Professional Documents
Culture Documents
Cost336 Report Use of Falling Weight Deflectometers in Pavement Evaluation
Cost336 Report Use of Falling Weight Deflectometers in Pavement Evaluation
COST 336
Use of Falling Weight Deflectometers in Pavement Evaluation
Final Report of the Action
Main Report
2nd Edition – April 2005
European Commission
Directorate General Transport
Preface
First Edition
COST Action 336 'Falling Weight Deflectometer' officially started in 1996. The project was a continua-
tion of the Falling Weight Deflectometer working group of the Forum of European Highway Research
Laboratories (FEHRL) that started in 1991. The goal of COST Action 336 was to develop a European
common code of good practice for the use of Falling Weight Deflectometers (FWDs) in pavement
evaluation. This involved:
• Development of a harmonisation proposal for the evaluation of flexible pavements at project level using
FWD-tests. This deliverable was seen as an update and an expansion of FEHRL Report No. 1996/1.
• Assessment of the potential for using FWDs for evaluation at network level
• Drafting common requirements for calibration of FWD measurements and devices
• Launch of a preparatory basis for possible European standardisation in the field of the use of Falling
Weight Deflectometers in pavement evaluation
The countries participating in COST Action 336 and the members of the Management Committee and
Task groups are listed in Annex B of this report. The management committee consisted of a wide variety
of members from in total 22 European countries. Due to their different background and disparity positions
in different types of organisation, this committee reflected all parties involved in manufacturing,
calibrating and using FWDs. Secondarily, this COST Action also resulted in a better co-operation amongst
the participants and the participating institutes and 22 participating countries, not only during this project,
but also for the future. All members of COST Action 336 are gratefully acknowledged for their work, their
time and their extraordinary devotion. Without their enthusiasm, this result and report would not have
been achieved.
Egbert Beuving
CROW, The Netherlands
Chairman of COST Action 336
April 2000
Second Edition
After the issue of the main results of COST Action 336 at the website of the COST Secretariat of the
European Union, a strong need was felt to publish paper versions of the final report. Unfortunately,
negotiations with the various parties involved did not lead to the desired result. In 2004, FEHRL offered
their assistance to have the findings of COST 336 and other unpublished COST Transport Actions
published in paper and digital format. Since quite some years passed since the issue of the first results, the
i
various parts of the final report needed some revision and editing prior to publishing. The major editing
consisted of converting the text into a similar lay-out for all chapters. Where necessary, chapters were
altered a little to make the information more up-to-date without departing too much from the original
contents of the report. Double appearances of similar text blocks were removed and errors were corrected
as much as possible.
The most extensive revision was applied to the calibration chapter, Annex F containing the calibration
protocols and Annex H with the description of the FWD Foundation Test. The revision of the calibration
was needed because the EC co-funded project SpecifiQ was on its way during the COST3 336 contract
period. One of the objectives of this project was to develop a set of FWD calibration procedures that could
be used as European Standard. Annex F consists of the calibration standards as developed by the SpecifiQ
consortium (draft 1) and corrected a year later (draft 2). These pre-draft standards will hopefully form the
new European Standards on FWD calibration. The text of the calibration chapter was amended accord-
ingly.
Since the finalisation of COST Action 336 quite some experience was gained with the FWD Foundation
test. New results and views were incorporated in Annex H.
ii
Abstract
Falling Weight Deflectometers (FWDs) are widely used throughout Europe for the evaluation of the
structural condition of road pavements. So far, the device is mostly used for assessment at project level.
However, the equipment has a potential for use at network level as well. COST Action 336 has carried out
inquiries on the current use of FWDs at these two levels, and based on this, produced guidelines for the
use of FWDs at these levels, with an emphasis on flexible pavements. To function properly, the measuring
instruments are in need of periodical calibration and verification, and the Action has therefore also coped
with the calibration procedures for FWDs.
The guidelines produced for the project level evaluation are based on an earlier FEHRL publication,
which during the Action is considerably expanded and updated. This is achieved between other means by
the use of a questionnaire on the 'Post-processing of FWD data' amongst the COST member countries. The
guidelines for the use of FWDs at project level covers requirements of measuring procedures, the
processing of measurements, and assessment of pavement condition and structural maintenance design.
Project level reporting is also addressed. A draft procedure is written for the determination of the stiffness
of the layers beneath the bound pavement layers from direct measurement on these layers by an FWD.
Even though the FWD is mainly a project level tool, it is widely used for network analysis, often with less
additional information. Information on the use at network level was achieved from a literature survey and
a workshop at LNEC, Lisbon in June 1997. These information together with the discussion of what is, and
what is the purpose of, network level analysis are the basis for the guidelines for the use of FWD at
network level.
The FWD calibration section of the report provides the findings of a state-of-the-art survey supplemented
by results of studies to fill the missing links between available and required FWD calibration procedures.
On the basis of these inventorial actions, a calibration approach with corresponding timetable was
developed. For reasons of user-friendliness, two levels of calibration were discerned, i.e. the easy-to-use
FWD user level and the more complex calibration station level. Detailed protocols are written for each of
the calibration and verification procedures.
The executed inquiries have shown the very wide range of practices in FWD measurement and processing
carried out in Europe. This has clearly emphasised the need for some harmonisation if the benefits of
European co-operation are to be fully realised.
iii
Kurzfassung
Fallgewichtsdeflektometer (Falling Weight Deflectometer, kurz FWD) sind innerhalb Europas zur
Bestimmung des strukturellen Verhaltens von Straßenkonstruktionen weit verbreitet. Das FWD wird
bislang hauptsächlich zur Bewertung von Straßenkonstruktionen auf Projektebene verwendet. Möglich ist
aber ebenso die Verwendung des Prüfgerätes auf Netzwerkebene. COST Aktion 336 hat Untersuchungen
zur Verwendung des FWDs auf den genannten zwei Ebenen durchgeführt und aufbauend auf der Analyse
der Umfrageergebnisse Anwendungsrichtlinien je Verwendungsebene aufgestellt. Bei den
Untersuchungen wurde die Betrachtung von flexiblen Straßenkonstruktionen in den Vordergrund gestellt.
Für eine einwandfreie Funktion des FWDs ist eine periodische Kalibrierung und Nachprüfung der
Meßgenauigkeit des Prüfgerätes erforderlich. Aus diesem Grunde wurden im Rahmen dieser COST
Aktion Kalibrierungsrichtlinien und Vorgehensweisen der Kalibrierung aufgestellt.
Die Richtlinien für den Einsatz des FWDs auf Projektebene basieren auf früheren FEHRL (Forum of
European National Highway Laboratories) Publikationen, die im Rahmen dieser Aktion auf den neusten
Stand gebracht und fortgeschrieben wurden. Dies wurde u.a. durch eine Umfrage mit dem Thema 'Analyse
von FWD-Meßdaten' unter den Mitgliedstaaten dieser Aktion erreicht. Die Richtlinie beschreibt
Anforderungen an die Messungen (Durchführung und Meßwertaufzeichnung), die Auswertung der
Meßdaten, die Bewertung des strukturellen Verhaltens der Straßenkonstruktion sowie die daraus
abzuleitenden Erhaltungsmaßnahmen. Die Vorgehensweise zur Bestimmung der Steifigkeiten (E-Moduli)
der Schichten unterhalb der gebundenen Schichten des Oberbaus, direkt aus den aufgenommenen
Meßdaten auf diesen Schichten, wird kurz beschrieben.
Obwohl das FWD in erster Linie ein Prüfgerät auf Projektebene darstellt, wird es immer häufiger für
Straßenzustandsanalysen auf Netzwerkebene eingesetzt. Informationen zum Einsatz des FWDs auf
Netzwerkebene wurden durch eine Literaturrecherche sowie durch einen Workshop zu diesem Thema im
Juni 1997, am LNEC in Lissabon, zusammengetragen. Diese Informationen, zusammen mit der
Diskussion der Ziele des Einsatzes von FWDs auf Netzwerkebene, bilden die Basis für die Richtlinien für
den Einsatz des FWDs auf Netzwerkebene.
Für die Entwicklung einer einheitlichen Kalibrierungsrichtlinie sind Informationen zum Stand von
Wissenschaft und Technik und neue Erkenntnisse zusammengestellt worden. Weiterhin wurden Studien
durchgeführt, um fehlenden Informationen zwischen zur Verfügung stehenden und geforderten
Kalibrierungsrichtlinien und –vorgehensweisen zu ergänzen. Auf der Basis der zusammengetragenen
Informationen wurde eine Richtlinie mit dazugehörender Zeittafel entwickelt. Hierfür wurden, unter
Berücksichtigung der Benutzerfreundlichkeit, zwei Ebenen von Anwendungsrichtlinien unterschieden,
eine vereinfachte Richtlinie für den FWD Benutzer sowie eine komplexere Anwendungsrichtlinie für
Kalibrierungsstationen. Detaillierte Protokolle für jede Anwendungsrichtlinie zur Kalibrierung und
Nachprüfung der Meßgenauigkeit von FWDs wurden erstellt.
Die im Rahmen dieser COST Aktion durchgeführten Nachforschungen und Untersuchungen zeigen einen
großen Spielraum für den Einsatz des FWDs und dessen Meßwertauswertung in Europa. Dies bestärkt die
Forderung nach einer harmonisierten Anwendungsrichtlinie für den Einsatz des FWDs zur
Vergleichbarkeit der Daten und damit zur Möglichkeit und Sicherstellung einer europäischen
Kooperation.
iv
Résumé
Les déflectomètres à masse tombante (FWD) sont largement utilisés à travers l'Europe pour l'évaluation de
l'état structurel des chaussées. Jusqu'à présent cet appareil était principalement utilisé pour l'auscultation
des chaussées au niveau étude (projet routier). Néanmoins, il pourrait tout aussi bien être utilisé à l'échelle
d'un réseau. Cette action COST 336 a réalisé des enquêtes sur l'utilisation actuelle des FWD à ces deux
niveaux (projet routier/gestion du réseau) et, à partir de là, a proposé des guides d'utilisation des FWD, en
insistant sur les chaussées souples. Pour fonctionner de manière correcte, les instruments de mesure
nécessitent d'être étalonnés et vérifiés périodiquement; l'Action s’est donc également intéressée aux
procédures d'étalonnage des FWD.
La guide qui concerne l'évaluation des chaussées au niveau du projet, a été rédigé à partir d'une publica-
tion antérieure du FEHRL. Au cours du déroulement de l'Action, cette publication a subi un développe-
ment considérable et a été révisée. Cela s'est fait, entre autres, par le biais d'un questionnaire sur 'le
traitement des données du FWD' adressé aux pays membres du COST. Le guide d'utilisation au niveau du
projet routier couvre les exigences relatives aux procédures de mesure, le traitement des mesures ainsi que
l’évaluation de l'état de la chaussée et des travaux d'entretien structurel. L'ensemble des besoins pour une
étude routière est ainsi couvert. Une ébauche de procédure est décrite pour déterminer la rigidité des
couches non liées directement à partir d'une mesure FWD sur ces couches.
Bien que le FWD soit un outil essentiellement adapté aux études routières, on l'utilise fréquemment pour
l'analyse d'un réseau (routier), souvent avec peu d'informations complémentaires. Des informations, sur la
gestion des réseaux routiers ont été recueillies dans une étude bibliographique ainsi qu'au cours de
rencontres de travail qui se sont déroulées au LNEC à Lisbonne en juin 1997. Ces informations, associées
à une discussion sur ce qu’est un réseau routier et sur le but de l’analyse d'un tel réseau, constituent
l'essentiel du guide d'utilisation du FWD.
La section du rapport sur l'étalonnage du FWD présente l'état de l'art en la matière, complété par les
résultats d'études visant à établir les relations manquantes entre les procédures d'étalonnage du FWD
existantes et requises. A partir de cet inventaire, une approche de l'étalonnage est développée, incluant des
tableaux d'enchaînement des différentes procédures. Par souci de convivialité, on distingue deux niveaux
d'étalonnages, à savoir, le niveau Utilisateur du FWD "utilisation simplifiée" et le niveau plus complexe
de station d'étalonnage. Des protocoles détaillés ont été rédigés pour chacune des procédures d'étalonnage
et de vérification.
Les enquêtes réalisées ont mis en évidence une pratique très diversifiée quant à la réalisation des mesures
et de leur traitement en Europe. Cela souligne bien la nécessité d'une certaine harmonisation pour
bénéficier pleinement de la coopération européenne.
v
Executive summary
Most the deliverables of COST Action 336 are a direct continuation of the different procedures and usage
investigated by a FEHRL FWD group from 1991 to 1996, with the aim of creating guidelines for making
measurements with FWDs. This COST Action 336 comprised originally three task groups, which have
contributed to fulfil different parts of the Memorandum of Understanding of the Action. In this executive
summary, the results of the task groups are summed up one by one.
The main use of FWDs has for many years been as one of the tools used to assess the need for mainte-
nance of a selected section of pavement. Thus, at the first level it is used to help to confirm or supervise
the need for maintenance suggested by network level surveys. At the next level, the FWD can be used to
diagnose any structural deficiencies in the pavement, and the probable causes of distress. It is also often
used to help design the appropriate strengthening treatments where needed.
FWDs are now widely used throughout Europe for the evaluation of the structural condition of road
pavements at project level. However, a questionnaire on the 'Post-processing of FWD data', answered by
31 respondents from 15 COST member countries and summed up in Annex C, has shown the very wide
range of practices carried out in these countries. For example, 22 different backcalculation programs for
calculating layer stiffness moduli were used by the respondents, and the definition of structural life varied
widely with criteria based on layer stiffness, roughness, rutting or cracking being used. This has clearly
emphasised the need for some harmonisation if the benefits of European co-operation are to be fully
realised.
Task Group 1 of COST 336 has therefore produced guidelines on the use of Falling Weight Deflectome-
ters for the evaluation of flexible pavements at project level. These are to be found in chapter 4 of this
report, and provide advice for this whole operation, covering:
• Measuring procedures, including layer thicknesses and temperature
• Pre-processing of measurements, including normalisation/standardisation of deflections and subdivision
into homogeneous sections
• Calculation of layer stiffness moduli, including pavement modelling, backcalculation programs and
normalisation of layer stiffness moduli
• Determination of stresses and strains
• Estimation of traffic
• Assessment of pavement condition, including direct stiffness interpretation and residual life estimation
• Structural maintenance design, including estimation of required overlay thickness
• Project reports
Task Group 1 has also written a draft procedure for the determination of the stiffness of the layers beneath
the bound pavement layers from direct measurement on these layers by an FWD. The results obtained
from these measurements can be used as an assessment of the bearing capacity of the road foundation.
This test is called the FWD Foundation Test and is to be found in Annex H.
vi
Task Group 1 has attempted to provide considered guidelines for the whole FWD measurement and
analysis procedure that will improve the reliability and consistency of the results. This should also provide
useful guidance for those countries less experienced in the use of such pavement evaluation techniques,
and help to enable effective exchange of FWD consultancy services within the Single Market. This in turn
should lead to more appropriate maintenance decisions and to the better use of maintenance funds for the
participating countries. In addition, the application of the guidelines should enable more meaningful
exchange of research results between countries. However, to fully realise these benefits, the work needs to
be further refined and developed into a form of Standard or Official Guide, possibly in the form of an EN
Standard.
The guidelines on the use of the Falling Weight Deflectometer for network level is one of the major
deliverables of Task Group 2 of COST 336. The guidelines are to be found in chapter 5, and describe how
and when FWDs should be used at network level, and can provide estimates for the timing of maintenance
and strengthening requirements to road agencies.
The information was achieved from a literature survey and a workshop held at LNEC, Lisbon in June
1997 and - last but not least - from the experience in the COST 336 member countries. The literature
survey and the workshop are reported separately (Annex E). From this literature survey and from the best
of our knowledge within COST 336, the workshop turned out to be not only a European, but also an
almost complete worldwide overview of where and how FWD and other deflection measuring systems are
used for network level analysis, mainly as part of pavement management systems (PMS's).
All European countries increasingly consider structural condition achieved from routine deflection
measurement as a parameter in their PMS systems. This is probably due to the fact that FWDs and other
bearing capacity devices are seen as a project level tool, and it seems difficult to focus the fairly detailed
project values to the much more aggregated network values. Anyhow, some systems presented at the
workshop and in the guidelines show the benefit by using one set of FWD measurements for both network
level and too, with additional information, for project level analysis of the weak sections.
The guidelines should be useful to the member countries to increase the benefit of their FWDs by using it
for both project and network analyses.
FWD calibration
For any structural (re)design procedure it is absolute necessary that the registered data are accurate and
unbiased. Inaccurate and biased data lead to incorrect conclusions about the structural condition of road
and airfield pavements. This in turn can imply that incorrect estimates are made of the remaining life.
Improper calibrated FWDs, errors in testing and data collection, inaccuracies in normalisation and
adjustment approaches for magnitude and duration of FWD loads, are all, among other parameters, factors
of influence in this process. A systematic error of 5% in deflection coincides with a similar error in critical
asphalt strain, but with an error of 25% in terms of pavement life.
Task Group 3 of COST 336 has therefore prepared a set of approaches and procedures for the calibration
and verification of individual components of the FWD but also for the FWD as a complete unit. The main
results are presented in chapter 6 of this report.
vii
The basis of the calibration procedures developed was laid in a state-of-the-art survey. Available
procedures and references were reviewed via a checklist on usefulness and completeness for the purpose
of a pan-European FWD calibration approach. The inventory resulted in a list of FWD calibration aspects
and procedures that had to be covered by the calibration guidelines. The inventory disclosed that some
calibration procedures were lacking in current practice. The verification of accuracy of a deflection sensor
reading at remote distance of the FWD load impact was the most missing feature.
The report presents the results of a COST financed study on the investigation of the feasibility of servo-
accelerometers for in-situ deflection sensor calibration. The findings of that study formed, together with
the state-of-the-art survey data and other research data, the basis of the new calibration procedures.
Two levels of calibration and verification of proper FWD operation are distinguished. The first level,
termed the FWD user level, contains all weekly, monthly and other periodic check-ups that can, may and
must be made by the FWD user. The procedures used at this level are generally easy-to-use, do not require
complicated auxiliary reference instrumentation and are not time-consuming. The second level, or the
FWD calibration station level, consists of the more complex calibration routines that require skilled
technicians and/or expensive instrumentation and equipment.
A calibration scheme and timetable was developed, basically to illustrate the interaction between
calibration actions, but also to provide an overview of all procedures, the level upon which they must be
used and the frequency of application per year.
In the calibration approach, distinction was made between reference calibration and relative calibration,
but also between calibration and verification. Some of the procedures developed address the issues of
repeatability and reproducibility. All these terms are explained in full detail in the report.
viii
Table of contents
Preface..................................................................................................................................................... i
Abstracts (English, Deutsch, Français)................................................................................................. iii
Executive summary............................................................................................................................... vi
1 Introduction.............................................................................................................................................1
2 Workplan.................................................................................................................................................9
3 FWD description and requirements..................................................................................................... 12
4 FWD project level guide...................................................................................................................... 20
5 Use of FWD at network level .............................................................................................................. 97
6 Calibration of Falling Weight Deflectometers .................................................................................. 116
7 Benefits to different users .................................................................................................................. 145
8 Conclusions and recommendations ................................................................................................... 148
Annex A Technical annex to the Memorandum of Understanding
Annex B Committee membership
Annex C Current practice of post-processing of FWD data in Europe - Results of a questionnaire
Annex D Current FWD analysis programs
Annex E Current practice of FWD use on network level
Annex F FWD calibration procedures
Annex G Owners of FWDs
Annex H FWD foundation test
Annex I Glossary of terms
Annex J Bibliography
More detailed information on the contents may be found at the beginning of the longer chapters 4-6.
ix
CHAPTER 1
INTRODUCTION
In a growing number of European countries the Falling Weight Deflectometer (FWD) is becoming the most
common device for the determination of the bearing capacity of different types of pavements e.g. roads,
airfields (runways, taxiways, aprons etc.) and industrial paved areas, both during construction (on the various
pavement layers) and after construction. Currently more than 300 devices are operational worldwide.
