-
ide a discrete number for any pavement profile. Hodges et al12 modified the
original structural number by including a subgrade contribution, which is
necessary when evaluating pavements on strong subgrades such as those
commonly encountered in Botswana.
The modified structural number (SNC) is calculated as1:
SNC = 0.04 Saihi + 3.51 log10CBR – 0.85 (log10CBR)2 – 1.43 For soft subgrades the AASHTO structural
number SN can be calculated using SN =
Where 0.04(a1h1 + a2h2 + a3h3+ ….+ anhn) where the
layer thickness h is in mm.
ai = material and layer strength coefficients (Appendix D)
hi = layer thickness (mm)
CBR = in situ CBR strength of subgrade (%).
Relationships have been developed between the peak deflection results Relationships similar to these will be offered as
obtained from FWD testing and the SN value13,14,15,16. an optional method of measuring the structural
number in HDM-4. Experience in Botswana has
shown that strong subgrades influence the
correlation between SN and deflection.

Deflection bowls can be used to calculate the stiffnesses of the various
layers within a pavement. Complex back-calculation methods are necessary
and the computation of E-modulus values in this way requires a good
understanding of the pavement behaviour, the layer materials and thicknesses
in order to avoid spurious results. The use of DCP data to assist with the
estimation of values for the material stiffness and layer thicknesses is of
great value.
8.2 Deflection measurements

The timing of deflection measurements is very As a loaded vehicle moves over a road, the pavement surface deflects
important. Experience in Botswana has shown downwards before returning to its original, or close to its original, position.
that the deflection in the dry season can be as
low as 65 per cent of that in the wet season. This is termed the deflection and is a function of the magnitude of the load
Deflection measurements should therefore applied, the pavement structure (layer thicknesses and strengths) and the
preferably be carried out in the wet season i.e. subgrade properties. The performance of bituminous surfacings is closely
between March and May.
related to the deflection in a pavement as repeated flexing results in fatigue
failure of the surfacing and the development of cracking. Not all of the
deflection is recovered after each load application and, with time, this
accumulates to produce permanent deformation.
TRL have very rarely found ‘traditional’ fatigue Some controversy still exists about whether cracking of asphalt is generated
cracking (i.e. initiating at the bottom of the layer)
in asphalt surfacings, but rather cracking from
from below or from above. Cracking from the top would explain why there
the top. is often a delay between the first appearance of cracking and development
of roughness - water finally entering the pavement after the crack develops
to the full depth of the asphalt layer.
Deflections are inversely proportional to the
relative strength of a given pavement, but ad-
ditional information such as radius of curvature 8.2.1 Types of equipment
and layer strengths is needed to quantify the
structure fully. Deflection measurements with a Benkelman Beam use a standard wheel
load although this varies from country to country. This equipment is usually
Deflectograph and FWD equipment are not very available within most Road Departments. The Deflectograph uses a stan-
common and the owner of the equipment would dard 80 kN axle load (40 kN dual wheel load) whilst the Falling Weight
carry out this specialised testing.
Deflectometer (FWD) can make use of variable loads.
An important component of deflection measurements is the degree of
curvature of the deflection bowl. This gives an indication of the stiffness of
the upper portion of the pavement structure. The radius of curvature can be
determined (using the Dehlen Curvature Meter) during Benkelman Beam
testing or can be calculated from deflection bowl measurements e.g. from
the Deflectograph or FWD.
Deflection bowl parameters from FWD tests include Base Layer Index (BLI),
Deflection beam testing.
Middle Layer Index (MLI) and Lower Layer Index (LLI) as shown in Ap-
pendix E, Figure E4. These give an indication of relative contributions to
More detail regarding calculation of deflection overall pavement strength from the surfacing and base (to 300mm depth
bowl parameters is provided in TRH 128. approximately), subbase and upper selected layers (300 to 600mm depth
approximately), and selected layer and subgrade (600 to 900mm depth
approximately) respectively. They are obtained from the nominal deflection
bowl, and can provide some guidance on where weak layers occur.
A number of instances have been observed where high peak deflections are
measured but DCP and test pit information has shown no weak layers in the
top 800 mm of the pavement. The inference is that weak layers occurring at
depths beneath 800 mm can affect the deflection measurement significantly.
This can usually be confirmed by evaluating the radius of curvature – weak
Deflectograph. layers at depth should be manifested by high radius of curvature results.

8.2.2 Test procedures
The strength of a road pavement is inversely related to its maximum vertical
deflection under a known dynamic load. Deflections measured under a slowly
moving wheel load, therefore, can be a good indicator of the overall strength
of a pavement. These measurements are very useful in helping to identify
the cause and extent of any differential performance along a road and,
subsequently, for the design of any maintenance or rehabilitation measures
required for different sections of the road.
The main methods of deflection measurement are: Any deflection readings must be adjusted to a
nominal 40kN wheel load, as this is regarded
n Benkelman Beam; as a standard test load in southern Africa.
n Deflectograph, and;
n Falling Weight Deflectometer (FWD).
The first two methods determine deflections under a slow moving vehicle
wheel load. The FWD gives a response to an impulse load of a falling
weight. Both latter methods also provide more information regarding the
shape of the deflection bowl. The absolute values obtained using each method
may differ significantly. It is therefore recommended that deflection values
be treated on face value, without attempting to convert to a different deflection
standard, as each method will certainly allow differentiation of structurally
dissimilar sections. The only proviso is that any deflection beam readings
be adjusted linearly to a nominal 40kN wheel load (80kN axle load) if Falling Weight Deflectometer showing geophones.
necessary. This processing will be done during the first stage of data
validation.
Both transient and rebound deflections can be measured with the Benkelman It should be noted that each method gives
Beam. In one case the loaded wheel moves towards the beam tip (transient) different specific deflection values, which can
be broadly correlated. However, the correlation
while in the other the wheel starts at the beam tip and moves away (rebound). will differ according to pavement structure, and
Slightly different responses are recorded but, in practice, the rebound method applied loading.
is far easier and quicker to apply. Detailed information on Benkelman Beam
and Falling Weight Deflectometer testing is included in Appendix E.
8.2.3 Test frequency

It is important to acquire sufficient data to yield statistically acceptable results Deflection testing can be either labour intensive
and a minimum of 20 results per uniform section should be obtained. For (Benkelman Beam) or capital intensive (FWD
and Deflectograph) and the cost of deflection
long uniform sections, testing at a maximum of 100 metre intervals should testing is thus high. The number of tests carried
be carried out. The Deflectograph has the advantage that testing is carried out is therefore critical to the cost of the
out on an almost continuous basis with readings taken at about 6 metre investigation.
intervals.
8.2.4 Analysis and presentation

Apart from correcting the data for any discrepancies in distance measurement Temperature corrections are usually not
necessary for the typical pavements evaluated
or wheel load, no specific further data processing is recommended for
in Botswana, as they generally do not have thick
deflection measurements. If accurate data is available, moisture correction of asphalt layers (see Appendix E, Sec 3.6.1).
deflection measurements can be considered. The data from different wheel
paths must be kept separate and no attempt must be made to combine them.

8.3 Dynamic Cone Penetrometer
(DCP)
8.3.1 General
In most cases, the pavement strength will be determined using a DCP as this
is a simple, almost non-destructive test that gives valuable information.
8.3.2 Test procedures

The standard DCP apparatus can be adapted The normal procedures regarding use of the instrument should be followed
to use disposable cones, which are left in the with particular attention being paid to:
hole on completion of the test. These have the
advantage of: n the condition of the cone (not worn or bent);
- Reducing damage to the equipment during
removal after testing; n the device being held vertically throughout the test;
- Ensuring that the tip of the cone is always
in good condition. n large stones that could affect the readings;
n the hammer just touching the upper stop prior to release;
n a minimum depth of 800 mm is achieved;
n the DCP holes shall be backfilled with fine-dry sand, and the upper 50
mm of the hole shall be patched with asphalt premix.
8.3.3 Test frequency

It is recommended that at least one DCP profile is obtained in each wheel
path every 500 metres. However, in order to achieve meaningful evaluation
at least 20 profiles per uniform section should be obtained. As noted in
Table 4.2, the actual frequency should be based on the observed condition
of the pavement.
1000
8.3.4 Analysis and presentation

100
The DCP data is usually evaluated in terms of CBRs of the in situ layers
CBR (%)
derived from well-established DCP/CBR correlations related to DCP

10
penetration rates. Such derived in situ CBRs represent the in situ condition
of the materials at the time of testing, and will not normally be similar to
1
laboratory derived CBR values which are for a specific prescribed (usually
1 10 100
DN (mm/blow)
regarded as worst case, i.e. soaked) test procedure. It is therefore essential
to relate the field DCP results to in-situ moisture conditions at the time of
DCP-CBR correlation. testing, primarily to determine whether these are representative of typical
field conditions, or of wetter or drier than normal conditions.
It is essential to relate the field DCP results to

in-situ moisture condition at the time of testing.

Typical examples of DCP curve, balance curve
and layer strength diagram.
An important aspect of the DCP survey is the determination of the subgrade
CBR as this will influence the overall performance of the pavement structure,
and define the overlying layer configuration needed to provide adequate
subgrade protection.
The DCP data should be plotted in terms of the penetration rate (mm/blow) The new Windows version of DCP saves its
through the individual layers, comparing the layer thicknesses specified in files in Microsoft Access® allowing importation
into Microsoft Excel®. In this way, specific
the pavement design with those identified by a significant change in the relationships between the layer thicknesses
DCP penetration rate. This allows an evaluation of the construction quality and strengths can be investigated.
(in terms of layer thickness) as well as determining an average in situ CBR
strength for the constructed pavement layers.
The raw field data will normally be processed using the DCP software currently
available at the Roads Department.
The derived CBR data characterising the spatial variation of the DCP profiles
along the project length should also be entered into a spreadsheet. This can
be used to graphically evaluate variations in layer strengths and thicknesses
along the project and allow identification of uniform sections.

Example of Windows DCP screen output.

9 TEST PIT PROFILING
AND SAMPLING
9.1 General
Only by observing the materials and layers in a road directly can exact
conditions of construction be determined. The excavation of pits through
the pavement with testing and sampling of the materials comprising each
layer is therefore an essential part of most structural pavement evaluations.
The excavation and testing of test pits is probably the most costly part of a
routine pavement evaluation. It is imperative therefore that the optimum
locations are tested and that the maximum information is obtained from
each test pit.
9.2 Location
The location of test pits should be carried out once sufficient information is
available to divide the road into a number of preliminary uniform sections.
This can be based on PMS information, rut depth surveys, deflection surveys
and visual condition evaluations in association with the as-built records if
available.
It is, however, important to ensure that sufficient test pits are investigated to
provide adequate information regarding the nature of the pavement and
subgrade materials and thicknesses of the pavement layers. It is suggested
in Table 4.2 that at least 2 or 3 test pits should be excavated per 10 km. No
hard and fast rules can be laid down for the test pit frequency, as it is depen-
dent on the nature of the problem. The situation, length of uniform sections, The location of the pits should be such that
pavement layer material use and design and mode of distress will all affect adequate warning is provided to traffic and the
the number of test pits. Ideally, however, test pits should be located in at area should be cordoned off with high visibility
least: cones. Warning signs and flagmen must also
be established to provide adequate advance
n the visibly best, and warning to the road user.
n the visibly worst areas of each uniform section.

This facilitates an indication of the variability of the materials, thicknesses
and construction quality within the uniform section.
The locations of test pits should be determined using all the available infor-
mation in order to ensure that the most valuable data is obtained.
At each location, test pits are usually excavated in the most distressed area of
the road, which in most cases will be the more heavily loaded lane. The outer
wheel path will be tested in nearly all cases while the inner wheel path and
centre-line area are usually only investigated if the cause of any localised
structural failure is being established. Test pits may also be excavated in the
inner wheelpath or near the centre-line, however, for audit of material
properties or construction quality. Comparison of the properties of materi-
als in the outer wheel path (trafficked under the worst prevailing conditions)
and centre-line area (generally untrafficked) is particularly useful for
investigating whether any material degradation has occurred under traffic.
Typical test pit.
Because of the general variability in material testing, it is recommended that
all samples, trial pits, etc should be at least duplicated (or if possible triplicated

where time and costs permit). Triplication allows the determination of at
least a mean and standard deviation for the material properties. Without
these parameters, it is not possible to statistically compare whether the results
come from the same population. Sampling size must be determined in
accordance with the maximum size of the material e.g. British Standards
give recommended minimum sample sizes. The difficulties with testing are
highlighted by the likely inaccuracies associated with carrying out single in
situ sand replacement density tests - usually several tests are needed to obtain
a reliable mean value.
Test pitting is a skilled operation which must be 9.3 Size and depth
done methodically and comprehensively to gain
maximum benefit. Test pits are more “traffic
friendly” if excavated at an angle to the traffic
The areal extent of the test pit should be as small as possible so as to cause
direction. minimal disturbance of the pavement surface but large enough to:
n Permit a sample of sufficient size to be collected. A 100 mm thick
layer of typical material would generally provide 180 to 200 kg of
sample per square metre.
n Allow enough space to carry out in situ testing and excavation at
depth.
At least one test pit at each site should be excavated to a depth at which the
in situ material can be inspected and sampled. Where the road is built on a
high fill:
n the top 300 mm of the fill material, and
n all the overlying layers should be
atested asampled and aprofiled.
9.4 In situ measurements