Because the pavement response and behaviour is dependent on load level and loading speed, various attempts
have been made to measure the bearing capacity, by means of deflection measurements, under a moving
truck. Because it was not possible to measure those deflections accurately under a high-speed truck, the
Benkelman beam was developed in the fifties. This system measures deflections under a rolling wheel of the
rear axle of a loaded truck at low speed. Since this technique had disadvantages as e.g. low speed and the
problem of finding the 'zero level', people started to search for another solution for measuring deflections
under a short loading pulse to simulate traffic travelling at high speed.
The idea of development of a Falling Weight Deflectometer started in France [1-1]. After a while they
decided to stop the development of the falling weight and France continued with the development of the
French moving Benkelman beam, the Lacroix deflectograph instead. Based on experiments from France, the
first attempts to produce an FWD were made in Denmark in 1964. Experiments of performing bearing
capacity measurements of pavements were carried out with a newly constructed apparatus at the National
Road Laboratory (now Danish Road Institute). Initial studies showed that the apparatus at the National Road
Laboratory was not as effective as the equipment developed in France as regards transfer of force. Several
modifications were made based on information obtained from France, and new equipment was constructed at
1
the Technical University of Denmark. With this new equipment, the goal was almost accomplished to
achieve a half sine wave as loading pulse.
Then tests were carried out and the apparatus verified against plate bearing tests. The first falling weight
apparatus was extremely difficult to use. Phønix A/S managed to construct a small series of falling weights in
a more handy construction. In 1975, S-shaped springs were replaced by three rubber buffers. The weight was
lifted hydraulically by means of a hand pump and mounted on a trailer. The deflection was measured
electronically.
In the United Kingdom, as in France, the development of the falling weight deflectometer was viewed
sceptically on the basis that the load duration in the deeper layers would be unnaturally short under a falling
weight loading. Furthermore, it was argued that the inertia in the road materials, which had been activated by
the falling weight, would also play a role. The objections raised by France and the United Kingdom resulted
in the fact that Denmark carried out comparative measurements between the falling weight and a tire moving
at high speed. The initial results were very positive. The deflections were similar for the two types of load
pulse generation.
Dynatest was formed in 1976 as a result of a co-operation between the Danish Road Institute and the
Technical University of Denmark. During that period, similar experiments were carried out in Sweden by
KUAB, who started routine operations in 1976.
In the Netherlands several companies started to build an FWD for there own use in the 80’s. These efforts
resulted in a wide variety of different brands of FWDs and also a variety of interpretation methods for the
measured deflections. CROW installed the study committee ‘Falling Weight Deflection Testing’ in 1988 to
solve this problem and speed up harmonisation of the FWD-equipment and the FWD-measurement method
including the interpretation method. This resulted in a Dutch guideline for FWD-measurements and for the
interpretation of FWD data in 1995. For the harmonisation of the equipment various FWD comparison tests
were held in the Netherlands, starting in September 1998. This resulted later into the Dutch guidelines for the
calibration of FWDs.
In 1987 Foundation Mechanics, Inc. started to produce JILS-FWDs in the USA. Foundation Mechanics, Inc.
is also the manufacturer of the Road Rater (since 1970). In Japan, Komatsu is building FWDs too.
Phønix /
Carl Bro
KUAB
Dynatest
Komatsu
JILS
2
Figure 1-3 Old manually operated FWD of TU-Delft (1978) with S-shaped springs
3
1.2 European co-operation
In the 80's FWDs were used more and more, in the beginning as a research tool and later for routine practice.
A lot of countries gained their own experience in this field in that period. In the beginning of the 90's there
was a need to exchange experience in a broader way.
In 1991 the Forum of European National Highway Research Laboratories (FEHRL) started with a FEHRL -
FWD Seminar in Delft. Two years later the second Seminar was held in France. This was the start of a
FEHRL - FWD activity group, which drafted FEHRL publication no 1: 'Harmonisation of FWD Measure-
ments and Data Processing for Flexible Road Pavement Evaluation at Project Level'. During the last Seminar
in Denmark, the final draft was discussed and accepted. This publication contains the findings and results of
the three FEHRL - FWD Seminars with contributions from 14 European countries. In this FEHRL - FWD
activity group, the idea of converting the research activities into a COST Action started. In 1995 FEHRL
officially initiated a COST Action on FWDs, and in 1996, the European Commission started the COST
Action 336 'Falling Weight Deflectometer' officially.
The COST framework was chosen as the most appropriate co-ordination funding mechanism in this area for
the following reasons:
• It is desirable that as many of the nominated COST countries benefit from the work as possible; not only
those within the EU at the present time
• In order to create an effective common code of good practice, it is desirable to have agreement between
the technical representatives of national governments
At network level, the FWD can be used to develop a bearing capacity data bank for budgeting, allocation of
budget per area, maintenance section selection and prioritisation (maintenance section order). The difference
between network level and project level is extensively described in chapter 5. Decisions on maintaining or
rehabilitating roads at network level are usually directed by pavement management systems (PMS).
PMS's at network level seek to determine the optimum maintenance strategies to be applied to network
clusters or links of road sections. Optimisation is a process that determines the best possible use of available
resources (maintenance strategies) taking into account budgetary and scheduling constraints. Maintenance
strategies are predetermined sets of maintenance standards associated with intervention thresholds that will
trigger the corresponding maintenance action. The criterion for optimisation is an expression called the
4
objective function that sets what should be maximised. Different approaches can be taken to define the
objective function.
One approach is 'economic' and seeks to minimise agency funding or total transportation costs (including
user costs). A PMS of this type uses simulations based on life cycle cost analysis to optimise maintenance
strategies. These PMS types usually take into account the user’s benefit goal, and will base the management
decisions on obtaining 'the most for the agency spending'. The aim is to treat and therefore remove as many
sections as possible from maintenance each year.
Another approach is technical and looks at maximising road performance or at meeting certain levels of
service standards. Remaining service life analysis is used in this case in the PMS. These PMS types will seek
to minimise the cost for the agency. The management decisions will be based on a 'worst first' strategy that
can be applied in this context. Its goal will be to treat and therefore remove as many as possible bad sections
each year, after which the medium condition sections can be treated.
To be complete, a broader optimisation could be achieved through multi-criteria analysis extending the
technical and economic considerations by taking into account factors such as environmental impact, overall
productivity of freight transport, safety and comfort etc.
All of these PMS's need performance prediction models in order to be run.
At the project level, measurements are performed on a road section with the aim of obtaining knowledge on
the behaviour of that section alone; these measurements may then be used to aid the design of strengthening
measures.
There are many possible uses for FWD measurements, but the main sections of this document deal only with
the use of FWD measurements for the evaluation of completed flexible and semi-rigid pavements at project
level (in chapter 4) and at network level (in chapter 5).
In this document the difference between flexible and semi-flexible (or semi-rigid) versus rigid pavement is
based on the measuring and data interpretation procedure. In the case of flexible or semi-flexible (or semi-
rigid) pavements, the pavements will be considered as homogeneous layers in the longitudinal direction of
the road and the selection of the test points can be done without restrictions. In the case of rigid pavements,
the slab dimensions and the location of the joints have to be taken into account when selecting the test points.
In the case of rigid pavements the load transfer at the joints should be measured as well.
The use of FWD measurements for rigid pavement evaluation at project level is not described by COST 336.
The evaluation of rigid pavements is more complex than with the flexible ones, mainly because of the finite
slab sizes, the joints between the slabs and the variable load transfer at the joints. The experience gathered
with the use of FWDs on rigid pavements is limited (in Europe). Recently (at the end of this COST Action)
some documents, that were able to serve as a basis for harmonisation of the use of FWD at rigid pavements,
became available. So the work in this field may be performed by another COST Action in the future.
During COST Action 336, a CROW-committee in the Netherlands was working parallel on an evaluation
method for rigid pavements at project level using an FWD. This resulted in CROW Publication 136 'Uniform
Evaluation Method for Jointed Cement Concrete Pavements' (in Dutch) [1-2]. This publication also contains
a computer program (in the English language) called 'UEC-Slab', that can be used for e.g. backcalculation of
FWD data and the calculation of stresses in jointed cement concrete pavements due to different loads,
residual pavement life, etc, etc. The decision to publish this software in the English language was partly due
5
to the efforts of COST Action 336. In this way, COST Action 336 also created extended knowledge in this
area.
Other documents that describe a/o. the evaluation of rigid pavements are the UK Highways Agency Design
Manual for Roads and Bridges [1-3] and the Spanish 'Norma de Ensayo del Centro de Estudios de Carreteras'
NLT-338/98 [1-4].
FWD measurements can be used for the relative comparison of the condition of pavement layers based on the
measured deflections only, and/or the shape of the deflection bowl, for:
• Subgrade
• Road-base and subgrade
• Entire pavement structure
Whilst these comparisons can be used for research purposes, for a road condition data bank or to make
subsections for other evaluation methods, it should be noted that, in road engineering, deflection measure-
ments are not a goal by itself. Stresses and strains in the pavement layers are the crucial elements in pavement
design. However, when only deflection levels are considered, it must be borne in mind that a comparison can
be misleading, particularly when different pavements with different layer thicknesses are considered, because
the deflection-strain/stress relations are different for different pavement structures. With a given deflection
bowl, different stress/strain levels may be obtained, if the pavement structure varies.
FWD measurements can also be used for the direct determination of the load-carrying capacity of:
• Asphalt pavements
• Cement concrete pavements.
This load-carrying capacity can be evaluated in different ways using nomographs or 'simple' computer
programs (based on e.g. the central deflection and/or the difference between certain deflections) enabling
classification of pavements according to their structural condition. In some countries, this classification is
used at network level; in other countries it is also used for overlay design, and sometimes for the acceptance
of a new construction.
The measured deflections obtained from the FWD can be used for the determination of the stiffness moduli
of the different pavement layers:
• Subgrade
• Road base and subgrade
• Asphalt layer, road base and subgrade
6
• Cement concrete layer, road base and subgrade.
The backcalculation used to determine stiffness moduli has to be carried out with considerable engineering
judgement. If the layers are too numerous or too thin, it is difficult if not impossible to backcalculate stiffness
moduli accurately. This is also the case when stiff layers form part of the pavement structure.
The backcalculated stiffness moduli derived from FWD measurements can be used:
• to assess the relative contribution of bound and unbound materials to the pavement strength
• to indicate any weak areas that need replacing or special consideration
• to identify the structural quality of a critical layer (or interface)
• to calculate stresses and strains in pavement layers due to the load imposed
• to calculate the estimated (total) pavement life, using the calculated stresses and strains in combination
with a fatigue curve or deformation criterion and the traffic history
• to determine the residual pavement life, using the calculated total pavement life and the predicted traffic
in the near future
• to calculate the overlay thickness if the residual pavement life is shorter than the required pavement
design life.
Although the use of a mechanistic model is common, strains in flexible pavements can also be calculated
without first determining the stiffness moduli. Strains can be determined using algorithms, provided that
there exists a relation between strains and deflections; for example between the difference of the central
deflection and the deflection at 600 mm offset and the horizontal strain at the bottom of the bound layers.
In the case of cement bound pavements, FWD measurements can also be used:
• to assess the severity of cracks in cement bound material (load-transfer)
• to examine the load transfer efficiency across joints in rigid cement concrete pavements
• to detect voids below slabs
• to examine the foundation support available beneath slabs
In the design of a pavement, design values of layer stiffness modulus are set for the different materials, such
as subgrade, sub-base, road base and asphalt layers. During the road construction, the degree of compaction
is verified and the (assumed) stiffnesses of the different pavement layers are only measured in some
countries. Until now, the static plate bearing test was mostly used for this purpose. This test method is time
consuming and therefore several consultants and road authorities started to use the FWD for measuring the
stiffness of the granular layers. Each consultancy and/or road authority had their own measuring procedure
and consequently various diameters of the loading plate were used. For that reason it was decided in COST
336 to develop a harmonised 'FWD Foundation Test'.
7
In Annex H to this report, a procedure is described for the determination of the stiffness of road base and
underlying substructure with the use of the Falling Weight Deflectometer. The objective of the procedure
described in this document is to gain experience and expertise with the test and analysis procedure. The
results obtained with this test serve as input for the assessment of the bearing capacity of the road foundation.
The stiffness of the road base and underlying substructure or the sub-base and underlying substructure is
computed from the test load and the measured peak value of deflection.
The test system consists of the towing vehicle with all electronic systems and data acquisition equipment, and
the test trailer. This trailer is in fact a regular FWD with minor modifications. This test can also be used on
(cement and/or bituminous) bound (road base) layers.
1.5 References
[1-1] Bretonnière, S., Étude d'un Déflectomètre à boulet. Bulletin de Liaison des Laboratoires Routiers,
No. 2, 1963.
[1-2] Uniformering Evaluatiemethodiek Cementbetonverhardingen (Uniformed Evaluation Method
Jointed Cement Concrete Pavements), CROW Publication 136, March 1999, CROW, Ede, The
Netherlands; ISBN 90 6628 300 9.
[1-3] UK Highways Agency Design Manual for Roads and Bridges; Volume 7: Pavement Design and
Maintenance; Section 3 Pavement Maintenance Assessment; Part 2 HD 29/94 Structural Assess-
ment Methods; Part 3 HD 30/97 structural Assessment Procedure; Published by Her Majesty's Sta-
tionery Office, London.
[1-4] Norma de Ensayo del Centro de Estudios de Carreteras NLT-338/98. Medida de Deflexiones con el
Deflectómetro de Impacto. Centro de Estudios y Experimentación de Obras Públicas (CEDEX).
Madrid 1998, Spain.
8
CHAPTER 2
WORKPLAN
To initiate COST Action 336, a Technical Sub-Committee was set up to define:
• Objectives and benefits
• Scientific programme
• Organisation and timetable
• Economic dimension
for the use of FWD in pavement evaluation.
These activities defined the content of the Technical Annex to the Memorandum of Understanding (MoU).
This MoU is shown in Annex A.
2.1 Objectives
The main objective of COST Action 336 'Falling Weight Deflectometer' is the development of a European
common code of good practice for use of Falling Weight Deflectometers in pavement evaluation.
According to the MoU (see Annex A) this goal has encompassed the following activities:
• Expand FEHRL harmonisation proposal to incorporate strengthening evaluation on the basis of FWD-
tests.
• Establish common requirements for calibration of measurements and machines.
• Describe the potential for use of FWDs in evaluation at network level.
• Establish a preparatory basis for possible European standardisation in the field.
The Action will also contribute to the achievement of a number of wider objectives. Examples are as follows:
• Lack of dependence on individual FWD types.
• A more harmonised market for organisations involved in bearing capacity testing by means of FWD.
• Extended knowledge of FWD testing on roads with flexible and rigid pavements.
The work of COST Action 336 is organised and supervised by the Management Committee. Four task
groups were set up and assigned responsibility for completing Tasks 1, 2, 3 and 4.
These 4 tasks are:
• Task 1: Post-processing of FWD data
• Task 2: Applicability of FWDs at network level
• Task 3: FWD calibration
• Task 4: Finalisation of deliverables and reporting
Management
Commitee
9
2.2 Tasks
The original tasks of the task groups are described in the MoU presented in Annex A. The tasks are
summarised below together with the additions and amendments made in later stages of the project.
This task started with a review of the FEHRL document, followed by providing a complementary description
of the post processing of FWD data at project level to the following items which had not been covered until
now:
a) Calculation of in-situ layer stiffness
b) Normalisation of data from different FWD types
c) Correction of layer stiffness to standard conditions
d) Calculation of critical stresses and strains
e) Estimation of residual structural lives and the required thickness of strengthening overlays.
Sub-task b) was previously part of the working package of Task Group 3, but was moved to Task 1 during
the COST Action.
This task extended the work on harmonisation of FWD measurement already carried out by the FEHRL
FWD Activity group for executing measurements at project level. It involved the description of how and
when FWDs can be used at network level, and provided estimates for the timing of maintenance and
strengthening requirements to road agencies.
This task focused on quality assurance of FWD measurements, from a calibration point of view, in two areas:
a) Calibration of the complete FWD measurement system
b) Description of an instrumented calibration station for calibration checks on real pavements.
The third original task of this Task Group, 'Normalisation of data from different FWD types', was performed
by Task Group 1, because this change fitted better to the activities of that group.
This task consisted of the concluding and production stage of the work programme. The work of Task
Groups 1, 2 and 3 were collated in the final report.
a) Guidelines for the post-processing of FWD data at project level
b) Proposal for applicable procedures for use of FWDs at network level
c) Calibration protocols for FWD machines
d) Requirements for FWD calibration stations
10
e) FWD Foundation Test
Item e) was added to the original plan, because of the need for an adequate description of the FWD
Foundation Test.
11
CHAPTER 3
FWD DESCRIPTION AND REQUIREMENTS
Besides good procedures for the interpretation of FWD data, it is of course extremely important to start
with good data, not only for the deflections but also for the load level, the pavement temperature and the
location of the measurements. This chapter starts with a general description of an FWD. After that,
requirements for an FWD and FWD related equipment are reported.
Falling Mass
Rubber Buffer
Load Cell Deflection Sensor
Deflection Bowl
(not to scale)
The design of an FWD may vary in detail depending on the manufacturer. Often, this implies that the
shape of the load pulses can vary from one machine type to another. The stiffness of an asphalt layer (and
sometimes other layers too) depends on the loading speed, and hence the response of a pavement depends
on the pulse shape of the applied load. If the load pulse is to simulate a moving vehicle it should have a
rise time, which is approximately equal to the rise time of a moving wheel load. Most FWDs have a load
rise time from start of pulse to peak of between 5 ms and 30 ms and have a load pulse base width in the
interval of 20 ms to 60 ms.
12
Figure 3-2 FWD loading plate
3.1.2 Deflections
FWD devices should have a sufficient number of deflection sensors to describe the shape of the deflection
bowl. There are two kinds of deflection sensors that are used in FWDs; geophones (see Figure 3-3) and
seismometers (see Figure 3-4). The type of sensor that is used depends on the make of FWD. For
pavements with thin asphalt layers, the deflection sensors have to be located closer to the load centre than
in the case of a thicker asphalt pavement.
13
Figure 3-4 Deflection sensors of the seismometer type
14
3.2.1 Load pulse specification
• The target peak load should be 50 kN ± 5 kN (for a completed asphalt road pavement). A target load,
different from 50 kN, may be chosen to match the standard wheel load used in the pavement design
method or for other reasons.
• The load rise time from the onset of the load pulse to the peak of the load pulse should be between 5 and
30 milliseconds.
• The load pulse base width should be between 20 ms - 60 ms (for the influence of the loading time see
Chapter 3.1.1).
• When deflection sensors with a maximum range of 2000 mm are used as deflection sensors, deflections
greater than that cannot normally be measured accurately. In that case the load magnitude has to be
lowered in steps of 10 kN to produce a measured (maximum) deflection of less than 2000 mm. The
applied load should, however, remain as close as possible to 50 kN and should kept constant during
testing as much as possible.