It is very difficult to carry out density tests The investigation of test pits should follow a standard procedure.
(sand replacement or nuclear) and material
profiling at subgrade level if the opening of the n The surfacing should be carefully described, rut depths measured,
pit is too small. and the general condition of the road surface, shoulders, and drainage
around the site should be recorded.
It is recommended that a DCP test be carried
out at each test pit site before opening. n The surfacing should then be carefully removed causing minimal
All moisture samples should be placed in double
disturbance of the upper base course.
plastic bags and properly sealed. Labels must n Immediately on exposure of the base to the atmosphere, an in situ
be placed between the two bags using a
durable labelling system. It is also good practice density determination should be carried out on the base course using
to write directly on the bags as well, with an either a nuclear method or sand replacement. With both methods, it is
indelible black marker, to minimise the possibility essential to obtain samples for accurate gravimetric moisture content
of confusion.
determinations. These samples should be at least one kilogram in mass
and should be oven dried to constant mass (not only for 24 hours). All
dry densities should be calculated using the gravimetric moisture
content and not the nuclear moisture contents.

n Once the density of the base has been determined, the material should
be loosened and sampled with sufficient material being collected for
the laboratory testing envisaged.
n The test pit should then be carefully trimmed and cleaned until the
next layer is exposed. If the materials in the two layers differ
significantly, it is easy to prepare a suitable surface on the next layer
for density testing. If the material comprising the underlying layer is
similar to the base, difficulty is sometimes experienced in locating
the top of the underlying layer. Well-defined compaction planes are,
however, commonly observed.
n Each pavement layer should then be tested and sampled in turn until
the subgrade is reached and this should be excavated until it is clear
that the in situ material has been reached.
n Test pits should be carefully reinstated with similar quality materials
to those removed and the hole sealed with cold-mix asphalt.
During sampling of test pits a summary of all samples collected with their It is often advisable to lightly stabilise (not
more than 3 per cent lime or cement) the base
depth and description should be made. This can be included on the soil pro- layer during re-instatement of test pits.
file description form. The samples removed from the test pits should be Adequate compaction equipment must be
carefully bagged and labelled prior to submitting to the laboratory for available to ensure sufficient compaction
appropriate testing. This testing usually entails grading analysis, Atterberg during re-instatement.
limits and bar linear shrinkage, compaction characteristics, strength testing

(CBR), and any other testing considered necessary e.g. durability, electrical
conductivity, etc.
9.5 Description and logging of the

soil profile
The seal removed from the trial pit should be closely inspected and
descriptions of the bituminous surfacing (type, binder condition, adhesion
to the base and to chippings, prime penetration, etc.) recorded.
Once the pit has been tested and sampled, one “wall” of it should be scraped
with a spade and the pavement profile described and measured. The
description should follow the standard method17. A summary sheet of the A photograph of the wall of the test pit (including
a scale) can assist with recalling details regarding
descriptions is included in Appendix F, together with a standard field data the pavement during later evaluation of the data.
collection form. Pavements that contain any chemically stabilised layers Identification of the location and details of the
should be sprayed with a phenolphthalein solution to determine whether pit should be included in the photograph.
any carbonation of the layer has occurred. Those areas of stabilised materi-
als that do not react with the phenolphthalein solution (i.e. turn a dark red
colour) should also be sprayed with a dilute hydrochloric acid solution and
the degree of any reaction (fizzing) recorded. If possible, similar material
that has not been stabilised should also be checked for the acid reaction and
whether the reaction is weaker or the same as the stabilised layer. This
indicates whether calcium carbonate occurs naturally in the material.
A photograph like this will help in recalling

details, such as the crack through the base layer.

The condition and shapes of the layer interfaces should be examined to
determine where rutting and failure originates.
n Deep ruts at the surface not reflected at the base/subbase interface
indicate that the rutting has taken place in the base course or asphalt
surfacing.
n Where the surface rut is mirrored at the base/subbase interface or the
subbase/subgrade interface, the surface rutting is a consequence of
compaction or shear at a depth below the interface. Shearing within
layers in the form of shiny shear planes (slickensides) can sometimes
be observed in specific layers indicating problems within that layer.
Phenolphthalein reaction in stabilised subbase.

The use of moisture prediction models provides The results of laboratory and field-testing carried out must be summarised
a good indication of whether the moisture on the standard form used in the Botswana Laboratory Management Sys-
contents of the pavement layers are higher
than expected and thus whether drainage tem. Graphic illustrations of the pit profiles including depths and material
problems contribute to any defects. Although descriptions must be provided. This can also be done using a standard
the moisture within the pavement theoretically spreadsheet.
stabilises at an equilibrium value, the season
during which the moisture determinations were Aspects such as the in situ moisture content can be evaluated in terms of
determined should be taken into account,
especially for moisture determinations in the
expected equilibrium moisture contents (EMC) and predicted equilibrium
outer wheel path areas of the road where to optimum moisture content ratios (EMC/OMC). Various models are
seasonal moisture fluctuations are common. available for the former, these being described in detail in Appendix G.

PART C PAVEMENT DATA
ANALYSIS
10 Data analysis
1 Introduction
3 Types of pavement
distress

5 Visual evaluation

methods
8 Pavement strength

sampling

data analysis
Part D References
Appendices

10 DATA ANALYSIS
10.1 Data analysis phases and
methodologies
Mechanisms should be installed that allow In carrying out pavement evaluations for rehabilitation or other purposes, it
pavement evaluation teams to gain the maximum is essential that the teams involved communicate closely with each other. It
benefit from the BRMS, which is an expensive
project that is continually being updated and frequently occurs that one team carries out the visual evaluation while field-
improved. The involvement of a key player in the testing, other teams carry out test pitting and sampling and the materials are
pavement evaluation team with the BRMS is thus submitted to the laboratory where a different team does the testing. The
recommended.
results are then collated by someone who may not have been involved with
any of the field or laboratory work and then presented to the rehabilitation
design engineer who, up to this stage, has had very little involvement with
the project. It is essential that an engineer or team be appointed to co-ordinate
and follow the total process and provide sufficient continuity to ensure that
all components of the evaluation are used to maximum benefit.
Once all the field and laboratory data have been gathered, they should be
compiled into a format that the rehabilitation design engineer can use with
both confidence and minimal referral to the pavement evaluators. The data
should be verified, validated and then presented in the required or a stan-
Example of X-Y plot.
dard format.
As noted previously it is essential for all collected Graphical representation of the data is the fundamental basis by which uni-
data to be referred to the actual field positions, form sections can be identified and classified. It also allows initial data
usually in terms of the milestones (kilometre
posts). This then allows direct comparison of data validation by clearly showing if there are outliers in the data sets that should
values and obviously means that uniform be removed. The approach recommended is the use of spreadsheets for data
sections can be readily identified on the road. processing, whereby all available raw data values are firstly entered against
lineal position. This gives the basis for X-Y plotting and data review.
Commercial programmes such as DOTPLOT© Data such as test pit profiles should be prepared to indicate the uniformity
can be used for describing test pit profiles in a
standard graphical format.
of pavement structures in terms of layer thicknesses and material descriptions.
10.2 Chainage adjustment

Although statistical methods can be used for data It is essential that all data presented be validated to ensure that:
validation, inspection of the data for obvious outliers
and input errors is recommended. n the information is correct;
n laboratory test results are correctly calculated, and
n all the information available is presented.
This is carried out during preparation of spreadsheets of all the data collected.
In line with the foregoing, the first part of data analysis is to make any
necessary adjustments to the reported position data (the X-axis values) to
tie in with milestones or other known fixed features. For data collected at
specific points along the road (typically DCPs, Benkelman Beam or FWD
deflections, rut measurements and test pits), there should be no real need
for such adjustment, as the positions ought to be recorded reasonably
accurately relative to the milestones.
Main adjustments will normally be needed for data collected continuously,
and initially referred to vehicle odometer readings (typically roughness

measurements and Deflectograph deflections). Often the start (or end) of
the project is taken as the zero reading. The process is straightforward,
arising from the fact that the measured distances for the collected data are
unlikely to be exactly 1 000 metres between the milestones (kilometre posts).
The kilometre post positioning is assumed to be fixed at 1 000 m intervals,
regardless if this is true or not. The field recorded distances between kilo-
metre posts are uniformly adjusted to 1 000 metres (or acceptable accuracy).
The ideal case is where kilometre posts are accurate and any odometer
measurement differences are consistent, because the whole data set can then
be adjusted by the same factor. An example for a short section is shown in
Table 10.1. It should be noted that the actual process is greatly streamlined
in the spreadsheet data manipulation, and the main data check is the
comparison of the adjusted chainage with the fixed reference (columns (V)
and (II)).
Comment on
(I) (II) (III) (IV) (V)
comparison
Recorded Fixed Initial Adjusted Adjusted
of fixed and
stationing reference comparison recorded chainage
adjusted
chainage Start + (I) stationing Start +
chainage
(IV)
values
0 36 + 65 36 + 65 0 36 + 65 Section start
0,35 37 37 + 000 a 0,35 37 + 000 No change
1,4 38 38 + 050 r 1.370 38 + 020 + 20 metre, OK
2,4 39 39 + 050 r 2.341 38 + 991 - 9 metre, OK
3,45 40 40 + 100 r 3.360 40 + 010 + 10 metre, OK
4,5 41 41 + 150 r 4.380 41 + 030 + 30 metre, OK
4,83 41 + 35 41 + 480 r 4.700 41 + 35 Section end
Total Total Correlation

4,83 km 4.700 km satisfactory -
Proportioned
adjustment Adjustment all recorded
by 4,35/4,48
needed needed from station values
from station
between 0,35 km 37 to adjusted by
0,35/km 37
and 4,83 i.e., 41,35, i.e. simple pro-
over 4,48 km over 4,35 km portionering
Table 10.1 Chainage adjustment example.
An initial comparison is made to see whether the recorded stationing can be

used without adjustment (column (III)). In this case, the comparison at km
37 is fine, but thereafter the difference is unsatisfactory and the recorded
stationing needs adjustment. In this example, a simple linear proportioning
is used based on recorded and actual distances. Where the difference is still
unacceptable after this process, say more than 30 metres, individual pro-
portioning over shorter fixed distances can be undertaken in the same manner.
All intermediate positions (indicated by r in the table) will be adjusted by
the selected proportioning applicable to the specific range.
Once the adjustments are completed to give a satisfactory accuracy, the
revised chainage positions are used as the x-axis for subsequent graphical
review of the data. Each data set should firstly be viewed for the whole
length to highlight anomalous data values and any outliers evaluated and if
necessary removed from the data set (with justification).

Accurate markings on the road surface can be It should be noted that this method is an expedient to preliminary data
used as semi-permanent reference points for
evaluation purposes and for subsequent
evaluation. It allows ready field identification of sections in relation to fixed
rehabilitation design and cost estimation. features, while acknowledging that different vehicles used will usually vary
slightly in distance measurement. Errors in chainages of marker posts will
be retained by this method. The method proposed above is considered the
most cost-effective for typical pavement evaluations, but it would be better
to have the site correctly marked out first, using an accurate odometer (and/
or surveyors).
Accurate marking is essential for subsequent rehabilitation design and cost
estimation.
10.3 Identification of uniform

sections
When considering rehabilitation and/or reconstruction strategies, the road
length must be delineated into uniform sections, based on a number of factors
influencing the pavement performance. Generally, these factors include:
n Pavement type;
n Construction history (incl. rehabilitation & maintenance measures);
n Pavement cross-section (materials, thicknesses, drainage);
n Traffic; and Pavement condition.
Ideally, an historic pavement database should incorporate all these factors.
Example of uniform sectioning based on deflection

data.

The as-built approach for delineating uniform sections consists of isolating
each already known factor and identifying all significant changes per factor,
along the road length. The road is subsequently divided in a number of
sections characterised by a unique combination of factors (see Figure 10.1).
Project Length
Start Chainage End Chainage
Bituminous Base ( B ) Granular Base (G)

Pavement Type
Overlay (Ov) Original (Or) Base milled & replaced (M&R)

Construction History
Overlay
Orig. Asphalt Orig. Asphalt Orig. Asphalt New Asphalt
BTB BTB Orig. Granular base New Granular base
Cemented Cemented Cemented Cemented
Cross-Section
CS 1 CS 2 CS 3 CS 4
Traffic Level 1 Traffic Level 2 Traffic Level 3

Traffic
TL 1 TL 2 TL 3
B B B G G G
Ov Or Or Or M&R M&R
CS 1 CS 2 CS 2 CS 3 CS 4 CS 4
TL 1 TL 1 TL 2 TL 2 TL 2 TL 3
UNIFORM SECTIONS
Figure 10.1 Delineation of sections using as-built

approach.
Pavement monitoring activities discussed in this document measure the
quality and quantity of the pavement condition. These include deflections,
rut depths, visual condition, riding quality, pavement strengths, etc, and are
most frequently used for delineating uniform sections. As overlaps and dis-
agreements between the different parameters may confound the breakdown
into uniform sections, it is recommended that the parameters are prioritised
with:
n peak deflection being the first choice, followed by;
n radius of curvature (when available);
n rutting;
n riding quality;
n visual condition index, and finally
n DSN.
The analytical approach consists, mostly, of plots of the measured pave-

ment condition parameter as a function of the distance along the project.
Once the plot has been generated, it can be used to delineate uniform sections
through several methods. A visual and, therefore, subjective examination of
the plot represents the simplest, primary method.

In addition, several mathematically defined methods can be used for
optimising the findings. For example, the "cumulative difference" method
Statistical approaches generally involve the as described in the AASHTO manual relies on the concept of area under a
processing of pavement condition data irre- curve. The simpler cumulative sum ("Cusum") technique is recommended.
spective of their positions along the project. Such
approaches normally apply to widely scattered The method is only useful when the data has been sampled continuously at
data, and require in-depth statistical knowledge regular intervals along the project length.
and advanced software. These techniques are
considered unnecessarily cumbersome. In the Cusum technique, plots of the cumulative sums of deviations from
the mean parameter value against chainage can be used to differentiate
homogeneous sections. The Cusum is calculated in the following way:
Si = xi - xm + Si-1
where
xi = Parameter value at chainage i
xm = Mean parameter value
Si = Cumulative sum of the deviations from the mean parameter value at
chainage i
Si-1 = Cumulated sum at the previous chainage i.e. chainage (i-1)
Using the cumulative sums, the extent to which the measured parameter on
the different sections of road varies from the mean parameter value of the
entire surveyed length can be determined. Changes in the slope of the line
connecting the plotted cumulative sums will indicate boundaries between
homogeneous sections.
The Cusum calculation technique is included in a spreadsheet (Appendix
H).
10.4 Data evaluation

A 90 th percentile is generally recommended and
implies that 10 per cent of the road is in a worse The data should be classified individually for each uniform section in order
state with respect to that parameter which for
most practical purposes is in line with the natural to define the representative characteristics of each section. The application
variability of typical engineering parameters. of simple statistics is generally appropriate. Mean values, maxima, minima,
and various percentiles of the condition parameters have been used in the
past. It is recommended, however, that the 90th percentile be generally used.
A different percentile value may be used if considered necessary by the
design engineer e.g. where light traffic conditions prevail or where axle
loads are closely controlled. The application of these simple statistical para-
meters assumes that the data complies reasonably closely with a normal
distribution.

11 INTERPRETATION OF
RESULTS
11.1 Simple methods of data
evaluation
An holistic evaluation of all the data obtained from the pavement evaluation During data analysis, past experience should be
will identify the uniform sections within the project. The data should then utilised. Practical experience often shows that rut-ting
and isolated shear failures frequently occur and
be compiled into a table in which the selected percentile values of relevant profilerate much quicker than theoretical analysis
parameters are tabulated by uniform section. The relevant parameters include indicates.
those directly pertaining to the structural condition of the pavement and
include deflection data, rut depths, riding quality, Visual Condition Index
and layer strengths. An example of an Excel spreadsheet for this purpose is
provided in Appendix I.
Once the data is tabulated, it is possible to highlight those parameters that
characterise the individual uniform sections. By referring to the visual survey
data, the test results and the test pit profiles, the reasons for any deficiencies
in the uniform sections can be determined. Aspects such as layer thicknesses,
material quality, construction quality (particularly compaction), prevailing
moisture contents, etc, need to be specifically evaluated.
It is useful to plot all of the field survey data on a summary sheet showing
the visual, manual and mechanical survey results with indications of the
condition rating (Appendix I).
11.2 Parameters obtained from

analyses
Once the parameters have been tabulated, relevant ones can be compared
with standard criteria for defining the need for any improvement or
upgrading. These criteria represent the boundaries between sound and
warning (X), and warning and severe (Y) conditions and are summarised in
Table 11.1. Road categories used in Botswana are shown as Category I and
II. These are equivalent to major rural roads and lightly trafficked roads
respectively.

Category of road
Category I Category II
Parameter Base course material X Y X Y
Gravel/untreated subbase 0,5 1,0 0,7 1,4

Gravel/treated subbase 0,4 0,9 0,6 1,3
Lightly cemented 0,4 0,8 0,6 1,2
Deflection (mm)* Bituminous 0,3 0,7 0,5 1,0
(Benkelman beam) Cemented 0,3 0,6 0,4 0,8
Radius of Gravel/untreated subbase 80 45 60 30

curvature Gravel/treated subbase 110 60 80 40
(m)* Lightly cemented 150 75 100 50
(Dehlen Curvature Bituminous 200 90 130 65
Meter) Cemented 200 100 150 80
D C P* All basecourse materials: 230 110 155 70

(DSN800 or number assumed that readings
of blows to penetrate are taken at equilibrium
800 mm) moisture content
Rut depth (mm)* 10 20 15 25
Riding quality (IRI)* 3,5 4,2 4,2 5,1
Cracking (%)**
10 20 15 25
- Crocodile
45 75 60 90
- Longitudinal
* Usually the 90th percentile values.
** As a percentage of length of the section being evaluated.
Table 11.1 Ranges of warning values for data interpretation.
11.3 Estimation of remaining life

Methods for estimating remaining pavement life The data presented to the engineer will allow an estimate of the remaining
are discussed in some detail in the SATCC Code life of each of the uniform sections. Various techniques are available for this
of Practice for Pavement Rehabilitation18 and
TRH 128 and additional reference should be and four of them are commonly used. The different methods invariably give
made to these. a range of results.
South African Mechanistic Design Method (SAMDM): This method requires
sophisticated computer modelling of the stresses and strains developed in
the pavement structure. In order to carry out these analyses, knowledge of
the moduli of the materials, the layer thicknesses and Poisson’s ratio of each
layer is necessary. This is usually obtained from back-analysis of deflection
bowl parameters. An approximation of the material moduli can be obtained
from the DCP analysis. The stresses and strains calculated are then related
through various transfer functions to the carrying capacity of the pavement.
A well-balanced pavement is one that shows a DCP method: This makes use of a comparison of the pavement structure in
progressive decrease in stiffness with depth. No
excessively strong or weak layers cause terms of the computer generated redefined layer strength diagram with an
deviation from a smooth DCP curve in the upper equivalency line developed from the predicted future traffic and the moi-
800 mm of the profile. sture regime. It should be noted that this is only reliable for a well-balanced
pavement structure.
AASHTO method: An estimate of the residual life of the pavement can be
calculated using the AASHTO structural number method. This requires inputs
of the subgrade modulus, the AASHTO structural number, the expected
change in riding quality and various reliability functions. This method is
known to be conservative in terms of its pavement thickness and is thus
equally conservative in terms of the estimated remaining life.
Asphalt Institute method: This relates the maximum annual deflection directly
to the structural capacity of the pavement. By subtracting the known previous
traffic, the remaining life can be estimated. The method is empirically based
and is valid for traffic up to about 7 million standard axles. It does allow,
however, for higher traffic counts by use of the pavement component analysis
procedure.
Other techniques such as the HDM-3 relationship between rutting and stan-
dard axles can be used but are often based on limited data and appear to be
excessively general. The data used is often from a localised evaluation with
material and climatic conditions dissimilar to those in Botswana and would
thus require calibration for local use.
11.4 Sensitivity analyses

The structural analysis and remaining life estimate techniques for pavements Sensitivity analyses are carried out to investigate
include many assumptions. The standard assumptions that the materials are the effect of changes in the input variables on the
outcome. They are particularly useful where
isotropic, uniform and elastic are known to be false for road construction confidence in the accuracy of the input para-
materials. Methods of overcoming this problem are being actively researched meters is poor as they can highlight critical and
but in the mean time, traditional analyses are utilised. less critical parameters.
It is recommended that pavement analyses and rehabilitation designs should

be based on a number of methods using a range of appropriate parameter
values. The parameter values should reflect conditions that are expected to
apply to seasons other than that during which field-testing was carried out
(particularly if field work was carried out in a dry season). The reliability of
parameters such as traffic and E-moduli should also be considered.
Comparison of the various outputs will allow the selection of rehabilitation
design parameters with a greater degree of confidence.

PART D REFERENCES AND
APPENDICES
1 Introduction

I
3 Types of pavement
distress

5 Visual evaluation

methods
8 Pavement strength

sampling

data analysis
Part D References
Appendices

REFERENCES
1. Paterson, W D O. 1987. Road deterioration and maintenance effects: models for planning and management. Johns Hopkins University
Press, Baltimore.
2. Emery, S J. 1992. The prediction of moisture content in untreated pavement layers and an application to design in southern Africa. CSIR
Research Report 644 (Bulletin 20), Division of Roads and Transport Technology, CSIR, Pretoria.
3. Committee of State Road Authorities (CSRA). 1992. Pavement Management Systems: standard visual assessment manual for flexible
pavements. Technical Methods for Highways (TMH) 9, CSRA, Pretoria.
4. Committee of State Road Authorities (CSRA). 1985. Nomenclature and methods for describing the condition of asphalt pavements.
Technical Recommendations for Highways (TRH) 6, CSRA, Pretoria.
5. Committee of State Road Authorities (CSRA). 1994. Pavement Management Systems. Technical Recommendations for Highways
(TRH) 22, CSRA, Pretoria.
6. Raters manual for the visual assessment of paved roads. 1994/96. Prepared by TTCS/VWL, Botswana.(Botswana Road Manage
ment System Manual).
7. Transport and Road Research Laboratory, 1999. A guide to the pavement evaluation and maintenance of bitumen-surfaced roads in
tropical and sub-tropical countries. Overseas Road Note 18. Transport Research Laboratory, Crowthorne.
8. Committee of Land Transport Officials (COLTO). 1997. Flexible pavement rehabilitation investigation and design.Technical Recom-
mendations for Highways (TRH) 12, COLTO, Pretoria.
9. Cundill, M A. 1991. MERLIN – A low cost machine for measuring road roughness in developing countries. Transportation Research
Record, 1291, National Research Council, Washington, D.C., pp 106-112.
10. Young, J C. 1977. Calibration, maintenance and use of the Rolling Straight-edge. Supplementary Report 290, Transport and Road Research
Laboratory, Crowthorne.
11. American Association of State Highway and Transportation Officials (AASHTO). 1986. AASHTO guide for design of pavement
structures. AASHTO, Washington, D.C.
12. Hodges, J W, Rolt, J and Jones, T E. 1975. The Kenya Road Transport Cost Study: Research on road deterioration. Laboratory Report
673, Transport and Road Research Laboratory, Crowthorne, England.
13. American Association of State Highway and Transportation Officials (AASHTO). 1993. Design of pavement structures, AASHTO,
Washington DC.
14. Jameson, G W. 1993. Development of procedures to predict structural number and subgrade strength from falling weight deflectometer
deflections.
15. Rohde, G T, 1994. Determining pavement structural number from FWD testing. Transportation Research Record 1448, Transportation
Research Board, Washington DC.
16. Roberts, J D and Martin, T C. 1996. Recommendations for monitoring pavement performance. Research Report ARR 293. ARRB
Transport Research Ltd, Victoria.
17. Jennings, JE, Brink, ABA, and Williams, AAB. 1973. Revised guide to soil profiling for civil engineering purposes in southern Africa.
The Civil Engineer in South Africa, January 1973, pp 3-12.
18. Southern Africa Transportation Communications Commission (SATCC). 1998. SATCC Code of Practice for pavement rehabilitation
(Draft). SATCC, Maputo, Mocambique.
19. Transport and Road Research Laboratory, 1985. Maintenance techniques for district engineers. Overseas Road Note 2, Transport
Research Laboratory, Crowthorne.
20. Smith, H R and Jones, C R. 1980. Measurement of pavement deflections in tropical and sub-tropical climates. TRRL Report LR 935.
Transport Research Laboratory, Crowthorne.
21. Jones, C R and H R Smith, 1980. Deflection temperature relationships for bituminous surfacings in Kenya. TRRL Report LR 936.
Transport Research Laboratory, Crowthorne.

APPENDICES page
Appendix A Visual evaluation forms and description....................................................................................65

Appendix B Visual Condition Index (VCI)..................................................................................................67
Appendix C Roughness measurement.........................................................................................................68
Appendix D Pavement layer strength coefficients for structural number after Paterson, 1985).......................74
Appendix E Deflection measurement.........................................................................................................75
Appendix F Test pit profiles.......................................................................................................................83
Appendix G Equilibrium and predicted moisture content.................................................................................86
Appendix H Cumulative sum method for identifying uniform sections.............................................................87
Appendix I Pavement condition rating form.................................................................................................95
Appendix J Definition of terms.....................................................................................................................96

Appendix A: Visual evaluation forms and
description

Appendix B Visual Condition Index (VCI)
Equation B.1 Equation B.2
N N
{
VCIp = 100 1 _ C * ( S F )}n C = 1/(S Fn
(max) )
n=1 n=1
where
VCIp = Preliminary VCI Fn(max) = Fn with degree and extent ratings set at
Fn = Dn*En*Wn maximum. Where two or more options are available per
distress, e.g. aggregate loss, use the average Wn for the
n= visual assessment item number (Table B.1) calculation of Fn(max).
Dn = degree rating of defect n
Range: 0 – 4 for functional defects Equation B.3 is applied to transform VCIp to a standard per
0 – 5 for other defects centage scale.
En = Extent rating of defect n Equation B.3
Range: Default 3 for functional defects
VCI = (a * VCIp + b * VCI 2p)2
0 – 5 for other defects
where
Wn = weight for defect n (Table B.1)
a = 0.02509
b = 0.0007
and 0 < VCI < 100
Item No. Assessment items Weight (Wn)
1 Surfacing failures 6,5
2 Surfacing cracks 5,0
3 Aggregate loss 3,0
4 Dry/brittle 3,0
5 Bleeding 3,0
6 Map/Crocodile cracks 10,0
7 Block cracks 6,5
8 Longitudinal cracks 4,5
9 Transverse cracks 4,5
10 Pumping 10,0
11 Rutting 8,0
12 Deformation 4,0
13 Patching 8,0
14 Failures/potholes 15,0
15 Edge breaking 3,5
16 Riding quality 5,5
17 Skid resistance 3,0
18 Surface drainage 3,0
19 Unpaved shoulders 3,5
Table B.1 Proposed weightings for VCI calculation (modified after TRH 225).