• For each project, (road section), one target load level has to be used.
• In special cases, e.g. an investigation on the non-linearity of the (unbound) granular layers, more than
one load level can be used on each test point. One of these load levels should be 50 kN. In these cases it
is also recommended to use the same load levels on each location (within a section). Some FWD devices
may have an automatic load targeting feature so that at every measuring location the load is adjusted in
such a way that the target load is reached.
• The load pulse on a completed asphalt pavement should be applied through a circular loading plate with
a diameter of 300 mm.
• A rubber pad with a thickness of at least 5 mm should be glued to the bottom of the loading plate. The
pad should be ribbed or have some other pattern allowing reshaping. Concerning the contact between the
loading plate and the asphalt surface, the following comment can be made: A uniform contact between
the loading plate and the pavement surface is needed to ensure a reasonably uniform contact pressure.
This is one of the reasons for the use of the rubber pad. Some FWDs have a 20 mm thick PVC plate
interfaced between the steel loading plate and the rubber pad to improve the load distribution. Some
FWDs have a segmented loading plate (two segments or four segments) with a pad up to 20 mm thick to
try to achieve uniform contact with the road surface when there is rutting (see Figure 3-6).
15
3.2.3 Deflection sensor specification
• The load pulse on a completed asphalt pavement should be applied through a circular loading plate with
a diameter of 300 mm.
• FWD devices should have at least six deflection sensors, which could be positioned at radial offsets from
0 mm to 2500 mm.
• The positions of the deflection sensors should be chosen from the following list:
0 - 200 - 300 - 450 - 600 - 900 - 1200 - 1500 - 1800 - 2100 - 2400 mm.
Deflection sensors must at least be mounted at the following offsets: 0 - 300 - 600 and 900 mm. The
location of other deflection sensors depends on the stiffness of the total pavement structure. The
stiffness of the subgrade has a major influence on the deflection bowl shape, and therefore there
should be at least two deflection sensors at such a distance from the load centre as to enable the stiff-
ness of the subgrade to be assessed. The ideal deflection sensor locations would be two deflection
sensors "in each equivalent `thickness' of the pavement layer". In chapter 5 a method is described to
calculate the equivalent thickness of each layer. In order to define the preferred deflection sensor
locations, the pavements are divided into three categories, depending on the reading of the deflection
sensor in the centre of the loading plate (d0). In cases where only six deflection sensors are available,
the recommended deflection sensor locations are listed in Table 3-1.
Table 3-1 Recommended deflection sensor offsets when six sensors in use
Sensor Recommended offset (mm) in case of FWD load 50 kN and expected centre deflection:
number < 500 µm 500 - 1000 µm > 1000 µm unknown
Sensor #1 0 0 0 0
Sensor #2 300 300 300 300
Sensor #3 600 600 600 600
Sensor #4 900 900 900 900
Sensor #5 1500 1500 1200 1500
Sensor #6 2100 1800 1500 1800
It should be noted that the use of more than six deflection sensors is recommended as there is less
need to change the position of the deflection sensors during the measurements, in that case. This leads
to fewer errors because every deflection sensor will keep its own fixed position. If only six deflection
sensors are used and measurements are performed on one day on different pavements with a wide
range of thicknesses, it may be necessary to change the positions of the deflection sensors during the
day to measure the most relevant parts of the deflection bowl shape.
• The load pulse on a completed asphalt pavement should be applied through a circular loading plate with
a diameter of 300 mm.
• Accuracy: systematic error within ± 2% of the reading ± 2 µm or less . The systematic (relative) error is
the difference between the mean measured value of a deflection and the absolute correct deflection value.
The random (absolute) error is the scatter around the mean value.
16
3.2.4 Load cell specification
• Accuracy: systematic error equal to or less than 0.5% of full scale range or 2% of the reading (whichever
is the greatest)
Some FWDs are unable to maintain the correct drop height. This means that successive peak loads may
vary due to the variations in drop height. The normalisation of the deflections for load obviates this effect.
Evidence of a satisfactory absolute calibration by the equipment manufacturer and consistency checks by
the operator should be provided as described in the following sub-chapters.
It is recommended that these FWD measurements should be carried out on a fixed section every month to
check the whole system when it is in use. Due to seasonal variations and temperature effects, the measured
deflections at any one location will not remain constant during a year. These quality control FWD
measurements cannot be used for calibration purposes; they are only intended to give an impression of the
FWD system as a whole by comparing the measured deflections with the previous history. Part 4 of
Annex F describes the exact procedure to be followed.
• Relative check-up:
- Prior to each test sequence, a check of the dynamic response of the machine as a whole should be
made by an automatic signal check on the background noise level of all deflection sensors.
- A deflection repeatability control check should be made at least each week of testing and should be
performed on a smooth, level, crack free asphalt pavement where a peak deflection, measured at the
centre of the loading plate, of the order of 250 µm to 600 µm can be generated. The standard
deviation of the normalised deflections of a series of drops, should, for each deflection sensor, be
equal to or less than 2 µm or 1.25% of the mean value of the reading + 1.5 µm (whichever is the
greater). (See Part 3 of Annex F for details).
- A load repeatability control check should be made at least each week of testing and should be
performed on a smooth, level, crack free asphalt pavement where a peak deflection of the deflection
sensor in the centre of the loading plate, of the order of 250 µm to 600 µm can be generated. The
standard deviation of the load should be less than 2% of the mean value of a series of drops.
- Relative deflection sensor calibration/comparison (stacking) should be made at least twice a year.
The deflection sensors are stacked vertically in the stand, one above another, so that all the
deflection sensors are subjected to the same pavement deflection. For this test the deflection level
should be between 250 µm and 600 µm. Difference between maximum and minimum of the
recorded normalised deflections should be within 4 µm. This test also provides a factor for fine-
adjusting the deflection sensors. Details are given in Part 5 of Annex F.
- The above mentioned relative control tests should also be carried out after any major service
involving replacement of essential components of the FWD.
• Reference control:
- The absolute calibration of the deflection sensors should be carried out once every two year.
- The absolute calibration of the load cell should be carried out once every two year.
17
- After any major service that could influence the calibration, the above mentioned reference control
should be carried out.
3.2.6 Reproducibility
As mentioned in sub-chapter 3.1, FWDs can vary in detail depending on the manufacturer. A consequence
of this is that the shape of the load pulse can vary depending on the manufacturer. Also different signal
control and processing techniques may be used in the deflection sensors and electronics of the various
models. This means that two different FWD brands, which produce the same peak load level, can give
different measured deflections on the same (asphalt) pavement at the same location. Ideally, different
brands of FWD should give the same deflections at the same load level. Part 10 of Annex F provides
guidance to match deflections generated by various types of FWDs as long as they all use the same class
of load pulse width. For larger differences in load pulse duration, no accurate conversion technique is
readily available.
The temperature of the bituminous layers of the pavement should be measured using a thermometer or
temperature probe with:
• Reading resolution: 0.5° or better
• Accuracy: ± 1.0°C in the range of -10°C to +50°C
For calibration of temperature probes, see Part 11 of Annex F. Calibration should be performed at least
once per year.
In order to locate the position of FWD test points, their distance from the beginning of a road section
should be measured. The requirements for distance measurement or oedometer are:
• Reading resolution: 1.0 m or less
• Accuracy: better than ± 0.5% of the measured distance
Comment: Calibration is performed with reference tape with an accuracy of 0.1%. The
comparison provides an adjustment factor, so the actual accuracy will be better than 0.5%.
To ensure that the locations of the test points can be found again after a certain period, it is recommended
to measure the distance from the beginning of the test section to certain fixed objects e.g. the centre of a
cross road, a viaduct or a bridge. For calibration of the oedometer see Part 12 of Annex F. This calibration
should be performed at least twice a year.
18
CHAPTER 4
FWD PROJECT LEVEL GUIDE
Preface
This chapter covers the use of a FWD at project level on flexible pavements from the planning of the
measurements through the methods of processing to the interpretation to finally the reporting of a project.
Figure 4-1 presents a flow chart with a simplified structure of this chapter. The flow chart also illustrates
the interaction between the different steps in the use of FWD in structural assessment.
Several COST countries have special procedures for testing and evaluation of FWD data. These
specialities do not form part of the generally accepted COST 336 procedure. The specialities are usually
indented but always marked with SÂ at the start and ÁS at the end of the text block.
19
Table of contents
20
4.6.5 Assessment of structural condition from stiffness moduli criteria .............................................. 68
4.6.6 Calculation of residual life from deflection criteria .................................................................... 69
4.6.7 Accounting for variances and uncertainties ................................................................................ 71
4.6.8 Comparison of calculations with observed pavement condition ................................................. 73
4.7 Structural maintenance design................................................................................................. 76
4.7.1 General ........................................................................................................................................ 76
4.7.2 Design criteria and additional design data................................................................................... 77
4.7.3 Determination of overlay thickness from strain or stress criteria................................................ 77
4.7.4 Calculation of required overlay thickness from deflection criteria ............................................. 79
4.7.5 Accounting for variances and uncertainties ................................................................................ 80
4.7.6 Use of the structural maintenance measure into practice ............................................................ 80
4.8 Project reports ........................................................................................................................... 81
4.8.1 Location and description of test site............................................................................................ 81
4.8.2 Description of test equipment and site test procedures ............................................................... 82
4.8.3 Methods and locations of measuring deflections and temperatures ............................................ 82
4.8.4 Choice of load level, time and date of testing ............................................................................. 82
4.8.5 Calibration status of equipment................................................................................................... 83
4.8.6 Methods of pre-processing data and delineation into homogeneous sections............................. 83
4.8.7 Calculation of moduli, strains and pavement lives...................................................................... 83
4.8.8 Other material tests...................................................................................................................... 83
4.8.9 Presentation of pre-processed deflections ................................................................................... 83
4.8.10 Presentation of construction and other information .................................................................... 87
4.8.11 Presentation of derived results..................................................................................................... 89
4.8.12 Comparison of available information .......................................................................................... 93
4.8.13 Design of maintenance treatment ................................................................................................ 94
4.8.14 References ................................................................................................................................... 94
4.8.15 Appendices .................................................................................................................................. 94
4.9 References .................................................................................................................................. 94
21
4.1 Measuring procedures
Before FWD testing can be performed on site, certain data have to be collected and operational choices
have to be made. First of all, it is important to determine the goal of FWD testing and what the collected
data will be used for. The following sub-chapters specify the data needed for preparing FWD testing. They
supply an instruction list and describe how temperature recording should be carried out. Criteria for
selection of lane for deflection measurements are listed, and procedures for assessing the minimum
number of test points per pavement section are described.
22
Plan FWD measurements (4.2)
23
4.1.2 Instruction list
The following list of instructions should be prepared for the FWD operator:
• Temperature limits for FWD testing (see 4.2.4.1)
• Choice of lane (see 4.2.5)
• Choice of survey alignment (see 4.2.6)
- In the nearside wheelpath
- Between the wheelpaths
- In the nearside wheelpath and between the wheelpaths
- Special
• Test point spacing (see 4.2.7)
• Load level (default: 50 kN) plus recommendations as to action when the target load level generates
deflections beyond the recording capacity of the deflection sensors (see 4.2.8)
• Deflection sensor offsets (see 3.2.3)
• Number of drops per test point (see 4.2.9)
• Special requirements (optional).
Where the results of the FWD testing are used for the assessment of the residual pavement life, additional
information will be needed (see 4.7.3)
The following sub-chapters contain practical information and tests that will be helpful to the FWD
operator, and therefore essential for satisfactory interpretation of the results.
Information about the existing pavement structure is important, not only in the backcalculation process,
but also in the preparation of FWD testing. The pavement section under investigation can consist of
different structures in the longitudinal direction and could have been widened in the past. In the case of
old(er) roads, the pavement structure details may be unknown or may not be well defined. Structural
details or construction details of newer pavements could be readily available from local authorities.
In general, information on the pavement structure and the pavement thickness can be obtained from
historical data, by coring, trial pits or Ground Penetrating Radar surveys (GPR) [4-2]. For the preparation
for FWD testing, GPR surveys can be performed to acquire a good impression of the layer thicknesses of
the pavement structure in the longitudinal direction and in the transverse direction when GPR surveys are
performed in two or more lanes. GPR has the advantage over coring that it can be performed at high speed
and the layer thicknesses can be estimated at short intervals e.g. 1 m spacing. It is recommended to take
some cores after a GPR survey. These cores will be used for the calibration of the GPR results and for the
identification of the different (asphalt) layers. After analysis of the GPR surveys it is easier to determine
where the FWD testing should be performed. After the FWD testing, the locations for coring can be
determined based on the FWD data or in combination with GPR-data when available. Accurate measure-
ments of layer thickness are essential for the analysis of FWD testing, particularly if stiffness moduli are
to be backcalculated. Sufficient cores should be taken to provide a reliable record of layer thickness data.
Besides the total thickness of the bituminous layers, it is also important to collect data of the types and
thicknesses of the individual asphalt layers at the interpretation stage of the data. Different construction
stages of the road in the past may be detected in this way. Only coring and historical data can provide the
latter information but normally GPR surveys are necessary to provide an indication of the longitudinal
variation in layer thickness. Figure 4-2 presents two types of GPR surveys. Figure 4-3 shows a raw result
of the survey.
24
Figure 4-2 Air-launched and ground-coupled Ground Penetrating Radar
25
4.1.3.2 Accuracy of layer thickness measurements
Accurate layer thicknesses are needed for the interpretation of the FWD data,. The following accuracy is
required:
• For asphalt (bituminous bound) layers the accuracy should be within 5% of the total thickness of these
layers
• For unbound road base and lower layers the accuracy should be within 10% of the total thickness of
these layer(s)
Since the stiffness moduli of asphalt layers are significantly affected by temperature, consequently the
measured deflections are affected too. This involves that temperature normalisation to reference
conditions is required when deflections are compared to each other. Ideally the full temperature gradient
of the asphalt layer is required but in practice it has been found that a more restricted range of measure-
ments are adequate.
The preferred temperature range for FWD deflection testing is between 0°C (freezing) and 30°C (or a
suitable alternative), where the temperature of the asphalt is measured at a depth of 40 mm or more. In
26
countries with a temperate marine climate the temperature limit of 30°C is mostly used. Countries in other
climatic zones may use the following guideline for the maximum temperature: '2/3 of the expected
softening point (Ring and Ball Temperature) temperature of the top asphalt layer expressed in °C'.
The roadbase, sub-base and subgrade should not be frozen, except when special measurements of bearing
capacity during frost and thaw-weakening periods are required.
It is important that the depths are measured and recorded together with temperature, times of measurement
and location. It can be useful to store information about the weather condition as: sunny, cloudy, rain. A
single in-depth temperature measurement is often found adequate when conducted at a depth between one
third of the total depth of the asphalt layers and mid-depth and with 120 mm as a maximum. If the
thickness of the layers are not known, then a depth between 40 mm and 100 mm can be chosen on the
basis of the expected structure thickness for the type of road under test. In case of layer thicknesses of less
than 50 mm, one mid-depth measurement is sufficient. Where asphalt thickness is too thin to drill a hole,
then the surface temperature alone might be adequate. Multiple in-depth temperature measurements can
provide more reliable information. For example: when the asphalt layer thickness ranges between 50 mm
and 100 mm, two holes can be used
• First hole at 25 mm below pavement surface
• Second hole at 25 mm from the bottom
27
When the asphalt layer thickness is greater than 100 mm, temperatures may be recorded at three levels in
the asphalt layer (similar to SHRP method [4-1]):
• First hole at 25 mm below pavement surface
• Second hole at mid-depth
• Third hole at 25 mm from the bottom
Estimated pavement temperatures for each FWD test point can be interpolated from the available test data.
• For two-way single carriageway roads, deflection testing can be made in one lane and therefore in one
direction, or in both lanes and directions using 'staggered' locations. This provides efficient coverage
when the traffic loading is similar in each lane. Figure 4-5 shows an example of 'staggered' locations but
they may not be suitable on the grounds of safety.
The measurements may be performed between the wheelpaths, in the nearside or verge side wheel path or
in both wheelpaths.
The disadvantage of measuring between the wheelpaths only is that these measurements take little or no
account of the degree to which trafficking has affected the pavement properties.
28
• No edge effect adjustments are needed for strain calculations if the nearside wheelpath is (very) close to
the pavement edge (see 4.7.4). When the deflections are measured at such a test point, the deflection
values will be higher than those captured between the wheelpaths, simply because of the influence of the
nearby pavement edge. If a multi-layer computer program is used to backcalculate layer stiffnesses, the
edge effect will not be considered by the program and the calculated stiffnesses will be effective
stiffnesses and not actual stiffnesses. If these effective stiffnesses are then used by the same multi-layer
computer program to calculate strains, then it is likely that in the modelling this effect is corrected and
the 'actual' strains are calculated.
• Measurement are conducted on the loaded area of the pavement
On occasions, testing has to be performed in the nearside wheelpath of the lane, between the wheelpaths,
and in the offside wheelpath. Such additional test, if performed on the same structure, can provide a
valuable indication of the condition of the largely untrafficked area of the pavement that has been
subjected to the same environmental conditions as the trafficked wheelpath. If only a rough indication of
the bearing capacity of the untrafficked area is required, testing in the wheelpaths and at every fifth test
point between the wheelpaths remains a possibility.
In order to assess the variation of the measured deflection bowls in a given road section, the following
specification should be complied with:
• Normally at least twelve test points per (sub)section are required. A (sub)section is a part of the road that
has the same uniform pavement structure and also a uniform bearing capacity in the longitudinal
direction. Normally one consistent rehabilitation design will be specified for the whole of a (sub)section.
When testing is performed at project level, sufficient measurements need to be performed to enable a
road (length) to be delineated into homogeneous subsections where necessary. Sufficient test points per
subsection are needed to provide a statistical basis for decision making.
• For project level investigations, the longitudinal spacing of the test points should not be more than
100 m.
• If the variation along the road is substantial, test point spacing may be shortened in order to discriminate
between uniform subsections.
29
4.1.8 Choice of load level
The normal target peak load should be 50kN (see 3.2.1). However, when the deflections measured with
the outer deflection sensors are low, e.g. less than 20 µm, the significance of any inaccuracy of the
measured deflections will increase. There are three solutions to overcome this problem:
• Increase the load level in steps of 10 kN to increase the outer deflection to a level above 20 µm
• In the case of stress dependent pavement materials, a target load equal to the standard pavement design
wheel load may be preferred. In this case, the outer deflection sensor can be placed closer to the load
centre (the two outer deflection sensors still have to measure the subgrade stiffness).
• Accept that the far field deflections are smaller than 20 µm, and therefore less accurate, knowing that the
subgrade is very stiff anyway.
For each project or pavement road section only one target load level should be used; all sub-sections
within that section will therefore be subjected to the same target load level. In special cases, e.g. an
investigation on the non-linearity of the (unbound) granular layers, more load levels can be used on each
test point. In these cases it is also recommended to use the same load levels on each site within a section.
During testing, several aspects have to be taken into account to obtain accurate data. It is also important to
store all the relevant data to be able to determine (e.g. during coring and the visual inspection of the
pavement and the interpretation of the data) the precise position of the FWD during the measurements in a
later stage.
• There should be no standing water (surface texture fully filled by water) on the road surface under the
FWD loading plate
• Care should be taken to ensure that the whole FWD loading plate is in contact with the road surface.
(This means that a segmented plate may be needed for measurements in the (deformed) wheelpath. When
a non-segmented loading plate is used, then the presence of significant rutting should be recorded,
particularly if it adversely affects the contact between the loading plate and the pavement surface.