Appendix C Roughness measurement
C1 INTRODUCTION
The standard measure of road roughness is the International Roughness Index (IRI). This was developed during ‘The International
Road Roughness Experiment’ in Brazil. It is a mathematical quarter car simulation of the motion of a vehicle at a speed of 80 kph over
the measured profile and can be calculated directly from road levels measured at frequent intervals. Devices for measuring levels are
usually either slow and labour intensive or fast, automatic and expensive. Hence, the roughness of the road is usually measured using
a Response Type Road Roughness Measuring System (RTRRMS) that must be periodically calibrated to allow the values of roughness
to be reported in terms of IRI. Methods of calibration include a rod and level survey or a standard instrument, such as the TRL Profile
Beam, the MERLIN (Machine for Evaluating Roughness using Low-cost Instrumentation), the Face Dipstick and the ARRB Walking
Profiler.
C2 OPERATION OF THE MERLIN

The MERLIN is available in Botswana and is thus discussed in detail below.
A diagram of the equipment is shown in Figure C.1. It has two feet, 1.8 metres apart which rest on the road surface along the wheel
path. A moveable probe is placed on the road surface mid-way between the two feet and measures the vertical distance, ‘y’, between
the road surface under the probe and the centre point of an imaginary line joining the two feet.
The result is recorded on a data chart mounted on the machine.

By recording measurements along the wheel path, a histogram
of ‘y’ can be built up on the chart. The width of this histogram can
Pointer
then be used to determine the IRI.
Chart
Handles To determine the IRI, 200 measurements are usually made at
regular intervals. For each measurement, the position of the poin-
ter on the chart, shown in Figure C.2, is marked by a cross in the
box in line with the pointer and, to keep a count of the total number
of measurements made, a cross is also put in the ‘tally box’ on the
chart. When the 200 measurements have been made the position
mid-way between the 10th and 11th crosses, counting in from one
end of the distribution is marked on the chart. The procedure is
repeated for the other end of the distribution. The spacing between
the two marks, D, is then measured in millimetres.
Wheel with marker in Weight Pivot Rear foot
Moving arm
For earth, gravel, surfaced dressed and asphaltic concrete
Probe
roads, the IRI can be determined using the following equation.
IRI = 0.593 + 0.0471 D
This equation assumes that the MERLIN has a mechanical

amplification factor of 10. In practice this may not be true be-
cause of small errors in manufacturing. Therefore, before the
MERLIN is used the amplification has to be checked and the
value of D corrected. To do this the instrument is rested with the
probe on a smooth surface and the position of the pointer
carefully marked on the chart. The probe is then raised and a
calibration block approximately 6mm thick placed under the
probe. The new position of the pointer is marked. If the distance
Figure C1 Operation of the Merlin.

between the marks on the chart is S and the thickness of the block T then measurements made on the chart should be multiplied
by the scaling factor:
Scaling factor = 10 T
S
Length of test section used in calibration

If 200 measurements (one at each wheel revolution) are taken using a MERLIN with a 26-inch (415 mm) diameter wheel then the length
of the section surveyed will be 415 metres. For shorter or longer sections, a different procedure will be required. The guiding principles
are:
I) The test section should be a minimum of 200 metres long
II) Take approximately 200 readings per chart. With less than 200 readings the accuracy will decrease and with more the
chart becomes cluttered. If the number of readings differs from 200, then the number of crosses counted in from each
end of the distribution, to determine D, will also need to be changed. It should be 9 crosses for 180 readings, 11 for
220 readings etc.
III) Always take measurements with the marker on the wheel in contact with the road. This not only prevents errors due to
any variation in radius of the wheel but also avoids operator bias
IV) Take regularly-spaced measurements over the full length of the test section. This gives the most representative result.
V) If taking repeat measurements along a section, try to avoid taking readings at the same points on different passes. e.g.
start the second series of measurements half a metre from where the first series was started.

TEST SECTION :______________
WHEEL PATH :________________
DATE:_______________________
OPERATOR :_________________
Tally box
1 2 3 4 5 6 7 8 9 10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figure C2 Merlin data collection form.

Three examples are given below:
1) For a 210-metre test section take the measurements in two passes, taking one reading every revolution of the wheel,
and offseting the second pass by half a metre.
2) For a 280 metre test section take the measurements in two off-set passes, taking one reading every wheel revolution on the
first pass and one reading every two revolutions on the second.
3) For a 500-metre test section take the measurements in one pass, taking one measurement every wheel revolution and
omitting every fifth measurement. Or, rather that omitting readings, enlarge the tally box and take all 240 measurements.
Measure the limits on the chart by counting in 12 crosses rather than 10.
C3 ROUGHNESS SURVEYS USING A RTRRMS

When roughness measurements are needed on more than a few short sections of road, a RTRRMS is recommended. The main
advantages of these types of systems are their relative low cost and the high speed of data collection. The systems are capable of
surveys at speeds up to 80 km/h, so many hundreds of kilometres of road can be measured in a day.
The TRL Bump Integrator (BI) Unit is a response-type road roughness-measuring device that is mounted in a vehicle. The instrument
measures the roughness in terms of the cumulative uni-directional movement between the rear axle and the chassis of a vehicle in
motion. The BI system comprises a bump integrator unit and a counter unit and is powered by the 12-volt battery of the vehicle. The
NAASRA meter, Linear Displacement Integrator (LDI) and the Mays meter are similar response-type road roughness measuring
devices and the survey and calibration procedures will be similar to that used with the TRL BI Unit, described below.
C3.1 Fitting the TRL BI unit

The BI unit is mounted in a rear-wheel drive vehicle as shown in Figure C.3. The unit is bolted to the rear floorpan of the vehicle directly
above the centre of the rear axle. A 25mm hole needs to be cut in the floorpan and a bracket or hook fixed to the centre of the differential
housing of the rear axle
Before each survey, the flexible metal cord from the cylindrical drum of the BI unit is passed through the hole in the floor and hooked
onto the bracket on the rear axle. This cord must not touch the sides of the hole. Tension in the cord is maintained by a return spring
inside the drum of the BI unit. The BI unit measures the unidirectional movement, in centimetres, between the vehicle chassis and the
axle as the vehicle is driven along the road. This is displayed on a counter box, usually fixed to the front passenger fascia.
C3.2 Survey procedure

I) A safe working environment should be maintained at all times. As the vehicle may be moving slower than the majority of
other traffic, it should be clearly signed and fitted with flashing lights.
II) The vehicle should be well maintained and in good working order. The wheels should be properly balanced and the steering
geometry correctly aligned. The tyres should not have flat spots or be unduly worn. Tyre pressures should be maintained
precisely to the manufacturers specifications and always checked cold. The load in the vehicle must be constant. Ideally
the vehicle should contain only the driver and observer, and no other load should be carried.
III) The engine and suspension system should be fully warmed-up before measurements commence. This can be achieved
by driving the vehicle for at least 5km before measurements start.
IV) The tension cord from the BI unit to the axle should only be connected during the survey. At all other times, the cord should
be disconnected to stop unnecessary wear of the BI unit. When attaching the cord to the rear axle, the cord should be pre-
tensioned by turning the BI pulley 2.5 turns anti-clockwise. The wire is then wound around the pulley 2 turns in the same
direction as the arrow. Note: the pulley must NOT be turned clockwise or suddenly released after being tensioned as the
internal spring mechanism could be damaged.

Figure C3 Diagrammatical representation of the TRL Integrator Unit fittet to a vehicle.
V) When measurements are being taken the vehicle should be driven at constant speed, avoiding acceleration, deceleration
and gear changes. This is necessary because the vehicle’s response to a given profile varies with speed. To improve
reproducibility it is best to operate the RTRRMS at a standard speed of 80 km/h. However, if this speed is unsafe for
reasons of traffic, pedestrians or restrictive road geometry, a lower speed of 50 or 32 km/h can be used. Calibration
must be carried out for each operating speed used in the survey.
VI) Readings are recorded at half kilometre intervals. This distance should be measured with a precision odometer fitted
to the vehicle. The use of the vehicle odometer or kilometre posts is not recommended for survey purposes.
VII) There are two counters in the recording unit, connected by a changeover switch. This allows the observer to throw the
switch at the end of each measurement interval so that the reading can be manually recorded while the other counter
is working. The first counter can then be re-set to zero ready for the next changeover.
VIII)The type of road surfacing and any landmarks should be recorded to aid future analysis of the data. On completion of
the survey, the wire cord should be disconnected from the rear axle.
IX) After the survey, the results should be converted into vehicle response roughness values (VR). The counts measured
by the BI are in units of cumulative centimetres of uni-directional movement of the rear axle. These should be converted
to vehicle response roughness values using the following equation.
VR = BI count x 10
Section length
where
VR = Vehicle Response (mm/km)
BI = No of counts per section (cm)
Section length (kms)
X) These vehicle response roughness values should then be converted to units of estimated IRI, E[IRI], using a calibration
that is unique to the RTRRMS at that time.
C4 CALIBRATION OF A RTRRMS
The RTRRMS must be regularly calibrated against an instrument such as the MERLIN. Calibration should preferably be carried out
before the survey and checked on ‘control’ sites during the survey period to ensure that the RTRRMS remains within calibration. The
calibration of the RTRRMS will need to be re-checked before any subsequent surveys or after any part of the suspension of the vehicle
is replaced.
The calibration exercise involves comparing the results from the RTRRMS and the MERLIN over several short road sections. The
relationship obtained by this comparison can then be used to convert RTRRMS survey results into units of E[IRI]. The recommended
practice for roughness calibration is described below.
I) A minimum of eight sections on the road under evaluation should be selected with roughness levels that span the
range of roughness of the road. The sections should have a minimum length of 200m and should be of uniform

roughness over their length. In practice it may be difficult to find long homogeneous sections on very rough roads.
In this case it is better to include a shorter section than to omit high roughness sites from the calibration. The section s
should be straight and flat, with adequate run-up and slow-down lengths and should have no hazards, such as
junctions, which may prevent the vehicle travelling in a straight course at constant speed along the whole section.
II) The roughness of each section should be measured by the RTRRMS at the same vehicle speed that is to be used
for the survey. The value of VR (mm/km) should be the mean value of at least three test runs.
III) The MERLIN should be used to measure the IRI in both wheel paths. The average of these IRI values is then
plotted against the vehicle response for each of the test sections. The calibration equation for the RTRRMS is
then derived by calculating the best-fit line for the points. This relationship generally has a quadratic form but,
depending on the characteristics of the vehicles suspension and the levels of roughness over which the RTRRMS
has being calibrated, has also been found to be logarithmic.
E[IRI] = a + b VR + c VR2
Where
E[IRI] = Estimated IRI (m/km)
VR = Vehicle Response (mm/km)
a, b and c = constants
The calibration equation can then be used to convert data from the RTRRMS (VR) into units of E[IRI].
C5 INTERPRETATION OF RESULTS
To divide the road into homogeneous sections, such as to minimise the variation in roughness within each section, it is
recommended that the cumulative sum method be used (Appendix H).

Appendix D: Pavement layer strength coef-
ficients for structural number
(after Paterson, 19851)
Pavement layer Strength coefficient ai
Surface course
Asphalt mixtures (cold or hot premix of low stability) 0.20
Asphalt concrete (hot premix of high stability)1
MR30 = 1 500 MPa 0.30
MR30 = 2 500 MPa 0.40
MR30 = 4 000 MPa or greater 0.45
Base course
Granular materials2 For maximum axle loading
< 80 kN > 80 kN
CBR = 30 %3 0.07 0
CBR = 50 % 0.10 0
CBR = 70 % 0.12 0.10
CBR = 90 % 0.13 0.12
CBR = 110 % 0.14 0.14
Cemented materials4
UCS = 0.7 MPa 0.10
UCS = 2.0 MPa 0.15
UCS = 3.5 MPa 0.20
UCS = 5.0 MPa 0.24
Bituminous materials5 0.32
Subbase and selected subgrade layers
(to total pavement depth of 700 mm)
Granular materials
CBR = 5 % 0.06
CBR = 15 % 0.09
CBR = 25 % 0.10
CBR = 50 % 0.12
CBR = 100 % 0.14
Cemented materials
UCS > 0.7 MPa 0.14
1) Applicable only when thickness > 30 mm. MR30 = resilient modulus by direct tensile test at 30 °C.
2) ai = (29.14 CBR – 0.1977 CBR2 + 0.00045 CBR3)10-4: the coefficient ai may be increased by 60 per cent if CBR >70
and the subbase is cement- or lime-treated. Note ai = 0 for CBR < 60 when axle loading exceeds 80 kN.
3) CBR = California Bearing Ratio (per cent) determined at equilibrium in situ moisture and density conditions.
4) ai = 0.075 + 0.039 UCS – 0.00088 UCS2; where UCS = Unconfined compressive strength in MPa at 14 days. “Cemented”
implies development of tensile strength through Portland cement or lime treatment, or the use of certain flyash, slag,
lateritic or ferricrete materials that are self-cementing over time.
5) Dense-graded bitumen-treated base of high stiffness, e.g., MR20 = 4 000 MPa, resilient modulus by indirect tensile test at
20°C.
6) ai = 0.01 + 0.065 log10CBR.

Appendix E Deflection measurement
E1 INTRODUCTION
The structural integrity of a pavement can be quickly and efficiently assessed by applying a load to the pavement surface and
measuring the resulting deflections. The numerous pavement deflection measurement techniques currently in use can be
categorised according to the applied load characteristics. Measuring the pavement surface deflection under a static or slow
moving load (Benkelman Beam) represents the first generation approach. The next generation involved the application of a
dynamic vibratory load (Road Rater and Dynaflect). The third generation deflection equipment (Falling Weight Deflectometer)
simulates the effect of a moving wheel load by applying a dynamic impulse load. Future equipment will attempt to measure
deflections caused by an actual wheel load moving at highway speeds.
This Appendix gives descriptions, procedures for use, and factors influencing the application of both the Benkelman beam and
the Falling Weight Deflectometer. These represent the main methods that may be used to obtain pavement deflections.
E2 BACKGROUND TO DEFLECTION MEASUREMENTS

Early use of deflection data implied the analysis of maximum deflection relative to empirical (experience and/or experiment
based) standards. Generally, some statistical measure of the maximum deflection was compared with a “permissible” deflection
level. Should the measured value have exceeded the permissible one, an empirical rehabilitation procedure (e.g. an overlay)
would have been applied to adequately reduce the measured deflections.
As the understanding of pavement behaviour progressed, the mechanistic approach developed. This approach involves the
application of laws of physics in order to understand how the traffic loads are being distributed through the pavement layers.
Certain fundamental properties of materials must be known together with the layer thicknesses and load characteristics.
The current mechanistic-empirical approach incorporates elements of both individual approaches. The mechanistic component
involves the computation of pavement structural responses (deflections, stresses, strains) within the layers through the use of
physical models. The correlation between these responses and the pavement performance is given by the empirical component.
The simplest physical model of a pavement structure consists

of a succession of layers, each of them characterised by an
elastic modulus, Poisson’s ratio and thickness (see Figure
E1). The elastic modulus (E-modulus) is mathematically
defined as the constant ratio of stress and strain for that pave-
LAYER 1 ment layer’s material. The elastic modulus is expressed in
MPa and can typically vary between 30 000 MPa for Portland
LAYER 2 Cement Concrete to 35 MPa for subgrade soils.
The pavement material is also physically characterised by the
Poisson’s ratio, mathematically defined as the ratio of
LAYER 3 transverse to longitudinal strain of a loaded specimen.
Poisson’s ratio is dimensionless and can theoretically vary
from 0 to 0.5. Generally, “stiffer” materials will have lower
Poisson’s ratios than “softer” materials (e.g. from 0.15 for Port-
land Cement Concrete to 0.45 for subgrade soils).
LAYER 4
Figure E1 Simplified pavement structure model.

E3 DEFLECTION BEAM (BENKELMAN BEAM)
E3.1 General
This is the least expensive instrument for measuring deflections, and was originally devised by A C Benkelman. It is a
mechanical device that measures the maximum deflection of a road pavement under the dual rear wheels of a slowly moving
loaded lorry. The beam consists of a slender pivoted beam, approximately 3.7m long, supported in a low frame that rests on
the road. The frame is fitted with a dial gauge for registering the movement at one end of the pivoted beam, the other end of
which rests on the surface of the road. It is shown in Figure E2.
2,44m 1,22m
1,3m
Elevation
Initial position Initial position Dial

of wheels in Tip Twin Single
Beam of wheels in Pivot gauge Frame
rebound test feet foot
transient test
Plan
Figure E2 Diagrammatic representation of the Benkelman beam.
E3.2 Deflection beam survey procedure

A safe working environment should be maintained at all times. Many organisations will have on-site safety procedures that
should be followed. Where there are no local safety procedures those described in Overseas Road Note 219 are recommended.
Testing can be minimised by only taking measurements in the outer wheel path, as this usually is the most heavily trafficked
and therefore gives the poorest results. If it is not evident, however, that deflections measured in the outer wheel path are
consistently higher than in the inner wheel path, deflection beam measurements should be made in both wheelpaths of the
slow lane on dual carriageways and in both lanes of a single carriageway road.
Tests can be made at any frequency, but when measurements are taken at closely spaced regular intervals (say 25 or 50
metres) the additional time and cost implications for the Benkelman beam survey will not normally be merited by gains in data
quality. In this respect the automated deflection beam measurements using a Deflectograph can provide a cost-effective
alternative, while providing far more comprehensive coverage and detail.
Consequently, when using manual deflection beam measurements, it is recommended that the following strategy be adopted.
I) Tests are carried out on a basic pattern of 100 or 200-metre spacing.
II) Additional tests should be undertaken on any areas showing atypical surface distress.
III) When a deflection value indicates the need for a significantly thicker overlay than is required for the adjacent
section, the length of road involved should be determined by additional tests.
E3.3 Timing of deflection surveys

In some cases the moisture content of the road pavement, especially the subgrade, changes seasonally. In these circumstances
the tests should be carried out after the rainy season, when the road is at its weakest.

E3.4 Details of test truck
The truck must have dual rear wheels and should be loaded to a standard rear axle load if possible. The axle load must in any
case be recorded as load-related corrections to readings may be required. The standard axle load recommended in southern
Africa is 80 kN, although TRL recommends the use of a 63.2 kN rear axle load. Over this range of loads the maximum
deflection is usually linearly related to the applied load but this may not be the case with all structures or with structures that
have suffered considerable damage. The important factor is that the test method and test conditions must be compatible with
the deflection criteria and design procedures adopted. The effect of any differences from the original procedures adopted in
the deflection design criteria must be established for the roads under investigation.
E3.5 Test method

There are two basic methods20 which are commonly used for operating the deflection beam. These are the transient deflection
test and the rebound test.
E3.5.1 Transient deflection test

In this test the tip of the beam is inserted between the dual rear-wheel assembly of the loaded truck. The dial gauge is set to
zero and the truck then drives slowly forward. As the wheels approach the tip of the beam, the road surface deflects downwards
(loading deflection) and the movement is registered by the dial gauge. As the wheels move away from the tip of the beam, the
road surface recovers (recovery deflection) and the dial gauge reading returns to approximately zero. The test procedure is
summarised below.
I) Mark the point, in the vergeside wheelpath, at which the deflection is to be measured and position the lorry so
that the rear wheels are 1.3m behind the marked point.
II) Insert the deflection beam between the twin rear wheels until its measuring tip rests on the marked point. Insert
a second beam between the offside wheels, if deflections are to be measured in both wheelpaths. It is helpful in
positioning the lorry and aligning the beams if a pointer is fixed to the lorry 1.3m in front of each pair of rear
wheels.
III) Adjust the footscrews on the frame of the beam to ensure that the frame is level transversely and that the pivoted
arm is free to move. Adjust the dial gauge to zero and turn the buzzer on. Record the dial gauge reading which
should be zero or some small positive or negative number.
VI) The maximum and final reading of the dial gauge should be recorded while the lorry is driven slowly forward to
a point at least 5m in front of the marked point. The buzzer should remain on until the final reading is taken.
Care must be taken to ensure that a wheel does not touch the beam. If it does the test should be repeated.
VII) The transient deflection is the average of the loading and recovery deflections. Because of the 2:1 ratio of the
beam geometry over the pivot point (see Figure E2) the transient deflection is calculated by either:
VII) Adding the difference between initial and maximum dial gauge readings to the difference between maximum
and final dial gauge readings, or,
VIII)Calculating the loading deflection, as double the difference between the initial and maximum values, and the
recovery deflection, as double the difference between the maximum and final readings and then calculating the
mean of the two deflections.
IIX)At least two tests should be carried out at each chainage and the mean value is used to represent the transient
test result. If the results of the two tests do not fall within the repeatability limits described in Table E1 then a
third test should be carried out.

Mean deflection Max. permissible difference
betw een the tw o tests
(mm) (mm)
< 0,10 0,02
0,10 - 0,30 0,03
0,31 - 0,50 0,04
0,51 - 1,00 0,05
> 1,00 0,06

Table E1 Repeatability of duplicate transient deflection tests.
E3.5.2 Rebound deflection test

This is probably the most commonly used method which, while not as comprehensive as the transient method, allows a
greater production rate with less need for repeat measurements (e.g. due to the tyre touching the beam when guide pointers
are not used on the lorry). Because the rebound deflection can be influenced by the length of time during which the loading
wheels are stationary over the test point care must be taken over the exact procedure used. The rebound test is not
recommended by TRL for use on roads that may ‘creep’ under the effect of the stationary loading wheels.
For the rebound deflection test the dual wheels are positioned immediately above the test point and the measuring tip of the
beam is placed on the test point and between the dual wheels. The beam is adjusted in the same way as for the transient test
and when the initial reading has been noted, the lorry is driven forward at creep speed until the wheels are far enough away
to have no influence upon the deflection beam. The final dial gauge reading is recorded and the ‘rebound deflection’ is twice
the difference between the initial and final dial gauge readings.
Whichever method is adopted for the deflection beam measurements, the possible effect of plastic flow upon the results
should be noted, although this is only likely to be significant for thicker or relatively fresh asphalts. When an asphalt surfacing
material flows plastically, it squeezes upwards between the dual loading wheels of the deflection truck which, in the transient
deflection test, reduces the transient loading deflection because the upward movement of the material counteracts the
downward movement of the pavement. The transient recovery deflection that is measured may be correct but further plastic
movement of the raised surfacing material can occur during the time taken for the wheels to move from the test point to the
final position, thereby causing an error in the recovery deflection reading. It is usually very clear from the test results when
plastic flow occurs and testing should be stopped to avoid recording erroneous data.
In the rebound test greater plastic flow will be induced in susceptible materials because of the time the wheels remain
stationary over the test point. When the truck is driven forward the road surface ‘rebounds’ but an indeterminate amount of
recovery of the displaced surfacing material can occur. There is thus no clear indication from the simple rebound test when
plastic flow occurs.
E3.6 Analysis of deflection survey data

Deflection readings can be affected by a number of factors that should be taken into account before the results can be
interpreted. These are the temperature of the road, plastic flow of the surfacing between the loading wheels, seasonal effects
and the size of the deflection bowl.
E3.6.1 Road temperature

The stiffness of asphalt surfacings will change with temperature and therefore the magnitude of deflection can also change.
The temperature of the bituminous surfacing is recorded when the deflection measurement is taken, thus allowing the value
of deflection to be corrected to a standard temperature. It is recommended that 35oC, measured at a depth of 40mm in the

surfacing, is a suitable standard temperature. Fortunately, it is often found that little or no correction is required when the road
surfacing is either old and age hardened or relatively thin, which is likely to be the general case in this country.
If, however, there is a need to adjust for temperature (thicker, newer asphalt surfaces) the following should be noted. The
relation between temperature and deflection for a particular pavement is obtained by studying the change in deflection on a
number of test points as the temperature rises from early morning to midday20. As it is not possible to produce general
correction curves to cover all roads, it is therefore necessary to establish the deflection/temperature relationship for each
project.
E3.6.2 Seasonal effects

In areas where the moisture content of the subgrade changes seasonally, the deflection will also change. For overlay design
purposes, it is usual to use values that are representative of the most adverse seasonal conditions and it is therefore normal
practice to carry out surveys just after the rainy season. If this cannot be done, an attempt should be made to correct for the
seasonal effect. However, this requires a considerable data bank of deflection results and rainfall records before reliable
corrections can be made.
E3.6.3 Size of deflection bowl

The size of the deflection bowl can occasionally be so large that the front feet of the deflection beam lie within the bowl at the
beginning of the transient deflection test. If this happens, the loading and recovery deflection will differ. The simplest way to
check whether the differences in loading and recovery deflection are caused by the size of the bowl is to place the tip of another
beam close to the front feet of the measurement beam at the beginning of the transient test. This second beam can be used
to measure any subsequent movement of the feet of the first beam as the truck moves forward. If feet movements larger than
0.06mm are observed only the recovery part of the deflection cycle should be used for estimating the value of transient
deflection.
E3.6.4 Data processing

After all measurements have been made, and any corrections applied to the raw data, it is then convenient to plot the
deflection profile of the road for each lane. When measurements in both wheel paths have been made, only the larger
deflection of either wheel path at each chainage is used. Any areas showing exceptionally high deflections that may need
reconstruction or special treatment can then be identified.
The deflection profile is then used to divide the road into homogeneous sections, in such a way as to minimise variations in
deflections within each section. The minimum length of these sections should be compatible with the frequency of thickness
adjustments that can sensibly be made by the paving machine, whilst still maintaining satisfactory finished levels. When
selecting the sections the topography, subgrade type, pavement construction and maintenance history should all be conside-
red.
A number of statistical techniques can be used to divide deflection data into homogeneous sections. The recommended
technique is the cumulative sum method, where plots of the cumulative sums of deviations from the mean deflection against
chainage can be used to discern the sections. Details of this approach are given in Appendix H.
E4 FALLING WEIGHT DEFLECTOMETER (FWD)

E4.1 General
The Falling Weight Deflectometer (FWD) simulates the effect of actual traffic-induced loads by dropping onto the pavement
surface a constant weight from variable heights. A diagram is shown in Figure E3. The FWD is generally placed on a semi-
trailer and equipped with its own power source (generator / batteries). It weighs about 1 tonne (1.5 tonne for the airport
version) and can comfortably travel, on surfaced roads, at 100 km/h. A distance-measuring wheel is also attached to the semi-
trailer, for relative and global distance measurements.