• At least three loading cycles, excluding a small drop for seating the loading plate, should be made at each
test point and checks for consistency should be made prior to analysis
SÂ It is possible to use a corrective approach to the strain to take into account the errors due to
contact loading with the edge of the loading plate when measuring in ruts with a non-segmented base
plate (see 3.2.2) ÁS
SÂ The target number of drops can be made dependent on the stability of the measured deflection. In
some countries the repetition of drops is stopped when differences in deflections of successive drops
for the central deflection sensor are less than 5% or less than 2 µm. In other countries, at least three
loadings are applied. An average deflection bowl (of at least three drops) is computed. This approach
will minimise the effect of random deflection measurement errors. Significant differences in
deflections measured on successive drops may occur on some pavements, and they may increase over
ten or more drops. This suggests that material condition may change during FWD testing. ÁS
30
- Drop number
- Local time (in hours and minutes)
- Surface temperature, if measured
- Peak values of load and deflections
- Preferably the whole time history of load and deflections should be stored for one of the test points
of each road section
• The following data should be stored for the complete file:
- Deflection sensor offsets
- Loading plate diameter
- Deflection sensor numbers and calibration factors
- Test program (e.g. which drops are stored)
- Carriageway name and location (e.g. A-4 South)
- Name of operator (at least in some file)
- Day of testing
- State of filtering/smoothing option and cut off frequency
For each deflection bowl used in further analysis, the asphalt temperature is calculated from the measured
in-depth and/or surface temperatures. This temperature is then used for asphalt layer modulus temperature
adjustment, as described in 4.4.7. When deflection bowl parameters or strains are calculated from the
measured deflection bowl, the measured deflection bowl is first normalised to the reference temperature.
For in-depth measurements the asphalt temperature is specified as the arithmetic mean of the available
measurement data. However, when only surface temperature measurements have been performed, the
BELLS3 equation [4-3] should be used together with information on daily air temperature to provide an
estimate of asphalt temperature. The exception to this is when this method of measurement was used
because the layers were too thin to permit a measurement hole to be drilled.
31
Unrealistic values, or obvious errors must have been removed (e.g. measurements on a bridge).
Computer programs [4-1] for the validation of data exist, and can be used for this purpose.
The deflections are normalised to the defined target load by linear extrapolation or interpolation. This
means that the measured deflections are multiplied by the ratio target load over measured load. In
general, the target load is 50 kN. The contact pressure equivalent of the 50 kN target load on a
300 mm diameter plate is 707 kPa. E.g., if the deflections of a specific drop are due to a 730 kPa
loading pressure, then the measured deflections are multiplied by 707/730 = 0.97 to provide
normalised deflections.
SÂ The EC co-funded consortium SpecifiQ developed various approaches that account for load
pulse shape [4-4]. A simple conversion formula was developed on the basis of regression analysis to
convert deflections measured with a KUAB device at 60 ms load pulse duration, to deflections at
25 ms load pulse duration. As the deflection sensors of the KUAB device are usually set at 0. 200,
300, 450, 600, 900 and 1200 mm offset, only the deflections at 0, 300, 600, 900 and 1200 mm offset
were used. The basic structure of the conversion equation is:
32
Table 4-1 Regression coefficients for normalisation of deflections from 60 ms to 25 ms load
pulse duration
Offset i Regression coefficients
(mm) ai bi ci di ei fi r2
0 20.544 1.080 -0.631 0.998
300 27.268 1.153 -0.675 0.992
600 22.021 -0.390 1.205 -0.223 0.989
900 2.605 0.640 0.983
1200 4.697 0.644 0.985
1500 4.468 0.627 0.987
1800 6.794 -0.350 1.133 -1.323 1.161 0.995
Table 4-2 gives an example of this conversion. It must be emphasised that other coefficients will apply
for other load pulse durations.
ÁS
The following equations are shown as an example. They are based on simulated data of Dutch asphalt
mix characteristics [4-5]. The magnitude of temperature normalisation mainly depends on the
thickness of the asphalt layer, the temperature and the distance to the FWD load centre.
Deflections at testing are divided by the temperature normalisation factor to predict deflections and
deflection differences at reference temperature.
⎛ a ⎞ ⎛ a ⎞
TNF = 1 + ⎜⎜ a 1 + 2 ⎟⎟ * (TA − 20 ) + ⎜⎜ a 3 + 4 ⎟⎟ * (TA − 20 )2
⎝ h1 ⎠ ⎝ h1 ⎠ (4-2)
The coefficients a1 through a4 are tabulated for the centre deflection and several deflection differences
indices. The reference temperature is +20°C.
33
Table 4-3 Coefficients of temperature normalisation of FWD deflections
Variable a1 (°C-1) a2 (mm/ °C) a3 (0.001 °C-1) a4 (mm/ °C)
d0 0.01661 -0.67095 0.28612 -0.01408
d0 - d225 0.05955 -2.73223 1.48011 -0.08171
d0 - d300 0.05398 -2.61130 1.28439 -0.07493
d0 - d450 0.04720 -2.39175 1.05022 -0.06371
d0 - d600 0.04190 -2.15168 0.87228 -0.05301
It has to be stressed that the material characteristics used for developing these equations are typical to
those used in the Netherlands. Various other normalisation formulas exist, and those proven or
reasoned to be the most applicable to local conditions (structures, materials, FWD equipment) should
be used.
34
4.2.3 Delineation into homogeneous subsections
A variety of eye ball methods and statistical techniques can be used for the delineation process.
Within a given road section, the measured deflections on one part of the section are often
significantly higher or lower than those measured on another part. In this case, it is desirable to
divide the main section into subsections, each with a significantly different load bearing capacity.
A homogeneous subsection is a part of the road in which the measured deflection bowls (or any
other variable) have approximately the same magnitude and where it is not possible to subdivide
the road into subsections with significantly different behaviour.
When continuous information on layer thicknesses is available, this information can be used for
the delineation of the project into homogeneous subsections. This data can be obtained using e.g.
the ground penetrating radar (GPR). These procedures are described in sub-chapter 4.2.3.
Figure 4-6 shows sample deflection data from a road section of 3 km in length. The FWD
measurements were conducted with 50 m test point spacing in the nearside wheelpath. The deflections
were normalised to the target load level of 50 kN. The centre deflection is plotted in Figure 4-6. From
the figure, a distinctive change in the average level of deflection can be seen at Station 1900, and a
slight change at test point 1400. The road section could therefore be divided in three subsections:
Stations 0 - 1400, 1400 - 1900, and 1900 - 3100.
35
500
400
Centre Deflection, µm
300
200
100
0
0 500 1000 1500 2000 2500 3000 3500
Stationing, m
Figure 4-6 Sample deflection data from a road section
S1 = x 1 − x m
S 2 = x 2 − x m + S1
(4-3)
Si = x i − x m + Si −1
The cumulative sum approach can ideally be used to determine to what extent a certain part of a road
section is different from the mean deflection of the whole section. Changes of slope of the line connecting
all cumulative sum values point at subsection bounds.
Figure 4-7 presents the plot of cumulative sums based on the data displayed in Figure 4-6. A section
change is clearly visible at test point 1900. Some change of slope could be identified at test point 1400.
The changes in slope of cumulative sum between stations 0 - 1400 would lead to very short subsections
with too few measuring points per section. Also, the significance of the change in deflection within the
short sections is questionable. Thus, in this example, the road section is delineated into three distinct
subsections based on centre deflection: test points 0 - 1400, 1400 - 1900, and 1900 - 3100.
36
0
-100
Homogeneous
Cumulative Sum, µm
-200 section 2
-300
-400
-500
-800
0 500 1000 1500 2000 2500 3000 3500
Stationing, m
Figure 4-7 Illustration of the Cumulative Sum Method for determining homogeneous
subsections of a road section based on FWD deflection data from Figure 4.6.
Centre deflection
Layer thicknesses
Subgrade
Traffic volumes
Final division
(4 subsections)
Figure 4-8 Illustration of delineation of a road section into homogeneous subsections based on
FWD deflection data, structural information and traffic.
When statistical analysis is carried out at a later stage, e.g. for the calculation of a mean deflection bowl, a
subsection needs to contain a minimum number of test points. A standard deviation calculated from only
three deflections/test points is not meaningful. It is recommended that each subsection include at least
twelve points if statistical analyses have to be performed. If a subsection consists of only one or two test
points, it should be considered in need of localised rehabilitation or maintenance actions.
37
4.2.3.5 Testing statistical significance of subdivision
After the delineation into homogeneous subsections, it is advisable to check if there is a statistically
significant difference between these subsections. If not, they can be considered as one. A Student t-test
can be performed to check whether the means of variables of consecutive subsections are statistically
different from each other. If this is not the case, the subsections should probably be merged into one. A
less accurate alternative for determining the level of homogeneity is the use of the coefficient of variation
(CV). This parameter is defined as the ratio of the standard deviation over the mean value per section. The
following list shows typical classes of CV.
The CV is a measure of the consistency of the measurements spatial variability within the individual
sections/subsections. Although the CV may indicate that a section or a subsection is not very homogene-
ous, it gives no indication of the possibility of subdividing it. CVs greater then 30% usually indicate a
highly skewed distribution, produced, for example, by a number of relatively 'stiff' test points within a
weaker subsection.
4.2.4.3 Selection of (nearest match to) representative and/or safe deflection profile
The information needed for calculating the stiffnesses of the different pavement layers depends on the way
in which uncertainty or 'risk' is dealt with. For a given subsection layer stiffnesses can be calculated for:
• Each test point.
• The actually measured deflection bowl (test point) which is the nearest match to the mean deflection
bowl.
• A mean deflection bowl could be derived from calculating the mean deflections at each sensor location.
However, it is strongly recommended that an actual measured deflection bowl is used, rather than an
artificial deflection bowl. An artificial deflection bowl may have characteristics that do not exist in
practice and this could result in an unrealistic combination of layer stiffnesses for that subsection.
• The actually measured deflection bowl, which is the nearest match to a 'statistical' bowl. For example, an
85th percentile bowl could be estimated as the bowl of which every point (d0 , d200, d300, d 450, etc.) is
calculated as the sum of the mean value and standard deviation of that subsection.
38
For the determination of a 'statistical' deflection bowl, the mean deflections plus b times the standard
deviation per sensor-distance are calculated. The factor b depends on the chosen confidence level and the
chosen failure criteria for the design method. In this case an actually measured (nearest match) deflection
bowl has to be chosen and not the artificial deflection bowl. If the subsections have a high CV a
logarithmic normalisation of the deflections may be a better method of determining the 'statistical' bowl,
as it will reduce the effect of skewness in the distribution.
Large diameter cores are needed for the investigation of the composition of the asphalt mix to avoid
shifting of the aggregate gradation to the finer end. However, many test set-ups for determination of
material properties can only handle cores of smaller diameter (Ø 100 mm).
39
theoretically predicted deflection bowl is equivalent to the measured one. Given an assumed pavement
structure, a theoretical deflection bowl can be predicted using assumed (seed) moduli values. The layer
moduli values can then be adjusted and the process repeated until the predicted deflection bowl matches
the bowl measured by the FWD. Some methods use a database of deflection bowls from which layer
moduli are calculated using regression equations. A wide range of programs is available to carry out this
analysis. This sub-chapter mainly addresses the straightforward linear elastic approach to backcalculation.
E0 = ⎢
( )
⎡ 2 * 1 − ν 2 ⋅ σ0 ⋅ a ⎤
⎥
⎢⎣ dr ⎥⎦ (4-4)
E 0( r ) = ⎢
( )
⎡ 1 − ν 2 ⋅ σ0 ⋅ a 2 ⎤
⎥
r ⋅ dr (4-5)
⎣⎢ ⎦⎥
The equivalent depth of layer #n, (equivalent distance from the surface to the top of the layer) or
equivalent thickness of the layers above layer #n, can be calculated from:
⎡ E1 + ) ⋅ 3 E2 + )... + En −1 ⎤
h e,n = f i ⋅ ⎢(...((h1 ⋅ 3 h2 h3 h n -1) ⋅ 3 ⎥
⎣ E2 E3 En ⎦ (4-6)
This calculation is based on a modification of the original theory of Odemark. For values of r which are
greater than the equivalent thickness of the total pavement, the surface modulus will approximate the
subgrade modulus.
40
Figure 4-9 shows two examples of a 'surface modulus' plot. The graph on the left-hand side shows an
increasing surface modulus with decreasing equivalent depth. This means that the stiffness modulus of the
lower layers is less than that of the upper layers. The stiffness of the subgrade will be around 100 MPa.
The graph on the right-hand side exhibits a pavement which has a 'soft' interlayer between the upper layers
and the subgrade. The stiffness of the subgrade is about 150 MPa in this example and the stiffness of the
'soft' interlayer will be substantially lower than 75 MPa. The stiffness of 75 MPa would be approximately
the equivalent stiffness of the aggregate system of soft layer and underlying layers including subgrade.
Another approach is to use artificial neural networks to calculate pavement layer moduli. The artificial
neural network can be 'trained' to map deflection basins onto their corresponding pavement layer stiffness
moduli using a database of FWD results. An advantage of this method is the speed at which calculations
are carried out once the neural network has been trained. Such systems can be trained for a variety of
pavement systems. However, the use of artificial neural networks is invalid if the modulus of any one of
the pavement layers falls outside the training database [4-12].
41
4.3.1.4 Automatic deflection bowl matching
This method involves forward calculation using an iterative approach. In this system, theoretical
deflections are calculated for a set of layer moduli which may or may not be user defined (seed moduli).
These layer moduli are then adjusted in an iterative manner until the error between the measured and
calculated deflections is sufficiently small. The user can generally set the acceptable difference. A
maximum number of iterations can as a rule also be set to account for situations where meaningful
solutions are not possible. To minimise unnecessary iterations the process can be stopped when the
difference between layer moduli on successive iterations is below a pre-set value or percentage change.
For estimation of layer stiffness moduli from the deflection bowl it is nearly always necessary to model
the pavement allowing predictions of deflections generated when a particular load is applied. This
modelling can be very simple, as for the surface modulus approach (see 4.4.1), or somewhat more
complex. The most common approach is to consider the pavement as comprising a number of layers of
finite vertical thickness but of infinite horizontal extent. Less common are finite element models where the
pavement is broken down into individual three dimensional elements to each of which properties such as
dimensions, stiffnesses, Poisson’s ratios etc. are assigned. The modelling behaviour can be further
complicated if dynamic loading is introduced rather than the more normal and simpler static loading.
Other complicating factors are the presence of non-perfect adhesion between layers, the asymmetrical
response of the pavement to loading etc. In most cases the choice of pavement model can have a
significant effect on the final result. Therefore, for the purposes of project level evaluation, recommenda-
tions are provided mainly for backcalculation procedures using simple elastic models under static loading.
In these models, each of the main pavement layers are considered as a horizontal layer of finite vertical
thickness and infinite horizontal extent apart from the semi-infinite half-space representing the lowest
layer of the pavement.
The upper layer of flexible pavements will usually be a bituminous bound layer. Bituminous bound
materials are visco-elastic and so stiffness is a function of a number of factors including loading time and
test temperature. The stiffness of a bituminous material can be measured in the laboratory by a variety of
methods. The test temperature should be quoted when reporting such stiffness measurements. Stiffness-
temperature relationships are discussed later in sub-chapter 4.4.7. One method is to carry out indirect
tensile tests on cores cut from the road surface. Great care must be taken when comparing these results
42
with those from FWD deflections due to the different testing conditions involved. Typical values for the
elastic moduli of bituminous materials under FWD loading conditions are given in Table 4-6.
Table 4-6 Typical values of stiffness modulus and Poisson's ratio of bituminous materials
Material Temperature Stiffness range Poisson's
(°C) (MPa) ratio
Hot rolled asphalt wearing course 15 4,500 - 7,500 0.40
Dense bitumen macadam base course 15 4,500 - 7,500 0.40
Hot rolled asphalt roadbase 15 8,000 - 10,000 0.40
Dense bitumen macadam roadbase 15 7,000 - 10,000 0.40
Dense bitumen macadam 50pen roadbase 15 10,000 - 13,000 0.40
Heavy duty macadam roadbase 15 11,000 - 15,000 0.40
Broken asphalt cement 20 1,000 - 6,000 0.25
Dense rolled asphalt 20 1,000 - 6,000 0.35
Crushed stone rolled asphalt 20 1,000 - 12,000 0.35
Typical values for the stiffness moduli of cement bound materials are given in Table 4-7. Pavements
which have a cement bound base overlaid with bituminous bound material can cause particular problems
in backcalculation. The cement bound layer will generally have a higher stiffness modulus than the
bituminous bound layer (provided it is intact). Therefore the bituminous bound layer will have a smaller
influence on the pavement deflections than in the case where there is no cement bound material present. In
some cases the stiffness of the bituminous bound layer will need to be fixed at a realistic value and only
the stiffness of the cement bound layer will be calculated. It is quite common for such cement bound
layers to contain cracks. Such discontinuities will cause the pavement to behave in an inhomogeneous
manner violating the modelling assumptions of most modelling programs. In these cases the results
obtained should be treated with caution.
Granular mixes are sometimes stress-dependent and therefore the measured stiffness is dependent on the
applied stress, usually increasing with increasing stress level. The temperature of granular materials does
not generally affect stiffness except in the case of freezing temperatures in combination with the presence
of moisture. However, because bituminous bound material is temperature dependent, the stress imparted
into the granular layer will vary and hence this layer could be considered to be temperature dependent.
Moisture content usually has a large affect on the measured stiffness of these materials. Some granular
materials such as limestone and concrete granulate can undergo cementing actions, which also have a
large effect on material stiffness. Typical values for the stiffness moduli of granular materials under FWD
loading conditions are given in Table 4-8.
43
Table 4-7 Typical values of stiffness modulus and Poisson's ratio of cement bound materials
Material Stiffness range Poisson's
(MPa) ratio
Concrete 30,000 - 70,000 0.20
Cement bound material intact 10,000 - 30,000 0.20
Cement bound material (primary cracking) 5,000 - 15,000 0.20
Cement bound material (primary and secondary cracking 500 - 5,000 0.30
Aerated concrete 500 - 3,000 0.15
Fly ash-cement stabilisation 1,500 - 6,000 0.25
Sand-cement mortar 2,500 - 7,000 0.25
Lean concrete 10,000 - 40,000 0.20
Crushed stone (strongly cemented, uncracked) 7,000 - 30,000 0.15
Crushed stone (strongly cemented, cracked) 500 - 3,000 0.35
Good quality gravel (weakly cemented, uncracked) 3,000 - 9,000 0.15
Poor quality gravel (weakly cemented, uncracked) 2,000 - 6,000 0.15
Good quality gravel (weakly cemented, cracked) 160 - 2,000 0.30
Poor quality gravel (weakly cemented, cracked) 90 - 1,200 0.35
Table 4-8 Typical values of stiffness modulus and Poisson's ratio of granular materials under
particular stress conditions
Material Stiffness range Poisson's
(MPa) ratio
Granular base (no cementing action) 200 - 500 0.30
Granular base (cementing action) 300 - 2,000 0.30
Granular sub-base 50 - 200 0.30
Rockfill 100 - 400 0.30
Sand 50 - 300 0.35
Lava 50 - 300
Mine spoil 50 - 300
Masonry granulate 50 - 250
Sand and gravel 70 - 300
Aerated slags 100 - 1,500
Stone slag 100 - 400
Mixed granulate 100 - 800
Blast furnace slag (before binding) 200 - 800
Concrete granulate 300 - 1,500
Crushed stone 150 - 350 0.30
Wet-mix roadbase 200 - 500 0.45
Capping layer 50 - 100 0.45
Gravel base (moisture content dependent) 100 - 375 0.40
Granular base 50 - 300 0.35
Granular sub-base (moisture content dependent) 50 - 300 0.40
Many subgrade soils are also stress dependent. As in the case of granular materials the measured stiffness
will be greatly influenced by the moisture content present at the time of test. Some work has been carried
44
out to relate soil stiffness to other well-known parameters such as CBR [4-1]. Typical values for the
stiffness moduli of subgrade and other materials under FWD loading conditions are given in Table 4-9.
Table 4-9 Typical values of stiffness modulus and Poisson's ratio of subgrade and other
materials
Material Stiffness range Poisson's
(MPa) ratio
Clay subgrade 30 - 150 0.40
Clay 20 - 80 0.40
Peat 10 - 40 0.45
Clay (very soft) 2 - 15 0.50
Clay (soft) 5 - 25 0.45
Clay (medium) 15 - 50 0.35
Clay (hard) 50 - 100 0.35
Sandy clay 25 - 250 0.35
Silty sand 7 - 20 0.35
Sand (loose) 10 - 25 0.35
Sand (dense) 50 - 80 0.35
Sand and gravel (loose) 50 - 140 0.35
Sand and gravel (dense) 100 - 200 0.35
Silt 2 - 20 0.30
Concrete blocks; pavers 500 - 1,000 0.30
Hard polystyrene foam 5 - 10 0.10
For all types of pavement modelling, each layer requires the assignment of an appropriate value of
Poisson's ratio. For some pavement modelling programmes, such as those based on the method of
equivalent thickness, this value has to be the same for all layers. For most models however, different
values can be assigned to each layer. Typical values for use with the above range of materials are given in
Tables 4-6 to 4-9.
Different programs can handle various numbers of layers usually up to four or five. Most programs tend to
work best however when the number of layers is restricted to three as shown in Figure 4-10. Therefore the
modelling of pavements will often require that layers of similar stiffness behaviour are grouped together in
order to reduce the overall number of layers. Some programs recommend that modular ratios be set in the
case of more than three layers. This method can be used in cases where there are two distinct granular
layers with different stiffness values. Generally, it is recommended that the model should contain only one
stiff layer (bituminous bound) and that moduli decrease significantly with depth (an Ei/Ei+1 ratio of greater
than two is sometimes recommended).
45
FWD Load
Layer 2 (Granular) h2
Layer 3 (Subgrade )
Layer 3 (Subgrade)
The existence of bedrock at shallow depth or similarly stiff material close to the pavement surface (5 m -
6 m approximately) will have a large influence on the calculated layer moduli. Some programs attempt to
take this into account when calculating layer moduli. The estimated depth to a stiff layer can be calculated
from the shape of the deflection bowl or is user defined. Once the depth to the stiff layer has been
calculated the pavement can be modelled by assigning an extremely high stiffness to the lowest semi-
infinite layer of the model. Some programs even allow the introduction of a rigid bottom layer.
Many pavement subgrades are non-uniform often being layered or at least increasing in stiffness with
depth due to the lower stresses encountered at depth. Some measure of this effect can sometimes be
successfully modelled by incorporating an imaginary stiff layer at depth.
46
4.3.4 Output parameters
or
2
1 ⎡ δ c,i − δ m,i ⎤
∑ ⎥ ∗ 100 0 0 ≤ Accuracy Limit (0 0 )
n
⎢ (4-8)
n i =1
⎢⎣ δ m,i ⎥⎦
The degree of accuracy is usually expressed as a percentage of the measured deflections but sometimes it
is expressed in absolute measurement units. A percentage accuracy of four or less is considered sufficient
in many cases. Preferably a cut-off accuracy of two percent or less should be aimed for. A percentage
accuracy of four may not be achievable in some cases, particularly where it is difficult to adequately
model the construction layers.
As part of the TG1 work program, the contents of a questionnaire were agreed at the final COST336
meeting in October 1999. At that time a list of 17 possible authors was distributed and this was adopted as
the target. Table 4-10 lists the authors and the publicly available computer programs for FWD analysis
available at that time. During November and December 1999 these authors were contacted. Fourteen
replied with answers to the questions (total of 17 programs). The results have been compiled and are
presented in Annex C. The results are summarised below using the headings under which the questions
were asked.
47
Table 4-10 List of publicly available computer programs for FWD analysis (status 1999)
Nr. Program name Contact name Company Email
KOAC•NPC
1 UEC-SLAB Marc Stet [email protected]
(Netherlands)
New University of
João Roche de
2 BAP (LEAD) Lisbon [email protected]
Almeida
(Portugal)
Washington State
3 EVERCALC 5.0 Linda Pierce DOT [email protected]
(USA)
Rijkswaterstaat
Arthur van
4 CARE DWW [email protected]
Dommelen
(Netherlands)
University of Zilina
5 CANUV Jozef Kamacka [email protected]
(Slovakia)
Bohuslav CVUT Prague
6 BACKLAY [email protected]
Novotny (Czech Republic)
LAW PCS, Reno
7 BOUSDEF Haiping Zhou NV [email protected]
(USA)
Cornell University
8 MODCOMP 4 Lynne H. Irwin [email protected]
(USA)
University of
UMPED
9 Waheed Uddin Mississippi [email protected]
PEDD
(USA)
Michigan State
MICHBACK Ronald
10 University [email protected]
MFPDS Harichandran
(USA)
Abatech Computer
11 DAPS Geoff Rowe Services [email protected]
(UK)
Australian Road
12 EFROMD2 Binh Vuong Research Board [email protected]
(AUS)
Dynatest
13 ELMOD Frank Holt [email protected]
(USA)
Texas Transporta-
14 MODULUS 5.1 Tom Scullion tion Institute [email protected]
(USA)
15 SISMOD Fuming Wang [email protected]
Waterways
16 WESDEF Don Alexander Experiment Station [email protected]
(USA)
17 CLEVERCALC Carl Lenngren [email protected]
48
4.3.5.1 Program Source
Of the fourteen authors there were seven from the USA, four from Europe, one with offices in both
Europe and the USA, one from China and one from Australia. Six of the authors could provide some form
support by fax, phone or e-mail.
The costs of the various programs varies from free (web download), nominal charges, to € 2,500. Training
is available from nine at the authors and the type of training varies. When asked if the program was
suitable for production (large-scale back calculation) or research (user control of many input parameters)
eleven authors answer both and three said it is more suited to research
To try and ensure some sort of consistency between analyses it is recommend that all analyses are carried
out using a simple straight forward multi-layer linear elastic backcalculation program using a minimal
number of layers. Other more complex analyses can be carried out but should always be supported by
justification for such complexity together with a comparison with analysis by the 'simple' approach.
49
4.3.7 Normalisation of layer stiffness for pavement temperatures
The recommended methods for measuring pavement temperatures are described in Section 4.2.4. The
stiffnesses of the bituminous bound layers depend on both the temperature of the bituminous materials at
the time of testing and the loading time. The loading time will be nearly constant for a given FWD device.
However, in order to compare deflections/layer moduli they should be normalised to a reference
temperature. This will usually be the design temperature for the country or region. A survey of design
reference temperatures was carried out as part of COST Action 333 [4-11]. The reference temperature
values reported ranged from 15°C to 25°C with the most common temperature being 20°C.
The normalisation procedure should be carried out on the measured deflections when they are to be used
to infer strain directly. However if the stiffness moduli of the various layers are derived from the measured
deflections then the bituminous bound layer stiffness can normalised to standard conditions. There are a
number of normalisation methods available. Figure 4-11 shows some of the temperature-stiffness
relationships currently used in Europe. The horizontal axis gives the pavement temperature and the
vertical axis provides the adjustment factor by which the estimated stiffness obtained at the given
temperature should be divided to provide the estimated stiffness at a reference temperature of 20°C.
Different relationships apply to different materials and often these also vary with the age of the material.
Clearly the further the test temperature is from the standard temperature the higher the uncertainty in the
normalised stiffnesses. Therefore the aim should always be to test as close as possible to the reference
temperature conditions.
2.5
ELMOD 4.0 [4.4.9]
UK [4.4.10]
Portugal [4.4.11]
Dutch Mix (S78, 20 Hz) [4.4.4]
2
1.5
Et/ E20
0.5
0
0 10 20 30 40
Temperature (C)
The loading time for the FWD is not necessarily the same as for design conditions. In some countries the
backcalculated stiffness modulus is therefore not only normalised for temperature but also for loading
frequency.
50
SÂ In the Dutch method, frequency and temperature normalisation are conducted simultaneously
according to the following procedure. For this purpose an appropriate asphalt concrete stiffness graph is
chosen. However, such a graph is valid for only one sinusoidal load frequency or load pulse duration. This
is solved by converting the actual temperatures into equivalent temperatures by means of a temperature-
frequency equivalency law.
The temperature normalisation principle is demonstrated in Figure 4-12. First, the temperature at the FWD
measurement is normalised for FWD loading frequency to stiffness graph frequency using the frequency -
temperature equivalency law. The expected stiffness value is read from the graph at this normalised FWD
measurement temperature. Confrontation with the backcalculated stiffness gives a stiffness shift. This shift
gives a good indication of the validity of the backcalculated stiffness.
Then the step to design temperature is made. This design temperature is again normalised from traffic
speed frequency to stiffness graph frequency using the frequency - temperature equivalency law. At this
normalised design temperature, the design stiffness is read from the graph, taking into account the
stiffness shift determined in the backcalculation process.
1 000
measurement design
temperature measurement design temperature
corrected for temperature temperature corrected for
FWD frequency traffic speed
temperature (°C)
100
0 5 10 15 20 25 30 35
1 1 1 ⎛ f sg ⎞
= - log⎜⎜ ⎟⎟
Tcorr + 273 Tactual + 273 11242 ⎝ f actual ⎠ (4-9)
An equation for converting design speed into an equivalent frequency is presented below [4-16]
51
where fdes = Design loading frequency (Hz)
hac = Asphalt layer thickness (m)
vdes = Commercial traffic design speed (km/h) ÁS
Great care must always be used when modelling pavement structures. All estimates of stiffness values are
based on the input parameters entered such as layer thickness and type. Therefore, errors in the accuracy
of input information will lead to errors in the output data.
Many regional roads, particularly in rural areas, are granular constructions with thin bituminous bound
seal coats (surface dressing). These roads can also be assessed using FWD. The pavement can be modelled
either as a two or three-layer structure. When the pavement contains a thin bituminous bound seal coat
(10 mm - 30 mm) this layer should be given a fixed stiffness value of say 3,000 MPa. This effectively
means that the backcalculation process focuses only on the granular materials and the subgrade.
Some pavement conditions can also be difficult to model effectively, particularly where they contain
discontinuities. One example of this is a pavement with lower cement bound layers, which have been
overlaid with thin bituminous layers. Very often these type of materials crack at irregular intervals due to
variations in material properties. Modelling of these type of pavements is therefore difficult due to the
inhomogeneity of the pavement structure.
For the determination of stresses and strains at the critical positions in the pavement the same pavement
model (e.g. linear elastic) has to be used as was used for backcalculating the stiffness moduli. The
calculated strain values are used as input to fatigue curves and combined with traffic information (sub-
chapter 4.6) to estimate pavement life, as described in sub-chapter 4.7.4. If the layer moduli cannot be
solved, the critical strains can be determined directly from the measured deflection bowl, based on neural
network models, regression equations, etc. Before calculating the strains, the bituminous layer moduli (or
deflections if no backcalculation is done) must be normalised to standard conditions. The normalisation
procedures are described in sub-chapter 4.3 for deflections and in sub-chapter 4.4 for layer moduli.
The fatigue life of pavement is related to the stresses and strains induced in the structure by the moving
wheel loads (see Figure 4-13). On pavements with 'thin' bituminous layers, the horizontal strain at the
bottom of the bound layers is considered critical. This strain can be calculated using backcalculated layer
moduli and thicknesses from FWD measurements. Usually, the maximum of the longitudinal or the
transversal strain due to a standard axle load is used. Sometimes, the longitudinal is used, as it easier to
measure in the field.
52
Figure 4-13 Critical stresses and strains in pavement structures
The vertical strain at the top of the subgrade and unbound base and sub-base layers, may need to be
considered. Repeated loading causes permanent deformations in unbound materials and subgrade. The
elastic strains at the top of the subgrade and unbound layers can be calculated using layer moduli and
thicknesses from FWD. There are equations that correlate these strains to the permanent (or irrecoverable)
strains in these layers.
There are also other failure mechanisms that occur in pavement structures that are not so easy to interpret
or predict from FWD measurements.
• On 'thick' bituminous layers, 'top-down' failure mechanism may become critical. This means that stresses
strains at the top of the bound layers (pavement surface) cause cracking that initiates at the top of the
layer, but modelling of these conditions are difficult.
• Environmental factors, most importantly climate, cause distress in the pavement. At low temperatures,
cracking occurs due to excess tensile stress in the bituminous layer, frost heaving of subgrades and/or
structures cause surface cracking, etc.
It is important, that whatever the type of computational core used, it is consistent throughout the whole
process. For example, if the layer moduli are calculated with a computer code using a linear elastic multi-
layer program, the critical strains should be calculated using a linear elastic multi-layer program, too!
4.4.3 Determination of critical stresses and strains from the backcalculated layer moduli
The backcalculated layer moduli and layer thicknesses used in backcalculation are used as input to a
forward calculation program. Usually, a linear-elastic multi-layer program is used. Before calculating the
strains, the bituminous layer moduli must be normalised to reference conditions (see 4.4). If the layer
moduli are calculated for a representative deflection bowl per section, then the strains and residual
pavement life will be calculated for one point per section, too. If the layer moduli are calculated for each
53
test point, then the strains can be calculated either for each test point or for mean (or other representative)
values of layer moduli and thicknesses. If strains and residual lives are calculated for each test point, these
and/or the variation of these can be used as one variable for delineation of the project into homogeneous
subsections (see 4.3.3).
For each calculation point, the following input parameters are needed:
• Layer thicknesses
• Thickness of the upper part of the subgrade (depth to a rigid layer or bedrock)
• Layer stiffness moduli
• Subgrade modulus
• Poisson's ratio of layers and subgrade
• Standard load (magnitude and contact area)
The output from linear elastic calculation will be (reference to further use in sub-chapter 4.7.4.1):
• Maximum of longitudinal and transverse strain at the bottom of bituminous bound layer
• Horizontal strain at the bottom of cement bound base
• Vertical strain at the top of the subgrade
• Vertical strain at the top of the unbound layers
The calculated strain values are used as input to fatigue curves and combined with traffic information (see
4.6) to determine pavement life, as described in sub-chapter 4.7.4.
4.4.4 Determination of critical stresses and strains directly from the measured deflection bowl
In some instances it occurs, that no reasonable set of layer moduli can be determined for a specific
deflection bowl. This may e.g. be due to lack of accurate information on the layer thicknesses variation in
the layer thicknesses. Also, alternating 'stiff' and 'soft' layers in the structure may cause errors in the
backcalculation process. In these cases, the critical stresses can be determined directly from the measured
deflection bowl. This offers a quick way of obtaining critical strain values. Methods to determine the
critical strains directly from the measured deflections can be based on:
• Neural network models
• Regression equations
• or other methods combining the measured deflection bowl with the values of the critical parameters
Usually such models are based on simulated deflection data. The simulated data is produced e.g. with
linear elastic multi-layer programs.
Similar equations were developed elsewhere. Before using such models, one should check the types of
structures to be analysed against the types of structures used for developing these models. The expected
range of layer moduli and thicknesses of the structures to be analysed should fall within the range of layer
thicknesses and moduli used for developing these equations. If the structures in the analysis are remarka-
bly different from the structures used for developing the equations, the equations should not be used, but
54
more appropriate equations are developed instead. Before calculating strains, deflections should be
normalised to reference conditions (see 4.3).
4.5 Traffic
4.5.1 General
The calculation of residual life and overlay thickness of pavements is based on traffic data because the
traffic loading is the major influence on the structural damage to pavement structures. Therefore almost all
methods use an estimation of the future traffic loading and, furthermore, most methods estimate the traffic
loading that has already been sustained by the pavement to calculate the degree of damage that is already
present.
The calculation of traffic load usually follows the same method as is used in the pavement design methods
of the various countries. The individual wheel loads, which vary largely in load values, axle and wheel
configuration, tyre pressures, suspension types, etc., are converted to the equivalent number of standard
axle loads (ESAL). A standard axle, P0 converts every axle load, P to an equivalent number of load
applications. It is very common to use power laws for the conversion of different axle loads to standard
axle loads. Sometimes the load is also corrected for non-standard axle and wheel configuration, dynamic
effects etc. The equivalency signifies a certain value of total fatigue damage that is inflicted by a defined
axle load, ESAL to the pavement.
At present there is no commonly accepted standard single-axle load (SAL) in Europe. In the AASHO-
Road test [4-19] a single twin wheeled axle with a load of 18,000 lbs. (about 82 kN) was used as a
standard. European countries use SALs from 80 kN to 130 kN. See Figure 4-14, adapted from COST 333
[4-18].
ESAL (kN)
0 20 40 60 80 100 120 140
AT
BE
HR
DK
FI
FR
DE
GR
HU
IS
IE
NL
NO
PT
RO
SI
ES
CH
SE
GB
55
Traffic load can be determined in different ways, for example by Weigh In Motion (WIM) data or traffic
counts. In the most favourable case, the equivalent traffic load can be determined from WIM data that
generate the axle load spectrum and sometimes also some other data such as axle configurations. In many
cases results will only be available from traffic counts, which can give the daily number of vehicles, of
commercial vehicles, or of commercial vehicles per vehicle class, depending on the type of counting. In
this case mean conversion factors will have to be applied.
The calculated results for cumulative standard axle loads are usually corrected for effects of lateral wander
of traffic (which is favourable for pavement performance), distribution of the traffic over several lanes and
effects of slope.
In the best case WIM data are available, preferably in conjunction with the determination of the axle and
tyre configuration. A WIM measurement will give the axle load spectrum over a measurement period
(usually some weeks or months). Each load from this spectrum is converted into an equivalent number of
standard axle loads. Usually power laws are used for this. For instance, the most widely used 4th power
law of the AASHO-Road test [4-19] implies that when an axle load is x times the standard axle load, and
if all other variables (wheel configuration, contact area etc.) are the same, the damage inflicted to the
pavement is equal to the damage by x4 standard axle loads. The exponent is often empirically derived but
can also be derived from the performance law of the material, which often also has the form of a power
law (see sub-chapter 4.7.4.1). This implies that the exponent in principle depends on the deterioration
mode which is considered ( e.g. fatigue cracking, rutting, etc) and can vary widely between 4 or less and,
in the worst case, 18 for fatigue deterioration for a cement bound base. In practice these different powers
are often neglected and the fourth power law is used without consideration for the most appropriate value.