Drop Weights a) FWD in travelling mode
Data transfer cables

Load Cell Distance measuring wheel
Road surface
Loading Plate Sensors beam Sensors (transducers)

(raised) Testing direction
Cyclic lift and drop
b) FWD in testing mode
Drop Height
150 150 Sensors beam

200 100 300 300 300 300 (lowered)
Figure E3 Diagrammatic representation of the FWD.
A number of detachable weights are locked on a hydraulic piston that facilitates their quick and precise lift. The weights are
thereafter dropped from a predetermined height. A circular, flexible, loading plate (150 mm radius) ensures the smooth load
transfer between the dropping weights and the potentially uneven pavement surface. A load cell, placed directly under the
dropping weight, accurately measures the loading level. The resultant pavement surface deflections are measured by 9
sensors / transducers placed under a sensors beam at the following offsets (from the loading plate’s centre): 0 / 150 / 200 /
300 / 600 / 900 / 1200 / 1500 /1800 mm. Multiple data transfer cables, also attached to the sensors beam, ensure the
communication between the load cell / sensors / FWD engines and the central computer.
E4.2 Measurement procedure

All the relevant “under traffic” test safety measures apply.
Establishing a concise but clear and consistent testing reference system prior to the commencement of testing is critical.
The reference system should include the following:
a) General information
Date, operator(s), FWD serial number, road ID (for network testing), measurement units (metric / imperial), test
start and end chainages, test spacing (distance between adjacent test points); sensors spacing (depending on the
pavement layer thicknesses);
b) Test point information and parameters
Number and sequence of drops (in terms of corresponding load levels); air, surface and in-depth temperatures; pave
ment cracking type, extent and magnitude; road profile (e.g. fill/cut, to reflect potential water ingress); change(s) in the
pavement structure; and underground structures (e.g. culverts, pipes, which can significantly affect the deflection
magnitude).

The number and sequence of drops can be set up differently in up to five series. The operator can apply any or all of these
series at a test point. Generally, one series of two drops (4 tonne each) is usually applied for all test points.
The air, surface and in-depth temperatures are usually determined at the start and end of a testing session. Should any
temperature change occur during testing, the operator should repeat FWD measurements. It is usually more beneficial to
continuously monitor the temperatures.
All relevant calibrations (see next section) must be undertaken as required. A large amount of deflection data could prove
incorrect and, therefore, useless should the system malfunction at any time.
While moving between two adjacent points the sensors beam must be raised, irrespective of travelling speed. Once the FWD
has stopped, the sensors beam is lowered together with the loading plate. The operator inputs the test point information and,
automatically, the weights are raised and dropped from a “test” height for an in-built, on-the-spot system check. Once the
operator is satisfied with the system pre-test data, the weights are automatically raised and dropped to and from the
predetermined height(s) as many times as required. After each drop, the relevant data is sent to the computer, which displays
it. The operator can interrupt the automatic testing sequence at any time and restart and /or continue it manually (drop by
drop), if so necessary.
E4.3 Calibrations
Three types of calibration of the sensors are done, namely absolute, reference and relative. Absolute calibration is done in the
factory, at the time of manufacture, while designated agents typically undertake reference calibration annually, also indoors.
The absolute and reference calibration results should be recorded by the agents in calibration certificates and be available for
inspection at all times.
The relative calibration is usually done monthly and/or at the start of every new project, in approximately 4 hours. During this
calibration, the sensors are placed one on top of each other and subjected to a standard vertical load. If all the sensors are
in good condition, their readings should be sensibly equal.
The load cell should be tested at the start and end of each testing session by plotting, on the computer screen, its output curve,
for a standard drop. This plotting option is available on most FWD equipment. If the load cell is in good condition, its output
curve shall have a continuous sinusoidal shape.
Generally, no other calibration is required, even when the equipment has to travel on rough roads or pull aside on grass.
E4.4 Output
At present, the FWD data acquisition software runs in an MS-DOS operating system. The testing output is stored in specific
text files with extensions such as: .f10, .f20 or .fwd (depending on the software type and version). These text files can be easily
viewed, in MS-DOS, with the “Edit file_name.f10/f20/fwd” command.
The general information is stored in the file’s header. The initial test drop parameters and results follow. Subsequently, all
drop loads and their corresponding pavement surface deflections are recorded separately, though grouped per test point.
The test point information is generally recorded per point, though it can also be recorded for each drop.
Microsoft Excel can be satisfactorily used to import and process these files. Once the potential user tries to open such a file,
the text import procedure is automatically started and a “delimited” file type is assigned. As this default file type is convenient,
the user can subsequently choose the delimiter type for converting the text to columns. For the purposes of FWD data
analysis, the most adequate delimiter type is ”space”. The resultant file can be saved as an Excel spreadsheet and used for
further processing.
It is however recommended that the FWD service provider processes these files according to the Client’s requirements.
Inherent measurement errors can be easily overlooked if the processing personnel do not have the required expertise.
E4.5 Deflection bowl parameters

The deflections recorded at a test point constitute that point’s deflection bowl (see Figure E4a). The maximum deflection,
measured directly under the load, can serve as a good indicator of the overall pavement strength. The inner deflections
(closer to the loading plate) relate to the upper layers (surfacing, base, subbase) strength whereas the outer ones relate to the

lower layers (selected, subgrade). For this reason, a number of deflection bowl parameters are derived from measured
deflections (see Figure E4b).
A number of software packages (e.g. IDMP) have been specifically developed for FWD data conversion and processing. The
output includes deflection bowl parameters, allowable traffic, etc.
E4.6 Backcalculation
Currently complete deflection bowls are used in an iterative procedure, known as backcalculation, to estimate the pavement
layer E-moduli. The straightforward goal of the backcalculation process is to estimate a set of layer E-moduli that best match
the measured and calculated deflections, at all offsets.
A physical model is assumed, with estimated E-moduli and Poisson’s ratios. The layer thicknesses are considered known.
A set of theoretical deflections is then mathematically derived (at the same offsets as the FWD sensors) based on the
estimated E-moduli and the traffic loading. This set of computed deflections is compared with the FWD measured one.
Based on the difference between the two sets of deflections, the estimated E-moduli are adjusted and the theoretical
deflections re-computed. This process is iterated until the difference between the computed and measured deflections is
being reduced to a minimum (that is, 5-10%).
SENSOR OFFSET (mm)
0 200 300 600 900 1200 1500

D6
DEFLECTION (micron)
D5
Strength of subgrade
(Overall pavement strength)
D0 - Maximum deflection
D4
D3
D2
Lower Layer Index (LLI)

- Strength of selected & subgrade.
D1
Middle Layer Index (MLI) - Strength of subbase & upper selected.
Base Layer Index (BLI) - Strength of surfacing & base
a) Deflection Bowl Parameters
SENSOR OFFSET (mm)
0 200 300 600 900 1200 1500
Strong subgrade
Strong base (Low D7)
DEFLECTION (micron)
(Low BLI)
Weak subgrade
(High D7)
Weak base
(High BLI)
b) Pavement Layers Strength
Figure E4 Deflection bowl parameters.

Appendix F: Test pit profiles
0
SI moist, dk brown, dense, compacted, sandy

gravel, crushed stone (felsite) base.
150
Very moist, white, medium dense, compacted

silty-sandy gravel, nodular calcrete.
300
Dry, light grey, very dense, compacted sandy

gravel, stabilised nodular calcrete.
450
Moist, brown, soft, intact sandy clay, in situ

transported top soil with roots.
620
Example of a soil profile prepared in Microsoft Excel.

Appendix G: Equilibrium and predicted
moisture content
As discussed in the main text, the in situ moisture content can be evaluated in terms of expected equilibrium moisture
contents (EMC) and predicted equilibrium to optimum moisture content ratios (EMC/OMCm)2. Various models are available for
the former, the simplest being:
Equation G.1
EMC = 0.68 (OMCm) + 0.031 (LS)(P4250.7) – 1.1(K) %
where OMCm = Mod AASHTO optimum moisture content

LS = bar linear shrinkage,
P425 = percentage passing 0.425 mm sieve,
and K = 1 for base and 0 for subbase and subgrade.
The EMC/OMCm ratio in Botswana will differ from layer to layer as follows:
All unbound subgrade materials

Equation G.2
EMC/OMCm = 0.13(LS) + 0.63(loge(100 + Im)) + 0.13(P75/OMCm) –
0.0071(LS)(P4250.7) + 0.011(PL) – 2.36 %
where Im = Thornthwaites moisture index,

P75 = percentage passing 0.075 mm sieve,
PL = Plastic limit, and
the other parameters are as defined previously.
Subbase
EMC/OMCm = 0.707
Base course
EMC/OMCm = 0.56

Appendix H: Cumulative sum method for
identifying uniform sections
H1 Introduction
Interpretation of the field data to define sections having uniform characteristics in terms of a given parameter (e.g. roughness,
deflections, etc) can sometimes be aided by the use of the cumulative sum (Cusum) approach. This is not always applicable,
although it can be attempted with any data set, but is most effective when data have been collected at regular and fairly
frequent intervals. Hence, it is most often applied to deflection, roughness and rut depth measurements.
From the outset, it should be kept in mind that this approach is just one of various methods that could be used, but that the
principal aim is to help delineate uniform sections. Thus if a graphical inspection of the data values already gives clear
practical delineation of sections, there is actually no need to use further processing.
It should also be noted that some experience helps in expediting the process where, for example, the data values can be
seen as being combinable to define a practical uniform section without concern for minor differentiations. An important point,
however, is that the standard deviation is normally used to define characteristic (typically 90th percentile) properties and this
can become unrealistically high if data which are significantly disparate are combined for a “uniform” section.
Consequently, the final check on reasonableness must always be a visual comparison of the calculated characteristic value
for a section with the raw data values. This will show whether the delineation is realistic.
H2 Cumulative sum definition

The cumulative sum is calculated in the following way.
Equation H.1
Si = xi - xm + Si-1
where xi = Parameter value at chainage i

xm = Mean parameter value for the whole data set
Si = Cumulative sum of the deviations from the mean parameter value at chainage i
Si-1 = Cumulated sum up to the previous position (chainage i-1)
H3 Using the cumulative sum method

Using cumulative sums, the extent to which the measured individual parameter values on sections of road vary from the
mean value of the whole surveyed length can be determined. Distinct changes in the slope of the line connecting the plotted
cumulative sums will indicate boundaries between homogeneous sections.
This will be seen more clearly in the examples below, and arises from the fact that relatively uniform sections will have
parameter values that differ reasonably consistently from a selected datum value, in this case the overall mean. Thus a
uniform section having higher than average values, for example, will cumulate an increasing positive divergence from the
mean, and one with lower than average values will give an increasing negative divergence. The more consistent the uniform
section values, the closer the divergences approximate to a straight line.
Consequently the Cusum method can clearly show the points at which changes in the field data values would suggest
different uniform sections.

H4 Final processing to define characteristic values of
uniform sections
The final stage of the procedure is to calculate the representative parameter value for each identified homogeneous section
of the road. This entails calculating the mean and standard deviation for each uniform section. These are then used to define,
most commonly, the 90th percentile value, which can be calculated as follows:
Representative value = mean + 1.28 standard deviation.
If such a procedure is used for parameters in which lower values represent poorer conditions (such as material CBRs), the
representative value will be given by subtraction of the standard deviation component.
Other percentile values could be used, which would use a different multiplier for the standard deviation. These come from
statistical tables (based on the assumption of normal, or Gaussian, distribution of the data values) with other commonly used
values being the 80th percentile and the 95th percentile. Multipliers associated with these different percentiles are 0.84 and
1.64 respectively. The proposed method will tend to separate out areas of poorer condition or areas that warrant special
treatment or reconstruction and therefore the distribution of the remaining parameter measurements will approximate to a
normal distribution
H5 Examples of the use of the Cusum method

This section gives some examples on the use of the Cusum method, and highlights some of the important points to note
when using such a method.
H5.1 Example 1, Plasticity Index (PI) values

Figure H.1 gives a complete data set for PI values determined over a 17 km section, in which one data value has been
measured for each one kilometre section. These are shown as the first two columns of data, giving the position (chainage)
and the PI value (xi). The average for all the PI values (the data set) is given at the foot of the data column: 24.41 (xm).
The next column then gives the difference between each individual data value and the mean (xi – xm), and the last column gives
the cumulative sum of the differences. The first graph to the right of the data set simply shows the data values plotted against
chainage on an x-y basis, and the lower graph shows the Cusum plot.
There are several things to note. First, this method of plotting
Cumulated
PI Difference differences
Chainage (xi) xi - mx (Cusum) Data values (PI)
1 20 -4.41 -4.41
2 18 -6.41 -10.82 70
3 7 -17.41 -28.24
4 20 60
-4.41 -32.65
5 22 -2.41 -35.06 50
6 24 -0.41 -35.47
7 30 5.59 -29.88 40
8 55 30.59 0.71
9 60 35.59 36.29 PI 30
10 45 20.59 56.88 20
11 62 37.59 94.47
12 10 -14.41 80.06 10
13 1 -23.41 56.65
14 3 -21.41 35.24 0
15 1 -23.41 11.82 0 5 10 15 20
16 2 -22.41 -10.59 Chainage
17 35 10.59 0.00
Average, mx = 24.41
Cumulative sum for PIs
100
50
-50
0 5 10 15 20
Chainage
Figure H.1 Data and plots for Example 1