In order to take into account the variations from the standard load, e.g. axle and wheel configuration,
dynamic behaviour etc., the axle load values can be normalised by correcting factors before converting
them to equivalent standard axle loads (ESAL). The calculation of traffic load from WIM data will
therefore usually imply the following formula:
NT n
⎛ P '⎞ D yr
⎝
∑
CumESAL = ⎜⎜ i ⎟⎟ ⋅
P0 ⎠ TWIM
⋅ G ⋅ L ⋅ f lat.wand ⋅ f slope ⋅ f lanes
(4-12)
i =1
where CumESAL= Cumulative equivalent number of standard axle loads over design period (ESAL)
P'i = Axle load i, in some cases corrected for non-standard configuration
P0 = Standard axle load
n = Power of power law used for conversion
NT = Total number of axle loads over the WIM measurement period
Dyr = Number of days in year1
TWIM = WIM measurement period (days)
G = Growth factor for commercial vehicles over design period L
L = Design period (years)
flat.wand = adjustment factor for lateral wander or lane width effect (see Section 4.6.4.1)
fslope = adjustment factor for slope (see Section 4.6.4.1)
flanes = adjustment factor for number of lanes (see Section 4.6.4.1)
56
1
Normally 365, but in some cases less, for example where commercial vehicles are banned over the
weekend. In this case 52 x 5 = 260 should be used.
Many variations of equation (4-12) are available. For instance, in some countries the lateral wander is not
included as a single factor but analysed for the actual pavement. Furthermore, in some countries, the effect
of wide base single tyres is taken into account by the use of another factor which depends on the
proportion of wide base tyres in the spectrum.
4.5.3 Determination of traffic load from traffic count data using conversion factors
Often there is no WIM information available and simpler information has to be used. In many cases the
annual average daily commercial traffic, AADTCV , is known from traffic counts. Commercial traffic is
defined in most countries (Austria, Belgium, Croatia, Finland, France, Germany, Portugal, Sweden,
Switzerland) as the number of vehicles with a gross weight of 35 kN or more [4-11]. AADTCV can be used
in combination with a commercial vehicle or truck conversion factor CCV, giving the average equivalent
number of standard axle loads per truck. Some counting systems supply the daily number of trucks per
truck class. This information can be used with a separate conversion factor per truck class j. In this case
traffic load can be calculated according to:
SÂ For example, in Table 4-11 the approximate national average truck conversion factors for seven
categories are given for the UK [4-1]. A rigid truck is a non-articulated commercial vehicle.
In many cases just the annual average daily commercial traffic AADTcv is known (without distinction of
truck classes) or estimated by multiplication of the annual average daily traffic AADT with an estimated
truck percentage. In this case only one single truck conversion factor Ccv will be used.
In some countries no explicit truck conversion factor exists but an average number of axles per truck and
an average axle conversion factor (the average equivalent number of standard axle loads per truck axle are
known. However, the product of both can again be considered as a truck conversion factor. The average
number of axles per commercial vehicle depends on the function of the road. For example on minor roads,
three axles per truck would be appropriate whereas on a major road a value of 3.5 would be applicable.
57
The axle conversion factor depends on the proportion of heavy trucks within the commercial traffic flow
and varies between 0.15 and 0.45.
4.5.4 Adjustments
In the equation for the cumulative equivalent number of standard axles, (see equations 4-12 and 4-13)
adjustment factors can be taken into account dealing with the effect of the number of lanes, lateral wander
or lane width effect and slope or longitudinal gradient. Furthermore non-standard load configurations can
be normalised before transferring them to equivalent standard axle loads.
SÂ For example in Germany the value for this factor is said to vary from 1.0 where the traffic has
been counted for the main lane only to 0.8 where the traffic has been counted cumulatively in three or
more lanes[4-20]. ÁS
The factor for lateral wander or lane width effect, flat.wand., describes the influence of lateral wander on the
damaging effect of traffic. This effect depends on lane width, because the narrower a lane is the more
concentrated the traffic is and the greater the damaging effect will be. There exists a Lateral Displacement
Measurement (LDM) system for estimation of tyre position and width [4-21].
SÂ In Germany this factor is set to 1.0 where the lane width is 3.75m or more. At the other extreme,
a value of 2.0 is used where the lane width is only 2.5 m [4-20]. Other countries may use other values.
ÁS
The third factor for slope effects, fslope , describes the effect of slope or longitudinal gradient. The steeper a
lane is, the more shear forces are applied to the pavement structure.
SÂ In Germany this factor varies from 1.00, for slopes less than 2%, to 1.45 where the slope is
greater than 10% [4-20]. ÁS
The axle load Pi (see Eq. 4-12) can be normalised with adjustment factors for one or more of these effects.
Factors due to axle configuration, wheel type, or suspension type can presumably be used as constants
within acceptable limits of accuracy. Table 4-12 provides OECD-values [4-24] for the factors for vehicle
properties.
58
Table 4-12 Adjustment factors for vehicle construction
Configuration Type Adjustment factor
Axle configuration Single axles 1.00
Tandem axles 0.60
Triple axles 0.45
Tyre type Dual tyre 1.00
Wide base single tyre 1.20
Normal single tyre 1.30
Suspension type Traditional 1.00
Improved 0.95
Factor for tyre properties such as inflation pressure [4-23], dynamic effects [4-24], loading area and
distribution of contact pressure are hardly ever used in practice.
4.6.1 General
There are several ways to assess the structural pavement condition. A frequently used method of
assessment of residual life consists of estimation of the traffic loads that the pavement was able to sustain
initially, and to subtract the traffic loads that have already been carried by the pavement. This assessment
is usually based on calculation of stresses and strains in critical positions in the pavement structure. These
stresses and strains are then entered in performance laws such as fatigue or permanent deformation laws
for the various layers of the pavement. This way of determination of the residual pavement life is in fact
the reverse of pavement design, and the methods and performance laws used in residual life calculations,
although different from country to country, often are similar to those used in the new design.
This sub-chapter was structured around this more or less standard procedure. However, there are
alternative methods. Some methods estimate the residual life directly from the stress or strain levels in the
present condition, without looking back how much traffic load was carried. Other methods compare the
present stiffness moduli with values which are supposed to be representative for the initial condition and
estimate from this the extent of fatigue damage and the residual life. Others again only look at deflection
levels and derive the residual life from these data. These alternative methods are mentioned where
appropriate but often in less detail.
Not only the calculation methods show many differences but also the intervention levels that are observed
in these calculations and the way they deal with the often substantial dispersion and uncertainties
associated with pavement analysis. Two remarks however, are relevant for all methods of condition
assessment:
• Before starting any analysis, it is necessary to collect sufficient, often very site-specific data. These data
often have great influence on the design criteria which are verified, on the decisions on material
properties and other assumptions and adjustments.
• Any method of residual life calculation will have its limitations and the results cannot automatically be
accepted. It is of utmost importance to compare these results with other indicators of the structural
condition such as visual inspection results and drilling core data and to verify if these different indicators
of the residual pavement life support each other.
59
4.6.2 Design criteria
The design criteria used in pavement design, evaluation and redesign methods are normally expressed as
requirements to the maximum damage based on one or more deterioration modes. This maximum damage
is usually not the damage at which the pavement fails but the intervention level, at which maintenance is
necessary to enable further use of the pavement.
Fatigue cracking due to critical strains or stresses in the asphalt concrete road base is the most widely used
design and redesign criterion. The asphalt layers of a pavement structure are subject to fatigue damage
under the influence of traffic. Due to this fatigue damage, the bearing capacity of the pavement structure
declines and longitudinal cracks occur at the bottom of the asphalt layer. If fatigue continues, this cracking
can propagate upwards. Over the course of time parallel cracks can occur on account of which crazing will
start principally in combination with narrow transverse cracks.
However, other criteria can be verified, depending on the pavement structure. Fatigue or instantaneous
cracking e.g. due to tensile strains or stresses can be relevant for cement bound road bases, while
permanent deformation resulting in longitudinal and/or transversal unevenness due to compressive strains
can be relevant for granular layers, particularly for pavements with relatively thin bound layers.
The decision which design criteria will be taken into account is again of influence on the additional design
data that must be collected (see 4.7.3).
It should be noted that, especially on heavily designed pavements, observed surface initiated cracking may
be difficult to explain from structural point of view. In many cases this surface distress is the dominant
deterioration mode for these pavements. Other surface defects such as ravelling do not seem to be related
to the structural condition. Therefore, these deterioration modes do not lend themselves for analysis on the
basis of FWD data.
Intervention levels should be specified for the damage addressed by these design criteria. For instance for
cracking of the asphalt concrete road base, intervention is usually desired at a moment well before serious
cracks develop, because structural repairs performed in a late stage will be very costly and give rise to
serious traffic hindrance. Some methods implicitly include margins for this in the calculations, while some
others attempt to account for the intervention in a more explicit manner.
Some extra safety margin is often incorporated in the comparison of the design criteria with the desired
intervention levels, as a number of principal uncertainties will be latent in the assessment of design
parameters. These uncertainties implicate that residual life predictions can deviate from reality. To avoid
overestimation of the residual life and therefore exceeding the intervention level, the residual life is often
calculated for a limited underpass probability.
Design data are necessary for analysis of the pavement condition. Some of these design data are assessed
in previous parts of this chapter:
• Structural data such as layer thickness and stiffnesses, which follow from the backcalculation process
(see 4.4)
• Design speeds and temperatures for purposes of normalisation of backcalculated stiffness moduli (see
4.4)
• Design load system, i.e. the load value and tyre - pavement contact geometry for which the stresses and
strains are calculated (see 4.5)
• Traffic design data (see 4.6)
60
However, a number of additional design data are necessary. Site specific data can influence certain
calculation assumptions, can necessitate adjustments to be made to the calculations or can provide extra
information when comparing the calculation results with the observed condition (see 4.7.8). Examples of
site specific data are:
• Function and use of the road
• Age of the structure
• Carriageway and lane geometry
• Pavement edge and verge conditions
• Drainage and moisture conditions
• Additional construction data
Lateral wander reduces the damage inflicted by the traffic to the pavement. This positive effect depends
not only on the transverse distribution of the traffic, but also on the thickness of the asphalt layer and the
modular ratio of the asphalt layer and the substructure. A relatively stiff layer will distribute the load
better over the substructure than a less stiff layer (see 4.7.4.1).
61
underestimated for pavements with a short distance between the nearside wheelpath and the pavement
edge. When the stiffness moduli are backcalculated from measurements in the nearside wheelpath, higher
deflections due to edge effects will automatically result in underestimated values of the stiffness moduli
when using multi-layer programs, and therefore the risk of underestimation of the strains is reduced.
The extent to which the pavement is susceptive to moisture influences is determined by the drainage
situation. This drainage situation determines how fast precipitation will percolate to the subgrade. Also, in
low terrain with high ground water levels, the drainage condition will influence the capillary water levels
and the sensitivity of the pavement structure to groundwater table variations. The drainage condition is
affected by several factors:
• Permeability of granular road base and subgrade (peat, clay, sand). Materials with a high permeability
appreciably reduce the chance of moisture induced damage.
• Elevation of the road above the surrounding terrain; i.e. on embankment or fills, in level construction or
in cuts. For roads on embankments, drainage is generally easier to ensure than for roads in level
construction or cuts. Drainage can be particularly poor for cuts in low permeability soil.
• Height of the road relative to the highest groundwater level or the water level in the ditches
• Presence of ditches and drains
• Condition of the verge (for example, if the verge is too low, water can remain in it)
• Overall indication of the extent of water discharge of the road
One should be aware that, in the case of a saturated subgrade with a poor pore water pressure dissipation
due to low permeability and/or due to confinement of water (e.g. by frozen layers below), the stiffness
under a FWD load can be quite high while the stiffness and especially the resistance to permanent
deformation under a slowly moving truck can be very low.
On roads that have been strengthened several times, it will be difficult to find out the actual years and
types of maintenance and rehabilitation measures. Compromises are inevitable.
62
SÂ In the Dutch design method, a maximum of two former life stages are used in the calculation.
The first former life stage is the period between construction and the time of the last strengthening
measure; any earlier structural maintenance can be accounted for by using an equivalent number of
standard axle loads that have passed in this first former life stage. The second former life stage is the
period that elapsed since the last strengthening measure. ÁS
The most common way to determine the residual life of a pavement is to calculate strains and stresses in
critical positions in the pavement structure. These strains and stresses are then entered in performance
laws, such as fatigue or permanent deformation laws for the various materials applied in the pavement
structure. Most methods aim at estimating the initial bearing capacity of the pavement, which is then
adjusted for the traffic loads carried already by the pavement. Some other methods aim at a direct
estimation of the residual bearing capacity, so no adjustment for past traffic is required.
As stated before, methods that explicitly correct for the effects of former traffic, aim at calculating the
initial bearing capacity of the pavement, while methods that do not explicitly account for former traffic,
aim at calculating the residual bearing capacity of the pavement at the time of FWD measurement.
Generally several design criteria have to be evaluated to determine which design criterion is prevailing.
SÂ In the Netherlands, the asphalt strain is multiplied by a factor for edge effect if the nearside
wheelpath is close to the pavement edge but the deflections have been measured further away from the
edge. This adjustment factor is 1.1 if the distance ranges between 500 mm and 700 mm and 1.2 if the
63
distance is less than 500 mm. In case of moderate or serious edge damage, the distance to this edge
damage is used. If the deflections from the nearside wheelpath were used for backcalculation, the edge
effect will automatically be accounted for and no strain adjustment will be necessary. Similarly, when
the measurements were conducted on a rutted pavement, the critical asphalt strain value is multiplied
by a factor 1.08 if a rut depth of about 10 mm or more is measured. If a smaller rut depth is measured,
or if the measurement was conducted with a segmented loading plate, no adjustment is applied. ÁS
The critical asphalt strain is entered into a fatigue law to estimate the number of load cycles that the
material is able to sustain. If maintenance was applied in the meantime, the fatigue resistance is sometimes
assessed per life stage.
Many methods use an equation of the following type:
cfat
⎛ 1 ⎞
N fail,AC = c 0 ⋅ ⎜⎜ ⎟⎟
⎝ ε AC ⎠ (4-14)
where Nfail,AC = Resistance of asphalt concrete to failure due to fatigue damage (load cycles)
εAC = Critical asphalt concrete strain (m/m)
c0 = Material parameter
cfat = Material parameter, gradient of fatigue law in this case
Another representation, as e.g. used in the Shell Pavement Design Manual [4-25], also provides a
correlation with the asphalt stiffness modulus according to Figure 4-15. However, this type of fatigue law
shows a similar gradient cfat of the strain-fatigue relationship when examined at constant stiffness values.
1.000
str
ain
[mi
cro
n/
m] 100
N = 10.000
N = 100.000
N = 1.000.000
N = 10.000.000
N = 100.000.000
10
1000 10000 100000
Most methods using this approach, consider the fatigue resistance according to this fatigue law to be the
initial fatigue resistance of new asphalt. Some methods use the same fatigue law to calculate the residual
fatigue resistance of asphalt concrete that has been subjected to fatigue already.
64
SÂ The Danish method uses the same fatigue law for new and old asphalt concrete. By assuming that
the fatigue law provide the residual fatigue resistance directly from the strain level, regardless of the
fatigue history, the need for assessment of former traffic loading is obviated and residual life
determination will be more simple. ÁS
The asphalt mix fatigue law that is used has a large influence on the calculation results and must therefore
be determined as reliably as possible. It is emphasised that, even within identical specifications for the
material composition, great differences in fatigue resistance can be expected. Nevertheless, in many cases
one starts with using a fatigue law based upon former fatigue analysis on similar materials. Visual
inspection of the cores extracted for the sampling and determination of the type of asphalt concrete road
base material and its condition can help in the selection of the fatigue law.
If visual inspection and visual assessment of the drilling cores lead to suspicion of unusual fatigue
properties or if the calculated structural condition on the basis of the chosen standard fatigue law seems
not to be correct (see 4.7.8), more detailed assessment of the fatigue law is necessary. The following
possibilities exist for a better determination of the fatigue laws:
• Calculation, using nomographs, on the basis of the composition of cores taken out of the actual
pavement. Fatigue resistance can be predicted on the basis of e.g. the composition, the bitumen hardness
and void content. The cores, which are drilled to determine the layer thickness, can be used for this
purpose.
• The most reliable and applicable fatigue law is obtained by carrying out fatigue analysis on untrafficked
test specimens from the survey section itself. However, the additional value of such an investigation will
have to be weighted against the costs and time involved. For practical reasons the use of a standard
fatigue law or a fatigue law based on easily measured parameters is generally to be preferred.
The fatigue resistance according to the above-mentioned fatigue law is usually based upon laboratory
tests, where a cyclic load pattern without rest periods is used. Under traffic shorter or longer rest periods
between the passages of axles and vehicles will occur. These rest periods seem to contribute to a better
fatigue resistance than obtained in cyclic laboratory testing, due to a recovery of microdamage in the
bitumen between two successive load pulses. How large this effect is depends amongst on the composition
of the mix, load conditions and temperature. Higher temperatures and longer rest periods contribute to a
greater recovery and subsequently to a greater fatigue resistance.
Together with lateral wander, this so-called 'healing' causes the asphalt structure to perform better under
traffic loads than indicated by the fatigue resistance based on laboratory testing. Adjustments for healing
are usually made by multiplying the fatigue resistance of the existing asphalt pavement according to
laboratory experiments, with an adjustment factor.
SÂ The Dutch design method uses a healing adjustment value of four for asphalt concrete specified
according to the Standard RAW Specifications [4-27]. For the 'richer' mixes used before 1978 this
adjustment factor is supposed to have a value of six to seven. ÁS
From the time that crack initiation occurs in the asphalt concrete road base to the moment that cracks have
propagated far enough to substantially affect the bearing capacity of the pavement, a limited number of
further loads can be carried. In some methods the calculated bearing capacity of the pavement is adjusted
for this effect.
4.6.4.3 Bearing capacity and cement bound road base fatigue cracking
For cement bound road bases, the critical strain at the bottom is evaluated. This evaluation is generally
carried out for a standard wheel load and reference temperature and vehicle speed conditions. This is
65
treated in more detail in sub-chapters 4.4 and 4.5. In some methods, the strains are not determined in
mechanistic calculations but are inferred directly from the deflections (see 4.5).
For some cement bound materials fatigue laws exist that relate the tensile strains or stresses in the cement
bound material to the maximum number of repetitions. Usually these laws are of the form
c2
⎛ 1 ⎞
N fail,CB = c1 ⋅ ⎜⎜ ⎟⎟
⎝ ε CB ⎠ (4-15)
where Nfail,CB = Resistance of cement bound material to failure due to fatigue damage (load cycles)
εCB = Critical tensile strain in cement bound layer (m/m)
c1 = Material parameter
c2 = Material parameter, gradient of fatigue law in this case
The parameter c2 determines the sensitivity for strain level of the cement bound material. For many
cement bound materials this parameter is considerably larger than the value of 4 to 6 often found for
asphalt concrete. For instance the gradient of the fatigue laws for sand cement is much steeper than that of
asphalt (sand cement would rather show a 17th power relation between strains and damage instead of a 4th
power). This means that the weighing of traffic loads with a fourth power law (see 4.6) is not pertinent for
assessing the bearing capacity of a cement bound road base. However, this also means that only the few
very heavy axle loads are of importance. In combination with the large uncertainties and dispersion in
pavement engineering, this can make a bearing capacity calculation less realistic. In these cases another
approach could be followed where the strain under extreme conditions is limited to a sufficiently low
value to prevent cracking problems. Reliability concepts can be incorporated in this approach.
SÂ The Dutch design method incorporates the assumption that for sand cement road bases strains
less than the 50 µm/m threshold value will never result into structural problems. ÁS
The vertical compression at the top of the subgrade is calculated as design parameter. Usually the
calculation of this compression is performed for a standard wheel load and reference temperature and
vehicle speed conditions (see 4.4 and 4.5). In some methods the strains are not determined in mechanistic
calculations but are inferred directly from the deflections (see 4.5). Sometimes adjustments are applied to
the calculated compression values before entering them into the performance laws.
SÂ In the Dutch design method, the critical subgrade compression value is multiplied by a factor
1.06 when a rut depth of about 10 mm or more is measured under a bar of 1.20 m length during FWD
testing. If a smaller rut depth is measured, or if the measurement was conducted with a segment
loading plate, no adjustment is made. ÁS
66
A deformation law for a subgrade material should ideally supply the maximum number of subgrade
compression repetitions until a critical rut depth will be attained. The dispersion over the length of the
road section would then give the longitudinal unevenness or roughness.
However, the most widely used law, which originates from the AASHO Road Test [4-19], only gives a
relation between the compression at the top of the subgrade and the number of equivalent standard axle
loads for obtaining a Present Serviceability Index (PSI) of 2.5. PSI is actually an index in which weighted
rut depth, longitudinal unevenness, cracking and repairs are included. However, the proportion of
unevenness, in both longitudinal and transverse directions, is predominant [4-28]. The value of the
secondary deformation itself cannot be determined from this law. The subgrade compression law
according to the AASHO Road Test is as follows:
4
−7 ⎛ 1 ⎞
N fail,SG = 4.1 x 10 ⋅ ⎜⎜ ⎟⎟
⎝ ε SG ⎠ (4-16)
where Nfail,SG = Resistance of subgrade to failure due to permanent deformation (load cycles)
εSG = Critical compressive strain in subgrade (m/m)
Similar to the subgrade, granular road base layers can fail due to accumulation of permanent deformations
under repeated loading. Similar permanent deformation laws can be used, again using vertical compres-
sion values at the top of the layer. Some approaches use the compressive stresses in the layer as design
criterion.
In practice, heavy traffic does not travel exactly in the same path but the position of the wheels shows a
transverse distribution over the lane width. This effect is beneficial for the pavement bearing capacity.
Some design methods use standard values for this factor, which is often applied to the traffic load (see
4.6). However, the lateral wander effect not only depends on the degree of lateral wander but also on the
layer stiffnesses of the pavement structure and the redesign criterion that used (e.g. asphalt strain or
subgrade compression). Therefore, some other methods conduct a 'tailor made' calculation of the lateral
wander effect which can be applied as an adjustment to the pavement bearing capacity. Some methods use
an adjustment to traffic loadings as discussed in sub-chapter 4.6.4.
where Ntraf,passed = Equivalent number of standard axle loads since construction or rehabilitation
Ntraf,res = Residual equivalent number of standard axle loads to be carried
Nfail = Initial bearing capacity of existing pavement expressed as number of equivalent
standard axle loads to be carried
Mmax = Intervention level of the damage ratio before overlay is necessary (optional)
Fr = Safety factor (optional)
67
The bearing capacity Nfail as presented in Eq. 4-17, can represent the initial bearing capacity based on
cracking of the asphalt, cracking of the cement bound road base or permanent deformation of the granular
road base or subgrade. This equation is called Miner's law and is a summation of damage ratios. As
presented above, the denominator is the same for both left hand terms. Therefore, the equation can be
regarded as a special case of the general Miner approach. However, if the pavement has been overlaid
before or if the residual life calculation is conducted for e.g. different seasons, terms with different
denominators will appear in the left part of the equation. This also applies when the overlay thickness is
calculated (see 4.8).
Explicit values for the intervention level and safety factor are not always used in Eq. 4-18. Very often a
value of 1 is used for the right hand term, while variability and uncertainty are accounted for in other ways
or are not taken into account at all. his will be dealt with further in sub-chapter 4.7.7.
The application of a Miner calculation for determining residual life is sometimes dependent on the
stiffness modulus of the asphalt concrete [4-29]. If the backcalculated stiffness modulus of the asphalt
concrete amounts more than e.g. 50 % of a value considered typical for the material, the procedure
explained above is followed. If the stiffness modulus is less, the asphalt concrete is considered to have no
residual bearing capacity and is treated as a good quality road base.
In this approach, the residual life is usually calculated for each test point. The residual life of the section is
accordingly determined for a preset percentile value (see 4.7.7).
In this approach, the value of the stiffness moduli is analysed for obtaining an indication of the structural
condition of the pavement. After normalisation to reference temperatures and loading frequencies, the
backcalculated stiffness moduli of the individual layers are compared to values that are considered typical
for the material under study. If the stiffness moduli match, the structural integrity of the layer will be
classified as good. A value somewhat below the typical value may indicate some deterioration, while a
much lower value will indicate a poor integrity of the layer. When this approach is used, care should be
taken that the allocation of the stiffness moduli over the individual layers is correct. In automatic
backcalculation, the overestimation of the stiffness of one layer is always compensated by the underesti-
mation of the stiffness modulus of another layer. The fit of the backcalculation process may generally still
68
be good in these cases. In strain calculations, the proportionally higher stiffness of a layer and the
compensation in the next layer have limited effect on the accuracy of the calculated strains.
SÂ In the British design method [4-30], a stiffness modulus of a bituminous bound layer (normalised
to 20°C) is considered to indicate:
• Good integrity if the stiffness modulus is higher than 7000 MPa
• Some deterioration if the stiffness modulus varies between 4000 MPa and 7000 MPa
• Overall poor integrity if the stiffness modulus is less than 4000 MPa ÁS
There are several methods that do not backcalculate stiffness moduli from the deflections but use the
deflections to predict the (residual) pavement bearing capacity in a more direct way. Although these
methods are usually more intended for network level than for project level, two of these methods are
discussed in short below.
N
SNC = 0.04 ⋅ ∑ a i ⋅ h i + SN SG
i =1
(4-19a)
SN SG = 3.51 log(CBR ) − 0,85 log(CBR )2 − 1.43
(4-19b)
To estimate the SNC from deflections, several regression equations have been developed (see Annex E).
In all cases the centre deflection plays an important role here but other deflections are also often used. The
SNC is used in models to predict the moment of crack initiation and the crack propagation (see Figure 4-
16).
69
14 2 100
3
4 90
12
6
expected age at cracking (years)
80
8
10
70
50
6
40
4 30
20
2
10
0
0
0 0.5 1 1.5 0 2 4 6 8 10 12 14 16
Annual traffic loading (million ESA4/lane/yr) Time since cracking initiation (years)
Figure 4-16 Pavement life to crack initiation (left); rate of crack propagation (right)
2899829
log(N10 ) = 7.169 − 0.0074 ⋅ SCI 300 −
SCI 300 ⋅ N10 Y (4-20)
2280264
log(N10 ) = 7.287 − 0.0067 ⋅ SCI 300 −
SCI 300 ⋅ N10 Y (4-21)
where N10 = Cumulative traffic loading at the initiation of cracking (100 kN ESALs)
SCI300 = Surface curvature index
N10Y = Annual number of traffic loading (100 kN ESALs)
SCI300 is the difference between the centre deflection and the deflection recorded at an offset of 300 mm
using a load level of 50 kN, normalised to 20°C, measured on a new pavement or between the wheelpaths
on a existing pavement. By calculating N10 from the SCI300 measured between wheelpaths and subtracting
the traffic carried already, it is possible to derive a residual number of traffic loads that can still be
sustained. The model above has been derived from data of different types of FWDs without conversions.
Similar equations were derived for semi-rigid pavements. It is emphasised that models like this should be
implemented per country or region to account for specific features such as climate, materials and
construction practice.
70
4.6.7 Accounting for variances and uncertainties
There are many sources of uncertainty in pavement evaluation. The backcalculation and normalisation of
structural data and the calculation of critical strains and stresses from these data are influenced by the
limited descriptive capability of the models used. The subjective assumptions by the designer due to
insufficient data play a role too. The fatigue resistance calculation from these strains, again usually
requires subjective selection of fatigue characteristics of the material. Traffic data may also be of limited
reliability. Traffic load data are often limited to permanent or periodic counts, with no axle load spectra
available. If expected mean section values are used for all parameters and no safety margins are included,
the consequence would be (provided that the models used are valid) that there is 50% probability that the
pavement will fail earlier than the calculated residual life predicts. Therefore, in practice many ways are
used to account for variances and uncertainties in the design calculations by application of different
reliability percentages such as 75%, 85% or 95%. However, as the reliability is introduced in different
stages by the various methods, these figures often have different meanings and cannot easily be compared.
SÂ For instance in the Dutch design method for the asphalt cracking criterion, the intervention level
Mmax is calculated from the estimated in-section variations to limit the extent of fatigue damage at the
end of the residual life period. A maximum extent of fatigue damage is defined for several classes of
roads. Usually a maximum extent of fatigue cracking initiation of 15% of the wheel paths length is
used for the calculation of the residual life of primary roads. For other roads however, other values
may be used depending on cost considerations. The aspect of uncertainty in the various estimations is
accounted for separately by means of a safety factor, also called reliability factor. ÁS
SÂ The safety factor in the Dutch design method is calculated from the estimated uncertainties and
serves to limit the risk that the intended intervention level of structural damage is exceeded before the
end of the residual life period. This reliability is set at 85% for primary roads, 75% for secondary
roads and 70% for lower class roads.
As the safety factor in fact reflects how far the calculations can be wrong due to the uncertainties in
the calculation input, it also reflects to which extent differences between calculation and observed
behaviour are explained. The safety factor also sets the limits between which the calculations of the
overlay thickness may be adjusted, on the basis of the observed behaviour. When the difference
between observed and calculated behaviour is more than indicated by the safety factor, further
investigation is necessary. For the subgrade compression criterion, this comparison technique between
calculated and observed behaviour unfortunately is not valid. ÁS
71
4.6.7.3 Determination of residual life as a fractile of calculated pavement residual life distribution
Residual life computations are often performed per test point. This approach provides a frequency
distribution of the residual life, as the deflections vary along the pavement section. To account for these
variations, not the mean residual life is taken but a value with a low underpass frequency. Examples are
the 25-percentile of the log-normal distribution of the calculated residual life, or the mean of the log-
normal residual life distribution minus one time the standard deviation. In equation:
where Ntraf,res,x = x percentile value of residual life (maximum traffic load) (ESAL)
Ntraf,res = Calculated residual life per test point (ESAL)
ξx = Control factor for x percent underpass (-)
mean = Arithmetic mean
std = Standard deviation
To determine the 25-percentile, ξx, or more precise ξ25 should have a value of 0.67 in case of a normal
distribution. When the mean minus one times the standard deviation is used, ξx will obviously have a
value of 1. In a normal distribution this corresponds with an underpass frequency x of about 15%.
This method does account for the variability in the stress-strain properties of the pavement (which are
reflected in the deflection) but does not account for further variations, e.g. in material fatigue behaviour,
nor does it account for uncertainties, as e.g. in traffic data. Therefore, it is hard to formulate the practical
meaning of this underpass frequency.
Although this method indeed gives a safer residual life calculation, the degree of reliability is not clear. It
may be possible that, by piling reliability upon reliability, an unnecessarily conservative residual life value
is calculated. On the other hand, if for instance only an 85% deflection bowl is used while expected mean
values are used for all other design parameters, the reliability of the residual life value will not be 85% but
considerably less.
In geotechnics it is common to multiply the characteristic parameter values with partial safety factors,
derived from full probabilistic calculations, to arrive at an explicit reliability level. This more rational
approach deserves to be considered for pavement analysis purposes too.
72
In principle, preference should be given to a full probabilistic calculation, but these calculations are time-
consuming and complex, because of which they are normally not used for routine purpose. They can be
used for research purposes, for instance for the calculation of maximum damage ratio and/or safety
factors. In geotechnics, full probabilistic calculations are used to determine the partial safety factors that
are applied on the characteristic values of the design parameters in routine calculations. In the Dutch
design method, a simplified probabilistic method is used.
A residual life analysis cannot be based solely on calculations. The results should always be compared to
the visual condition of the pavement. It should be noted that, when safety is included in the calculation
implicitly or explicitly, e.g. by the use of characteristic values, the calculations will in most cases predict a
condition that is worse than the actual condition.
There are several problems involved in comparing calculation results to observed condition:
• Most calculation methods only predict crack initiation at the bottom of the asphalt layer. This cracking is
not necessarily visible at the surface; as some time is necessary for the cracks to propagate to the surface.
Only a part of these cracks will be visible.
• Cracking observed in the wheelpaths can partially consist of surface initiated cracking. For assessing the
residual structural life, only structural cracking is of importance.
• Maintenance may have been applied in the past. Structural maintenance and rehabilitation, but also
surface treatments hamper verification of the predicted fatigue damage in situ. On the other hand, factors
such as water penetration through cracks can appreciably accelerate the development of cracking and the
deformation of the pavement structure.
Many countries have a method for assessment of the extent of cracking. Especially the cracking in the
wheelpaths due to insufficient bearing capacity is of importance for matching the calculated structural
condition. Sealed cracks must be counted too. The visible damage registered on inspection forms can
generally provide valuable clarification in the matching of the condition with the figures produced by the
residual life calculations. In particular, detailed inspection provides additional support when processing
the deflection data. For instance, locally high values of the deflections can often be explained from local
cracking.
As described in sub-chapter 4.7.3.7, cores cut for the backcalculation process will also provide informa-
tion about the type and condition of the asphalt layers and provide an indication of the adhesion between
the layers. The cores are sometimes used for further material investigation. In cases where crack formation
is visible at the pavement surface, additional cores will be cut on these cracks to determine if they are
surface cracks or cracks over the full depth of the asphalt layer. Unfortunately it is not possible to take
cores specifically in places where crack formation has occurred solely at the bottom. Figure 4-17 however
illustrates how, by analysing a line of cores cut close together, attempts can be made to deduce whether
the cracks have originated at the bottom or at the top of the asphalt layer.
73
TOP VIEW
coring scheme
height of crack
SIDE VIEW
cracking from
bottom upwards
SIDE VIEW
cracking from
top down
depth of crack
In some countries, wear of studded tyres can also be of importance for surface unevenness. However, this
phenomenon is not related to permanent deformation.
Permanent deformation of the granular road base and/or subgrade indicates inadequate bearing capacity of
the structure. This form of damage frequently occurs locally, is often accompanied by cracking and can be
caused by:
• Too thin asphalt structures for the given traffic load over the service life of the pavement structure. In
many cases, the compressive strain at the top of the granular layers will be rather high.
• Crack formation in the asphalt structure, inevitably reducing the load distributing capacity and increasing
the compressive strain in the subgrade
• Unsatisfactory groundwater level and/or water susceptibility of the road base and subgrade. Insufficient
drainage may also play a role. Geotechnical investigation and groundwater level measurements can be
supplementary activities in analysing the cause of the damage.
In the second case the secondary deformation is not so much the cause of cracking of the asphalt layer but
rather the result, which means it is an indicator that cracks in the asphalt pavement are of a structural
nature.
74
4.6.8.2 Use of additional information
Supplementary information will often be used in the comparison of calculated to observed condition,
because of the various difficulties involved in the assessment of structural damage. Examples of
supplementary information are:
• File information from construction stage and final inspection
• Maintenance history
• Weather information (including extremely hot summers, severe winters)
• Design details, construction details, type of materials
• Geometry of the road, drainage conditions
• Core material investigation results, auger probes
• Deflection differences between longitudinal measuring paths
Surface initiated cracks and crazes can have a great number of causes. Crazing is usually preceded by the
surface cracking, appearing as longitudinal and transverse cracks of limited length.
75
damage; in this case only a qualitative comparison can be made. Some other methods attempt to predict
the extent of structural damage. This enables a somewhat more quantitative comparison.
In both cases the calculated residual life will probably not be in full accordance with what is observed in
situ. Within certain limits, differences are explainable. There are various sources of uncertainty in the
calculations and the observations. The question arises how far off the difference between calculated and
observed condition is given the reliability of the calculations and observations. If the difference in
predicted and observed pavement behaviour is of the same order of magnitude as the uncertainty of the
residual life calculations, a supplementary analysis of the cause of the damage is usually not necessary. In
this case it is justifiable to make some well-considered adjustments to the calculation parameters, such as
the fatigue law, to improve the degree of correspondence. The purpose of this approach is to obtain
parameter values for the overlay thickness calculation, supplying the best possible match with the
observed condition.
SÂ The Dutch design method provides for a quantitative prediction of the extent of fatigue crack
initiation as a means for a comparison between calculated and observed condition. A field calibration
factor can be derived from the difference between the calculated and observed structural condition,. If
this factor is within the limits indicated by the safety factor as used in Eq. 4-18, this field calibration
factor can be accepted and used in further overlay calculations instead of the safety factor. If the field
calibration factor falls outside the window indicated by the safety factor, the mismatch between
calculations and observation cannot be explained from the uncertainties in the residual life
calculations and further investigation is necessary. ÁS
If the difference in predicted and observed behaviour is greater than can be explained from the uncertainty
of the residual life calculations, a more extensive analysis should be performed. More in-depth investiga-
tion of pavement structure, material and/or traffic data will be necessary to achieve an acceptable
correspondence between calculated residual life and observed condition. This further investigation can
refer to traffic intensities and standard axle loads, but also to assessment of the extent of fatigue damage
actually present at the bottom of the asphalt layer.
In a proper investigation to the cause of the differences between the expected pattern and the damage
observed, assumptions must be checked iteratively and possibly adjusted in order that a final decision can
finally be given with sufficient certainty in respect of structural condition and expected damage
development.
4.7.1 General
Strengthening of an existing pavement is necessary if the residual life is less than the desired residual life.
As a general rule, only one measure is effective to extend the residual structural life, i.e. strengthening the
pavement structure by an asphalt overlay. As the pavement thickness increases by this overlay, the critical
strains in the asphalt concrete and/or cement bound road base and the critical compression in the granular
road base layers and subgrade will decrease, consequently leading to an increased residual bearing
capacity of the pavement structure.
76
The actual effect of strengthening is slightly different for fatigue related damage such as cracking of
asphalt concrete and cement bound road bases on one hand, and permanent deformation of granular layers
on the other hand.
For asphalt fatigue damage, the overlay prevents the initiation of new cracks at the bottom of the asphalt
layer and retards the growth of existing cracks up to the surface. However, the fatigue damage that has
already occurred is not eliminated in this way, and fatigue cracks which already exist, are not eliminated.
If the existing pavement shows extensive fatigue damage, it is even possible that no residual bearing
capacity is left and no overlay thickness can be determined. In this case the existing asphalt layer can be
regarded as a good unbound road base. A new asphalt pavement structure can be redesigned. However,
care should be taken that cracks from the existing asphalt pavement do not reflect into the new pavement.
For subgrade deformation, the existing deformation damage is actually eliminated when an overlay is
applied. The deformation process more or less starts from scratch again, be it at a slower rate due to the
overlay. The damage that has occurred in the former service life period does not have to be accounted for
in the calculation of the overlay thickness. In any case, for deformation damage, an overlay thickness can
be calculated independently of the damage that has occurred already.
For a number of practical considerations, the overlay thickness finally applied will in many cases be
different from the calculated value. This will be discussed in sub-chapter 4.8.6.
Usually, design criteria and design data (structural and traffic data, temperatures, speeds, function and use
of the road, age of the highway structure, carriageway and lane geometry, pavement edge and verge
conditions, moisture conditions etc.) will be similar to those observed in the residual life calculation (see
sub-chapters 4.7.2 and 4.7.3). However, sometimes differences may be found. If e.g. the function of the
road changes for the next service life stage, lanes may be added, extensive repairs can be made before
strengthening the pavement or the drainage condition can be altered. For the overlay design, the evenness
or profile of the road, as appropriate, will also have to be assessed to determine where filling/profiling is
necessary and which (minimum/maximum) layer thicknesses are required for this. If necessary, profile
measurements must be carried out. Filling and/or profiling can influence the overlay thickness. This issue
will be addressed in sub-chapter 4.8.6.