the raw data, while simple to do, is somewhat misleading in that each value is shown at a specific chainage value. For this
type of data it is assumed that any PI value in fact represents the typical value over the whole sub-section of road, i.e. that the
first value (20) applies from chainage 0 to 1, the second (18) from chainage 1 to 2, and so on. Thus it would probably be more
realistic to either show the data values at the mid-point chainage positions (i.e., 0.5 km, 1.5 km, etc) as an x-y plot, or
alternatively as bars from chainages 0 to 1, 1 to 2, 2 to 3, etc. While this is not critical, it is important that the significance of the
data representation is understood, as it will affect subsequent processing.
The second point to note, from the raw data value plot, is that from observation it is clear that there are at least four likely
divisions into uniform sections. It can be seen that, with the exception of the significantly lower value for chainage 2 to 3 (PI
7, plotted at chainage 3), the PIs up to km 7 are effectively in the range 20 to 30. This immediately highlights the need for a
further consideration of the data values, and clarification of the validity and representativeness of the lower value. This is
because the practical difference between a PI of 7 and PIs of 20 to 30 is highly significant. The next section, plotted at points
km 8 to 11, has much higher PIs (45 to 62), with those for km 12 to 16 low, and a last high point being PI 35 for km 17.
The Cusum representation (the lower graph) shows two distinct nominally linear sections, from km 7 to 11, and from km 11
to 16, with a distinct change in gradient. These represent the sections having the high and the low PIs noted above. The
gradients of the lines indicate that the first of these sections, with a relatively high positive gradient, is reasonably uniform with
higher than average PIs (positive gradient). Conversely, the second of these sections has a distinct negative gradient,
indicating lower than average values.
The first part of the Cusum plot shows a downward trend line to km 7, with a slightly higher dip from km 2 to 3, due to the PI
value of 7 for the section km 2 to 3. As the gradient for this section, apart from the dip, is reasonably uniform and slightly
negative, the plot indicates that the data values are typically somewhat lower than the overall average.
It might therefore be inferred from the Cusum plot that uniform sections could be taken as 0 to 7 km, 7 to 11 km, and 11 to 16
km, with the last single point (PI 35) indicating a final section. Two factors should be considered prior to making such a
conclusion: firstly, is this a realistic delineation; secondly, which values should be averaged to provide characteristic data
(including standard deviations).
Reviewing the raw data value plot provides this basis, and immediately indicates the need for caution when interpreting the
data. It is apparent that the low PI value for km 3 (representing km 2 to 3) could represent a totally different condition to the rest
of the section. Consequently, it should not be included in the averaging process as it will tend to give a lower (unrealistic)
average for the section generally, with a higher standard deviation (indicating greater scatter). As highlighted earlier, further
investigation of the validity of this value and possible extent of the different condition is actually needed.
Thus for the first uniform section, which in fact covers km 0 through to km 7, an average for characterising the section would
use raw data for chainages 1 to 7 in the table, excluding the value for km 3.
The second section (km 7 to 11 from the Cusum plot) represents very high PI but it is clear that the actual value for km 7 should
not be included in the average.
The third section indicated (km 11 to 16) would not include the value for km 11, in line with the previous comments, and also
calls for some judgement. The reason for this is that four of the data values (representing actual chainage from km 12 to 16
inclusive) are very low (PIs from 1 to 3), while the value for point 12 (representing actual chainage from km 11 to 12) is
markedly higher (PI 10). In reality, this difference is significant and clarification of the higher PI would be required while the rest
of the section could be characterised by averaging the lower uniform values.
Now this simple example, which is not perhaps a typical practical application, has been used to highlight the importance of
some understanding of the data, and to avoid blind usage of a method to aid practical delineation of data sets. Similar
practical consideration should be given to any data processing, but is especially true when data are limited as in this
example, and are effectively already representative values for substantial lengths. Other practical cases where this might
occur are for delineating sections based on CBR data or on roughness data when each data value is regarded as representing
finite lengths of the road.
H5.2 Example 2, deflection data values

This example is where each value actually represents a distinct point on the road where the measurement was taken. In this
case, only a few of the actual readings taken are shown (there are actually several thousand data points) for deflection data
obtained by Deflectograph for the Rasesa – Monemetsana section.
First of all this illustrates that the amount of data is immaterial once it is in a spreadsheet: processing follows the same
procedure and the time taken is little different as the simple formulas are copied for all data. Second, it is probably a more
typical application but, again, it should be emphasised that the Cusum process is just an aid to delineating uniform sections
and should not be considered other than a guide. Final classification will still normally require some judgement.

Figure H.2 gives some of the data for example 2, together with a complete plot of the data values over almost 25 km, and the
Cusum graph. First it should be noted that the data has already been processed ready for graphical representation. In this
case, this means all the chainage values have been corrected to tie in with kilometre posts, and the actual deflection values
for the lefthandside (LHS) beam have had 1.5 mm added. The last process is simply an expedient to show both beam values
against the y-axis rather than have them effectively plot on top of each other. Thus the zero deflection datum for the upper plot
(the LHS beam values) is actually 1.5 mm.
It should also be noted that the average used in the Cusum analysis is simply the average of the plot values, including the
added 1.5 mm on each value. This has no effect whatsoever on the Cusum process. The only important point is that any
subsequent averaging and evaluation of standard deviations must be done using the real data values.
As in the previous example, the numeric data is given as position (x-axis value), and in this case two columns of deflection
data (y-axis values), with the LHS values being the real value plus 1.5 mm as just discussed. The next two columns then show
the cumulative sum data: the average values of the two data sets (to which each individual value is compared) are at the top
of these columns. These are shown as 0.24663364 and 1.724723805 for right-hand and left-hand sides respectively, and are
simply the spreadsheet function averages. These would normally be notated as 0.25 mm and 0.22 mm (when 1.5 mm is
subtracted for the LHS beam) in any discussion. However, for data processing purposes the use of these values (showing 9
decimal places) is perfectly acceptable.
All the deflection data are shown on the upper plot and, in line with the very low overall average deflections, it will be seen that
most deflections are less than half a millimetre. The graph also clearly shows that both beams give similar patterns as would
be expected. It should be noted that there is actually a gap in the data from about km 54.3 to km 55, although a line is shown
joining the values. The reason for the gap should be established in actual processing but is outside the scope of this
example, which is only to provide insight into the use of the Cusum method.
The lower graph gives the Cusum results. This definitely gives a very distinct delineation of several major sections, the first
major change occurring at about km 44.9 (the actual point can be identified from the plot using the latest version of Excel, or
from examination of the data values). Reference to the upper plot suggests that this would satisfactorily define a uniform
section. Other distinct changes (ignoring the data gap from km 54. to km 55) are evident at approximately km 50.8, km 51.6,
km 53.6, and km 58.8.
Prior to detailed analysis, it should be recalled that gradients of the Cusum plot reflect relative deviations from the mean and,
most important, that similar gradients represent similar data values. This fact provides further assistance in selection of
uniform sections.
For this example, the first section is well defined to km 44.9 and, ostensibly, a second section would be to approximately km
50.8. A closer examination of the Cusum plot, however, indicates that apart from the slight rise in gradient from km 50.8 to
about km 51.6, the overall gradient from km 44.9 to km 53.6 is similar. This suggests that, although there is a definite need to
treat the section from km 50.8 to km 51.6 as a different uniform section, if these values were excluded from the calculation, it
would be appropriate to consider all the other data from km 44.9 to km 53.6 as a uniform section.
The implication is that, instead of averaging two separate sets of data on each side of km 50.8 and 51.6, only one set is used
once the values pertaining to this short intermediate section are excluded. While there is technically nothing wrong in treating
them as different data sets, which are likely to give similar results, the proposed approach appears more realistic: the section
is basically uniform with a short intermediate section having significantly higher deflections which might relate to some
specific field conditions. This would be clarified from other survey data.
Extending this reasoning, it will also be seen that the last section from about km 59.0 also has much the same gradient
(confirmed by straight edge or ruler), apart from a very localised deviation at about km 62.2, indicating similar conditions.
However, as this is already a substantial length it can be treated separately even though the results should be similar.
Other sectioning from the Cusum plot is suggested from about km 53.6 to km 55.6 (ignoring the slope change where there is
no data, but looking at the gradient, and the raw data), and from about km 55.6 to km 59.0.
This would give a total of nine uniform sections based on deflection, including the very localised high deflection area around
km 62, as shown in Table H.1 following. It should be noted, however, that in this case given the general low deflections from
km 44.9 onwards, a more practical approach might be to regard the whole section as uniform with the exception of the
significant higher deflection sections. These would be from km 50.8 to 51.6 (section 3), km 53.6 to 55.6 (section 5) and around
km 62.2 (section 8). The results from this delineation are also given in Table H.1 for comparison.

Data for graph Average y1 Average y2
LHS beam 0.246663364 1.724723805
Position RHS beam plus 1.5mm Cusum 1 Cusum 2
x-axis y1 y2 y1 y2
37.9830585 0.49 2.33 0.243336636 0.605276195
37.98870567 0.34 2.26 0.336673272 1.14055239
37.99435283 0.41 2.23 0.500009908 1.645828586
38 0.4 2.16 0.653346544 2.081104781
38.00564717
38.01129433
38.0169415
0.4
0.41
0.38
2.19 0.806683181 2.546380976
1.91 0.970019817 2.731657171
1.85 1.103356453 2.856933366
Small part of the actual
38.02258866
38.02823583
0.32
0.45
1.84 1.176693089 2.972209562
1.79 1.380029725 3.037485757
data set,
38.03388299
38.03953016
38.04517732
0.52
0.5
0.48
1.77 1.653366361 3.082761952
1.77 1.906702997 3.128038147
1.84 2.140039633 3.243314342
prepared for plotting
38.05082449 0.45 1.91 2.34337627 3.428590538
38.05647165 0.38 1.79 2.476712906 3.493866733
38.06211882 0.49 1.75 2.720049542 3.519142928
38.07341315 0.37 1.73 2.843386178 3.524419123
38.07906031 0.32 1.83 2.916722814 3.629695318
38.08470748 0.38 1.83 3.05005945 3.734971514
38.09035464 0.28 1.85 3.083396086 3.860247709
38.09600181 0.35 1.95 3.186732722 4.085523904
38.10164897 0.33 1.91 3.270069358 4.270800099
38.10729614 0.35 1.96 3.373405995 4.506076294
38.1129433 0.37 1.88 3.496742631 4.661352489
38.12423763 0.45 2 3.700079267 4.936628685
38.1298848 0.34 2.05 3.793415903 5.26190488
38.14117913 0.31 1.99 3.856752539 5.527181075
38.14682629 0.29 1.94 3.900089175 5.74245727
38.15247346 0.27 1.89 3.923425811 5.907733465
38.15812062 0.4 1.8 4.076762447 5.983009661
………etc
Rasesa - Monemetsana
Processed deflection data - complete surveyed length
3
2.5
2
Millimetres
1.5
0.5
0
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
Corrected chainage (km)
Rasesa - Monemetsana
Cusum for deflection data - complete surveyed length
120
100
80
Section 3
Section 5
Millimetres
60
40 Section 8
20
Section 1 Section 2 Section 4 Section 6
0
Section 7 & 9
-20
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
Corrected chainage (km)
Figure H.2 Deflection data example.

Mean deflection and standard deviation (mm)
Initial delineation Alternative processing
Section number/chainage LH S RHS LH S RHS
1: start to km 44,86 0,34 (0,13) 0,35 (0,12) 0,34 (0,13) 0,35 (0,12)
2: km 44,86 - km 50,79 0,16 (0,07) 0,19 (0,06) 0,17 (0,07) 0,20 (0,07)
3: km 50,79 - km 51,62 0,25 (0,07) 0,30 (0,08) 0,25 (0,07) 0,30 (0,08)
4: km 51,62 - km 53,59 0,16 (0,07) 0,19 (0,06) 0,17 (0,07) 0,20 (0,07)
5: km 53,59 - km 55,64 0,32 (0,09) 0,34 (0,09) 0,32 (0,09) 0,34 (0,09)
6: km 55,64 - km 59,04 0,22 (0,07) 0,24 (0,07) 0,17 (0,07) 0,20 (0,07)
7: km 59,04 - km 62,19 0,16 (0,05) 0,17 (0,05) 0,17 (0,07) 0,20 (0,07)
8: km 62,19 - km 62,44 0,21 (0,08) 0,20 (0,07) 0,21 (0,08) 0,20 (0,07)
9: km 62,44 - km 63,39 0,16 (0,05) 0,17 (0,05) 0,17 (0,07) 0,20 (0,07)
Comment Data combined for Data combined for

sections 2 and 4, and sections 2, 4, 6, 7 and 9
for sections 7 and 9
Table H.1 Possible uniform sectioning for deflection data.
The processing requires identifying the actual point at which the uniform section ends by direct examination of the Cusum
data sets. For example, the first section is marked by a Cusum peak of about 110 on one beam and about 104 on the other.
Paging through the data will quickly find where these values peak to identify the chainage and, the row number. This will
define the range extents for averages and standard deviations.
The most efficient method for processing is then firstly identifying all change points, in terms of chainage and spreadsheet
data row number, and only then calculating means and standard deviations. The reason for this is that all the calculation
functions can be kept together and simply edited for the appropriate data ranges, rather than placing these all through the
spreadsheet adjacent to the relevant tabulated spreadsheet data.
Without going into great detail on the specific values, it will be seen that the RHS values are generally the higher and in fact
represent the outer wheel path. The higher values will in any case be used to define the characteristic deflections. For the
particular data in this example, which are actually very low deflections overall, either method of processing in Table H.1 is
basically sound. For this example, however, the initial delineation will be used to define the uniform section characteristics
shown in Table H.2

Deflection (mm)
Section no. Mean Standard Representative

deviation deflection
(90th percentile)
1 0,35 0,12 0,50
2 0,19 0,06 0,27
3 0,30 0,08 0,40
4 0,19 0,06 0,27
5 0,34 0,09 0,46
6 0,24 0,07 0,33
7 0,17 0,05 0,23
8 0,21 0,08 0,31
9 0,17 0,05 0,23

Table H.2 Summary of deflection analysis.
The best way to review the selection is to show the characteristic uniform section data superimposed on the actual deflection
values, which readily confirms the validity of the selection or indicates further processing is required. This is shown in Figure
H.3, in this case showing only the outer wheel path deflections (RHS), and on a larger vertical scale.
Figure H.3 Outer wheel path data and 90th percentile characteristic deflections.

Figure H.3 generally confirms the selection of uniform sections based on deflection data, but also shows that there are a
series of significantly higher deflections near the start of section 1 (around km 38.4 to km 39 or so) that may need closer
examination.
The result of the deflection survey will in any case be considered in conjunction with other data such as:
(I) rut depths
(II) layer thicknesses and layer strengths (typically determined with a DCP)
(III) assessments of drainage conditions, and
(IV)visual assessment of condition.
Thus the prevailing condition of the section can be fully evaluated.

Appendix I: Pavement condition rating form
N Culvert Fort Jackson I/C
Eastbound C/W Eastbound C/W

Features
King Williams Town Westbound C/W Westbound C/W
Distance (km) 28 29 30 31 32 33 34 35 36 37 38 39 40
Highway Section N2/15

Uniform Section 15 W1 15 W2
Reseal
Surfacing 30 AS (1986)
Pavement
Structure
Upper Base 100 ETB (1986) (Insitu Qtz 1%Bit./1%Lime) 175 ETB (1986) (Insitu Qtz,1%Bit/1%Lime) 100 ETB (1986)
Lower Base 100 ETB (1986) (Insitu Qtz 1%Bit./1%Lime) 100 ETB (1986) (B / L)
Subbase C3 150 G5 (1986) 150 C3 (1986) (Insitu Qtz/Shale, 3% OPC) 150 C3 (1986) 150 G2 (1972)
Upper Selected 150 G5 (1972) 150 C3 (1986) (Insitu Shale /Dol, 3% OPC) 150 G5 (1972)
Lower Selected 150 G5 (1972)
CTO Traffic Data

(AADT/Heavy/%/Daily E80 per 7351 / 337 / 5% / 315 / 1997
lane/Year of Count)
Fast shoulder
Section
Cross
Fast lane
Slow lane
Slow Shoulder
Distance (km) 28 29 30 31 32 33 34 35 36 37 38 39 40
Roughness
measurements
(1998)
Instrument
Rutting
(1995 , update 1997)
Deflection
(1993)
Aggregate loss
Bleeding
Longitudinal/Transve
rse cracking
Visual assessments
Crocodile cracking
Pumping
Failures/Patching
Deformation
Drainage
Summary of
Test Pit
Results
Distance (km) 28 29 30 31 32 33 34 35 36 37 38 39 40
Legend : Condition Ratings :

Good Warning Severe
Figure II Pavement fondition rating form.

Appendix J: Definition of terms
Asphaltic concrete A group of hot bituminous mixtures used for surfacing. They normally consist of a
well graded mixture of coarse aggregate, fine aggregate and filler, bound together
with penetration grade bitumen.
Base course The layer(s) occurring immediately below the surfacing and above the subbase or, if
there is no subbase, above the improved layers.
Behaviour The function of the condition of the pavement with time (see also performance).
Bitumen emulsion A binder in which bitumen has been dispersed in finely divided droplets in water by
the aid of mechanical means and an emulsifying agent. Bitumen emulsion is made in
an anionic and a cationic type depending on the particle charge of the bitumen in the
material and the physical properties related to their behaviour during construction,
(See also break).
Bitumen stabilized material A material made of natural- or crushed aggregate with a bituminous binder admixed.
Used in pavement layers – primarily for base course.
Bitumen rubber A binder in which bitumen is modified with more than 15% ground rubber. (See also
modified binder).
Bituminous seals A general term for thin bituminous wearing courses made of surface treatments or
slurry seals, or a combination of these.
Borrow pit A borrow pit is a site from which natural material, other than solid stone, is removed
for use in construction of the works. The term borrow area is also used.
Break of emulsions ‘Break’ of a bitumen emulsion is when the water and bitumen separates so that the
water will evaporate, leaving behind the bitumen to perform its function.
Buses All buses with a seating capacity of 40 or more.
Crushed stone Crushed stones. Min 50% by mass of particles larger than 5 mm shall have at least
one crushed face. Made from crushing of stones, boulders or oversize from natural
gravel. Max 30% of the fraction passing the 4,75 mm sieve can be soil fines. The
material is compacted to a specified relative density of BS-Heavy.
Curing membrane A bituminous binder, usually made of bitumen emulsion, applied immediately after
construction of a completed surface of modified or stabilized materials with lime or
cement. Its purpose is to prevent early drying out of the cemented layer and to minimize
adverse effects of the stabiliser’s contact with CO2 in the air.
Cutback bitumen A penetration bitumen which viscosity has been temporarily reduced by blending
with solvents. The solvents are expected to evaporate during the early part of the
pavement’s service life. Classified according to their viscosity.
Cutting A cutting is a section of the road where the formation level is below the original
ground level.
Deflection The recoverable vertical movements of the pavement surface caused by the application
of a wheel load.
Deformation A mode of distress, unevenness of the surface profiles.
Degree of distress A measure of severity of the distress.
Distress The visible manifestation of deterioration of the pavement with respect to either the
serviceability of the structural capacity.
Dynamic Cone Penetrometer (DCP) An instrument for assessing the in-situ CBR strength of granular materials/soils.
Earthworks A general term describing all processed materials below formation level including
improved subgrade layers, fill and prepared roadbed.
Embankment An embankment is a section of the road where the formation level is above the
original ground level.

Embankment, shallow A shallow embankment is defined as a section of the road where the formation level is
between 0 and 0.3 m above the original ground level.
Equivalent standard axle (E80) Defined as an axle loaded to a weight of 8 160 kg, in the design concept meaning a unit of
measuring the damaging effect to road pavements caused by axles of any load.
Fill Material placed below the improved subgrade, but above the roadbed.
Fogspray A light application of bitumen emulsion, sprayed on top of surface dressings. Its purpose
is to improve retention of the aggregate in new seals. On old roads its purpose is to arrest
any loss of chipping and to water-proof and rejuvenate the bituminous surfacing.
Formation level The final level upon which the pavement layers are placed.
Geometric life The life of a road before the functional capacity to carry additional traffic is exceeded.
Granular materials Pavement materials made from crushed or natural sources, where no addition of any
stabilizer has been made. (Term NOT to be used: Unbound materials).
Gravel wearing course The uppermost layer of a gravel road, which provides the riding surface for vehicles.
Highway Design and Maintenance
Standards Model (HDM) HDM-3 and HDM 4 are computerized management systems for road networks developed
by the World Bank.
Heavy goods vehicles (HGV) All goods vehicles having 3 axles.
Heavy vehicles A general term describing vehicles with un-laden weight of 3 tonnes or more. Heavy
vehicles are further sub-grouped into Medium Goods-, Heavy Goods- Very Heavy
Goods Vehicles and Buses for the purpose of determining design load in pavement
design.
Improved subgrade The uppermost layer(s) of the subgrade, consisting of material of controlled quality.
(e.g. terms not be used: selected borrow – selected subgrade – capping layer – topping).
Life cycle cost The cost of a pavement structure or surfacing discounted over its expected service-
able life.
Light vehicles A general term describing vehicles with un-laden weight of less than 3 tonnes and
includes buses with a seating capacity of less than 40.
Medium goods vehicles (MGV) All goods vehicles having 2 axles and an un-laden weight of 3 tonnes or more.
MERLIN Simple apparatus to measure road roughness.
Modified binder A binder in which bitumen is modified with a prescribed percentage of polymers or
other approved chemical constituents, alternatively with less than 15% ground rubber
(See also bitumen-rubber).
Modified material A material where the physical properties have been improved by the addition of a
stabilizing agent but in which strong cementation has not occurred.
Natural gravel Material from natural gravel sources. The term also includes crushed material where
less than 40% of the mass of particles larger than 5 mm have a crushed face. Classified
according to their minimum CBR strength. Used in pavement layers.
Natural gravel/soil Material from natural sources. Classified according to their minimum CBR strength.
Used in improved subgrade layers and fill.
Pavement behaviour The function of pavement condition with time.
Pavement evaluation The assessment of the degree to which the pavement fulfils its structural and functional
requirements.
Pavement layers The combination of material layers constructed above the formation level in order to
provide an acceptable facility on which to operate vehicles.
Performance The measure of satisfaction given by the pavement to the road user over a period of
time, quantified by a serviceability/age function (see also behaviour).

Prime An application of low viscosity bituminous binder to an absorbent surface, usually the top of
the base course. Its main purpose is to protect the surface of a granular material during con-
struction and to improve the bond between granular materials and bituminous mixes or seals.
Quarry A quarry is an open surface working from which stone is removed for use in construction
of the works.
Reflection cracks Cracks in asphalt overlays or surface treatment that reflect the crack pattern of the pave
mentstructure underneath.
Rehabilitation design period The chosen minimum period for which a pavement rehabilitation is designed to carry
the traffic in the prevailing environment, with a reasonable degree of confidence,
without necessitating further pavement rehabilitation.
Roadbed All in-situ ground after bush clearing, removal of topsoil and excavation of any cut tings, and
before placing any layers, whether these layers are fill, improved subgrade or pavement
layers.
Sand seal A surface treatment made of sand aggregates of crushed or natural material. Can be
constructed in single- or multiple layers.
Serviceability The measure of satisfaction given by the pavement to the road user at a certain time,
quantified by factors such as riding quality and rut depth.
Skid resistance The general ability of a particular road surface to prevent skidding of vehicles.
Slurry seal A cold premixed material of creamy consistency in a fresh state, made of crusher-dust,
bitumen emulsion and cement filler. Water is added for adjustment of the consistency. If
constructed in combination with a new surface dressing, it is named a Cape seal.
Structural capacity The ability of pavement to withstand the effects of climate and traffic loading.
Structural design The design of pavement layers for adequate structural strength under the design
conditions of traffic loading, environment and subgrade support.
Structural distress Distress pertaining to the load bearing capacity of the pavement.
Structural evaluation The assessment of the structural capacity of pavement.
Subbase The layer(s) occurring below the base course and above the improved sugrade layer.
Subgrade The completed earthworks within the road prism before the construction of the pave
ment layers.
Surface dressing A surface treatment made of single sized aggregates of crushed material. Can be
constructed in single- or multiple layers.
Surface treatment A general term for thin bituminous wearing courses made by lightly rolling aggregate
into a sprayed thin film of bitumen. Aggregates can be made of crushed or natural
material with a grading depending on the desired type of surface treatment to be
produced. Can be constructed in single or multiple layers.
Surfacing integrity A measure of the condition of the surfacing as an intact and durable matrix (it includes
values of porosity and texture).
Surfacing bituminous The uppermost pavement layer(s), which provides the riding surface for vehicles.
Includes bituminous wearing course and bituminous binder course where used.
Tack coat An application of bituminous binder to a bituminous surface subsequent to placing a
bituminous layer. Usually made of bitumen emulsion with the purpose to improve
the bond between bituminous layers.
Terminal level A minimum acceptable level of some feature of the road in terms of its serviceability.
Types of distress The sub-classification of the various manifestations of a particular mode of distress.
Vehicle equivalency factor The total number of equivalent standard axles calculated for one vehicle. The average of
all these values within one vehicle category is subsequently calculated for ease of refe
rence to traffic count data.
Very heavy goods vehicle (VHGV) All goods vehicles having 4 axles or more.

Pavement Testing

Uploaded by

Copyright:

Available Formats

Pavement Testing

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pavement Testing

Uploaded by

Copyright:

Available Formats

MAY 2000

ISBN 99912 - 0 - 315 - X

2 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 3

Gaborone 15th May 2000

4 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

Photographs were provided by:

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 5

6 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 7

6 RUT DEPTH MEASUREMENT

9 TEST PIT PROFILING AND SAMPLING

PART C PAVEMENT DATA ANALYSIS

8 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

PART D REFERENCES AND

LIST OF TABLES page

Table 2.1 Composition of typical Botswana bituminous pavements......18

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 9

LIST OF FIGURES page

10 Roads Department Pavement Testing, Analysis and Interpretation of Test Data

Part A General 2 Road Pavements

4 Planning and selection

Part B Evaluation 6 Rut depth measurement

9 Test pit profiling and

Part C Pavement 10 Data analysis

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 11

Part A General 2 Road Pavements

4 Planning and selection

Part B Evaluation 6 Rut depth measurement

9 Test pit profiling and

Part C Pavement 10 Data analysis

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 13

1.2 Purpose and scope of guideline

14 Roads Department Pavement Testing, Analysis and Interpreta of Test Data

PART B (Practical aspects)

PART C (Analytical aspects)

The process discussed in this guideline is summarised in the flow chart

Other sources Strength Profiles

PART D (References and Appendices)

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 15

2.1 The road function

2.2 Principles of road maintenance

16 Roads Department Pavement Testing, Analysis and Interpreta of Test Data

2.3 Typical pavement structures

Cut back slope Roadway

Original ground level

Cutting Pavement layers Formation level

Figure 2.2 Cross section terms and elements.

Pavement Testing, Analysis and Interpretation of Test Data Roads Department 17

Surfacing Asphalt, bituminous surface treatment

Crushed stone, natural gravel, stabilised

Crushed stone, natural gravel, stabilised aggregate or

Selected layers Natural or stabilised gravel

SAND SEAL SINGLE OTTA

Fig 2.3 Schematic illustration of various types of bituminous surfacings.

18 Roads Department Pavement Testing, Analysis and Interpreta of Test Data

Asphalt (hot mix) 10 - 14 15 - 20

Double seal 6-8 9 - 13

Single seal 4-6 5-9

Double Otta seal 7 - 10 12 - 15