Overlay thickness calculation principally means determination of what extra pavement thickness is needed
to increase the residual bearing capacity of the pavement sufficiently to survive the next service life stage
without facing a condition falling below the intervention level.
For fatigue related damage, again two approaches can be distinguished for determination of the required
overlay thickness from strains or stresses; the Miner calculation based upon the initial bearing capacity,
and the direct calculation of the overlay thickness from the stress or strain level. Both methods require
calculation of the bearing capacity of the strengthened pavement. In the first case this bearing capacity is
considered to represent the initial bearing capacity, which must be adjusted for the dissipation of bearing
capacity already taken place since construction, while in the second case this bearing capacity is
considered to directly represent the residual bearing capacity. The bearing capacity of a strengthened
pavement is calculated along the same lines as followed in the residual life calculation. This was described
in detail in sub-chapter 4.7.4.
77
4.7.3.1 Miner calculation of required overlay thickness for fatigue related damage
In this approach, the initial bearing capacity of the existing pavement is estimated in terms of sustainable
traffic load, and the traffic load already carried. The ratio of passed traffic over initial bearing capacity
shows which percentage of the bearing capacity has been consumed already. In a next step, the required
overlay thickness is determined in a way that the residual bearing capacity is increased sufficiently to
survive the subsequent service life stage. Of course, this is dependent on the number of years that the road
still has to last, the amount of traffic that is expected, the percentage of bearing capacity that has already
been consumed, as well as the intervention level that was chosen for the total bearing capacity consump-
tion at the end of the second service life stage.
The determination of the required overlay thickness is solved iteratively. For each overlay thickness value,
the 'initial' bearing capacity of the strengthened pavement (the bearing capacity the pavement would have
had, if it had been constructed as new in the corresponding total thickness) is calculated. The actual
consumption of bearing capacity after overlaying is calculated as the traffic load after overlaying, divided
by this 'initial' bearing capacity of the strengthened pavement. The iteration process stops at a thickness
where the sum of the consumption of bearing capacity before and after overlaying is sufficiently close to
the maximum percentage of bearing capacity dissipation. The general form of this Miner calculation is:
Again, the intervention level for damage ratio and the safety factor are not always used explicitly in this
calculation.
SÂ In the Dutch method the maximum damage ratio is based on the maximum extent of fatigue
initiation. For overlay design purposes, this extent is usually set 5% higher than for residual life
calculation. For the safety factor, the same value is used as has been applied in the final residual life
calculation. As mentioned in Chapter 7, this safety factor can have been adjusted within certain limits
during the residual life calculation on the basis of the observed practical behaviour, giving it the
character of a field calibration factor. and ÁS
Some countries do not use weighted climatic conditions but calculate the fatigue damage, e.g. per season.
In this case the first and the second term of Eq. 4.23 are composed of several contributions with another
bearing capacity of the pavement for each condition.
4.7.3.2 Direct calculation of required overlay thickness for fatigue related damage
In this approach it is assumed that the residual bearing capacity of the pavement can directly be calculated
from the present stress or strain level meaning that it is not dependent of the fatigue history. The overlay
thickness is calculated directly, by iteratively adapting the layer thickness to arrive at an allowable strain
level, without accounting for the consumption of bearing capacity due to the traffic load that was carried
78
already. This means that the first term in Eq. 4-23 is not used (the procedure is in fact similar to that used
for permanent deformation, as in Eq. 4-24). Usually, the overlay thickness is calculated for each test point
and a percentile value is then used for introducing some degree of reliability.
The resulting overlay thickness is dependent on the number of years in the redesign period, the traffic
intensities and the extent of deformation that is acceptable at the end of the second service life stage.
The determination of the required overlay thickness is an iterative process. For each overlay thickness, the
bearing capacity with respect to permanent deformation is determined in terms of traffic load that can be
carried after overlaying, using the resulting total asphalt thickness. This result is compared to the traffic
load that should be carried in the second service life stage. The iteration process stops at the thickness
where both traffic values agree well enough. The overlay thickness is the difference between the
calculated total thickness required for the second service life stage and the thickness of the original asphalt
layer. The general form of this Miner calculation is:
where Ntraf,expected = Expected traffic load in next service life period (ESAL)
Nfail,strengthened = Initial bearing capacity of strengthened pavement (ESAL)
Mmax,strengthened = Intervention level of damage ratio (ratio of consumed bearing capacity
and total bearing capacity) for strengthened pavement (optional) (-)
Fr = Safety factor (optional) (-)
SÂ In the Dutch method, the maximum damage ratio is set to a value of 1, because the permanent
deformation law used for the subgrade already incorporates an (implicit) limitation of the deformation
to an acceptable intervention level, as it limits the Present Serviceability Index (PSI) (see sub-chapter
8.4.1). The safety factor is the same as used for the residual life calculation ÁS
Sub-chapter 4.7 presented two examples of calculation of the residual life from deflection data; i.e. a
method where the Modified Structural Number (SNC) is estimated from the deflections and a method that
uses the Surface Curvature Index (SCI) (see sub-chapter 4.7.6).
The SNC method as presented in sub-chapter.4.7.6.1 estimates the residual life directly. Therefore, this
method requires no adjustment for former traffic in the residual life and overlay thickness determination
(see sub-chapter 4.8.3). Application of Eq. 4-19 allows easy calculation of which additional layer
thickness is necessary to increase the SNC to the level necessary for the required service life.
The SCI-based method as presented in sub-chapter 4.7.6.2 estimates the initial bearing capacity of the
pavement from measurements between the wheelpaths or on an untrafficked part of the pavement.
Therefore, this method requires adjustment for former traffic in the residual life and overlay thickness
79
determination, by means of a Miner calculation (see sub-chapter 4.8.3.1). The overlay thickness can be
determined by iteratively adapting the pavement thickness in linear elastic multi-layer calculations until
the SCI has decreased sufficiently to arrive at the required bearing capacity (see sub-chapter 4.7.6.2) for
the second service life stage. However, this approach will necessitate modelling of the existing pavement.
Accounting for variances and uncertainties can be accomplished in a similar manner as described in sub-
chapter 4.7.7 by:
• Use of an intervention level for the damage ratio (consumed bearing capacity divided by total bearing
capacity)
• Use of a safety factor
• Determination of the overlay thickness as a percentile of the calculated overlay thickness distribution
• Calculation of a characteristic values
• Full probabilistic calculation accounting for variances and uncertainties
In the case of determination of the overlay thickness as a percentile of the calculated necessary overlay
thickness distribution, a percentile greater than 50 % will now be taken instead of less than 50 %, to
achieve a low frequency of excess (i.e. a low risk that the necessary overlay is thicker than calculated).
The design of the overlay cannot be solely based upon theoretical calculations. Usually a large number of
practical conditions must be observed. As a first consideration, practical minimum and maximum limits
set to the layer thicknesses must be observed. These limits depend on the specific characteristics
(especially on the (maximum) grain size) of the applicable materials. Which materials are applicable,
again depends upon the function of the road, traffic loads and other conditions. Another important
consideration can be the presence of existing cracks. Retarding the growth of existing cracks requires a
certain overlay thickness, which is often greater than the overlay thickness required to prevent initiation of
new cracks.
However, another way to retard the growth of existing cracks to the surface is the use of inlays. In this
case the original wearing course and, if necessary, underlying courses as well, are removed by milling.
Edge damage can also be removed by milling. The structure is then built up again with one or more inlays
and the new wearing course. This method reduces the extra overlay thickness for crack growth retardation
and also corrects all surface damage, and thus prevents cracks that did not start at the bottom of the asphalt
from propagating through the overlay.
If the overlay thickness is high due to a limited number of places in a section where the bearing capacity is
very poor, local reconstruction may be an option to reduce the overall overlay thickness. This may e.g.
implicate local road base improvements. The use of reinforcing geogrids may also be suggested in the case
of very weak subgrade conditions such as peat. Sometimes, soil stabilisation with cement is used to
improve the stiffness and stability of granular layers.
Another important consideration can be driven by unevenness. The remedial action may be combined with
the structural overlay in cases of minor unevenness or where the unevenness does not occur over short
distances. A separate fill/profile layer will be necessary for higher levels of roughness or substantial level
differences.. This can result into a total overlay thickness with much spatial variation. However,
unevenness can also be remedied by milling 'heads' (the highest parts of an uneven pavement). This may
80
implicate that the thickness after application of the overlay is locally hardly greater than the original
thickness and care must be taken that layer thicknesses remains sufficient.
Various other factors can be of influence as well. In urban environment, the thickness of the overlay will
be restricted by kerbs and street furniture. The overlay thickness may also be limited by minimum vehicle
clearance, e.g. under bridges. In these cases existing material needs to be milled replaced by new material.
When this appears to be no option, a decision must be made either to overlay including raising adjacent
objects or to mill and inlay consequently avoiding raising adjacent objects.
The effect of treatments on road drainage should be taken into account too. Care must be taken to ensure
that drainage channels are not blocked by surface treatments.
In summary, the final overlay design requires finding an optimal combination of 'head' millings, filling
lower parts, inlays, possibly local reconstruction and overlay thicknesses, considering:
• Resistance to initiation of new cracks
• Resistance to propagation of existing cracks
• Roughness
• Balance between production of old and use of new material
• Practical limitations of layer thicknesses
• Surroundings and adjacent objects
This section should contain a description of the test site. The length, width and number of lanes in each
direction should be stated. The presence of kerbs, hard shoulders, bus lanes, etc., should also be included.
The general location of the test site, including local authority area or positions relative to network
chainage markers, help to identify the site. A simple sketch showing popular landmarks can sometimes be
useful. An example of this is shown in Figure 4-18.
Where cores, trial pits or GPR have been used to establish the pavement layers, the site test data is usually
best represented in tabular form. The presence of surface defects, such as cracking, rutting and patching,
should be included in this section. Photographs of cores can be very effective.
The drainage method used should be described briefly. The effectiveness of run-off or gullies can be
described using visual inspection of ditch cuttings, gully pots, crossfall, etc.
81
Ch 980
Ch 550 Ch 660
N07 -
0565
FWD Test Site on Old N7
N7 South N7 North
N.T.S
Some of this information will be the same for most FWD surveys. There is usually little change to the
equipment set-up. Much of this information can therefore be located in appendices to the FWD report. The
other test procedures should be clearly outlined to reduce ambiguity.
The means by which the load is impacted on to the pavement using FWD should be described along with
the positions at which the resulting deflections are measured. A simple diagram showing deflection sensor
spacing is useful at this point. The method of measuring pavement temperature (i.e. drilling, depth, etc.)
should also be described.
The test set-up used including number and magnitude of loads should be stated as well as the load used in
the analysis of the pavement. The date and time of testing should also be included in the report. This is
generally more important when testing rigid pavements such as concrete slabs. It may be useful to show
the testing sequence in cases where there were several visits to the site. This could happen where FWD
surveys were carried out at various stages during construction. An example is shown in Table 4-13.
82
4.8.5 Calibration status of equipment
A description of the calibration methods used should be included in the annexes. Both relative and
absolute calibration methods are important. The frequency of calibration is important and should be stated.
The steps involved in converting field data to that which is used in the backcalculation process should be
outlined. The method used to delineate test lengths into homogeneous sub-sections should be explained. A
graphical representation of the FWD data is sometimes useful to illustrate the variations in deflection
behaviour at the various sub-sections. It is also useful to plot FWD deflection parameters that have been
measured on both sides of a road length. This compares the deflection response across the width of the
pavement structure. Once the test length has been delineated into homogeneous sub-sections the bowls
that will be used in the analysis must be chosen. These can sometimes be the average, characteristic or all
the deflection bowls.
The method used in calculating moduli, etc. must be stated. The manner in which the pavement is
modelled is also of paramount importance. It is common practice to model flexible pavements as three
layer structures. Layer 1 is the bituminous bound layer(s), layer 2 is the granular layer(s) and layer 3 is the
subgrade soil. Design assumptions such as fixing the stiffness of thin asphalt layers, should also be clearly
stated. When using software with possibilities of varying layer interface friction, the mode of friction
should be mentioned. When proprietary software is used in the backcalculation process the name and
release number of the package should be given. The computer code used in the software package should
also be stated, e.g. BISAR, WES5.
It is important that all the results of laboratory tests be included in the FWD report. The type of laboratory
tests carried out will vary from job to job. In many cases it is a simple procedure to cut cores from the
bituminous bound layers. These can then be tested in stiffness or fatigue etc. Indirect tensile tests can be
carried out directly on cores to establish the stiffness of the bound layers. The laboratory tests carried out
on granular and subgrade materials are usually compaction and strength related. Again, it is important that
these are reported in a clear manner. A table format is usually useful in reporting trial pit results.
It is a good idea to show the deflection parameters such as centre deflection, d0 - d600 and an outer
deflection in the same graph. This allows the interaction of the deflection parameters to be clearly shown.
Examples of FWD deflection plots are shown in Figures 4-19, 4-20 and 4-21. Some landmarks should be
indicated on the plots to make the deflection plots unambiguous. All of the permanent features along the
test length should also be shown in tabular form as shown in Table 4-14. The average deflection
parameters for the homogeneous sub-sections can be shown in tabular form as shown in Figures 4-22
More detailed statistical analysis can be carried out on the results of each sub-section. An example of this
is shown in Figure 4-23.
83
Table 4-14 Comments of FWD survey
Chainage (km) Comment
0.000 Junction with Marine Road
0.615 Plate marker N31 - 0154
1.000 Hospital Entrance
1.225 Plate Marker N31 - 0159
1.825 Junction with N11
400
300 D1(Roadbase)
200mm
200
100
0
14.5 14.7 14.9 15.1 15.3 15.5 15.7 15.9 16.1 16.3 16.5
Chainage in km
200
Surface Curvature Index Plot (D1 - D2)
Deflection in microns
150
D1 -D2 (Roadbase)
100
50
0
14.5 14.7 14.9 15.1 15.3 15.5 15.7 15.9 16.1 16.3 16.5
Chainage in km
15
D9 (Roadbase)
10
5
CUT FILL
TO TO
FILL CUT
0
14.5 14.7 14.9 15.1 15.3 15.5 15.7 15.9 16.1 16.3 16.5
Chainage in km
84
Ch. rev.
3000
2500
Déflexion [10 mm]
2000
-3
1500
1000
500
0
0.000 0.500 1.000 1.500 2.000
Station [km]
Ch. rev.
3000
2500
Déflexion [10 mm]
2000
-3
1500
1000
500
0
2.000 2.500 3.000 3.500 4.000
Station [km]
Ch. rev.
3000
2500
Déflexion [10 mm]
2000
-3
1500
1000
500
0
4.000 4.500 5.000 5.500 6.000
Station [km]
85
Figure 4-21 Example of FWD deflection plot - Alternative #3
Area:
Location:
Test Information Average Deflection Values Construction Traffic Average Moduli (MPa) Overlay thickness Values ( Comments
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
D1 Bitumino Design
E1 E2 E3 Average 85th
Granular
Test No. (Under D1 -D2 D9(2.1 us Bound Layer T2 traffic (Bitumi (Granula (Subgr Overlay in Percentile
Test Site Length Tests Load) (SCI) m) T1 (mm) (mm) (MSA) nous) r) ade) mm Value Comment
86
Figure 4-23 Statistical presentation of FWD deflections
The pavement structure may or may not be well defined. In the case of newer pavements, details of the
structure should be readily available from the local authority. Older pavements may not have been
formally designed and so the pavement construction has to be measured. This is usually done by coring,
trial pits, or Ground Penetrating Radar (GPR).
The method of presenting structural information may depend on the type of information that has been
obtained. When cores or trial pits have been used it is usually best to represent the data in tabular form
including comments. An example of a table containing core results is shown in Table 4-15. Photographs
can also very useful in showing the condition of cores. Examples of core photographs are shown in
Figures 4-24 and 4-25. Other information such as the results of Ground Penetrating Radar (GPR) should
be presented in graphical form. An example is presented in Figure 4-26.
87
Figure 4-25 Cores from Chainage 3.8 km to Chainage 5.3 km
88
4.8.11 Presentation of derived results
The results of backcalculation analyses can be presented in tabular form. The use of surface moduli was
discussed in sub-chapter 4.4.1 as a means of indicating the overall pavement condition. An example of a
surface modulus plot is shown in Figure 4-27. The various sub-sections can be compared in this manner.
The stiffness moduli and overlay calculations can be summarised in tabular form as shown in Table 4-16.
Stiffness moduli should be rounded to significant digits.
Table 4-16 Summary of calculated stiffness modulus, residual life and overlay thickness
Layer stiffness modulus (MPa) Theoretical
Bound Theoretical overlay
material Granular residual life thickness
@25°C material Subgrade (year) (mm)
Minimum 3400 190 135 4 0
Mean 6500 300 195 20 8
Maximum 10900 460 260 48 34
Standard deviation 2050 85 30 11 10
Coeff. of Variation 0.32 0.28 0.17 0.54 0.29
Backcalculated moduli values can also be plotted to show the variation in calculated values. Examples of
such plots are shown in Figures 4-28, 4-29 and 4-30 for the same set of data. An alternative method of
displaying this information is shown in Figure 4-31. Comparing the measured and calculated deflection
bowls can assess the reliability of the backcalculation process. An example of such a comparison is shown
in Figure 4-32.
89
Residual life calculations and overlay design thickness values can also be shown in graphical format. An
example of a chart, which combines these parameters, is shown in Figure 4-33. Figure 4-34 is an overlay
design plotted against distance along a test length. The 85-percentile value has also been included in this
plot. This type of visual representation is useful when deciding on the most appropriate type of rehabilita-
tion. It can also be useful to show the effect of an overlay application by carrying out a repeat FWD
survey.
La 10000
Bituminous Bound Moduli
ye
r 9000
M E1,MPa
od 8000
uli
in
7000
M
Pa
6000
5000
4000
3000
15 15.5 16 16.5 17 17.5 18 18.5 19 19.5 20
Chainage in km
400
200
0
15 15.5 16 16.5 17 17.5 18 18.5 19 19.5 20
Chainage in km
90
Estimated Stiffness Moduli for Motorway
10000
Stiffness Moduli in MPa
1000
100
E1,MPa
E2,MPa
E3,MPa
10
15 15.5 16 16.5 17 17.5 18 18.5 19 19.5 20
Chainage in km
Figure 4-29 Example of layer stiffness moduli plot - Alternative #2
10000
Stiffness Moduli in MPa
1000
100
E1,MPa
E2,MPa
E3,MPa
Low E1
Low E2
Low E3
10
15 15.5 16 16.5 17 17.5 18 18.5 19 19.5 20
Chainage in km
Figure 4-30 Example of layer stiffness moduli plot - Alternative #3 (with limits)
91
Modules élastiques [MPa]
10000
9000 7840
8000 6482
5612
Modules élastiques [MPa]
7000
6000
5000 E3 Sol
E2 Grave
4000 E1 [25°C]
3000
2000 383
304
1000 279 E1 [25°C]
224 282
0 195 E2 Grave
144
AB11S 121
E3 Sol
HMT16S
Colbase 22 N 20
Type de structure Grave
92
140.0
135
120.0
100
100.0
92
80.0
30
40.0
20.0 0 20
150
Overlay in mm
100
50
0
3.804
3.849
3.895
3.944
3.999
4.049
4.099
4.148
4.222
4.272
4.321
4.371
4.421
4.474
4.525
4.573
4.673
4.724
4.774
4.823
4.876
4.919
4.961
4.999
5.021
5.041
5.073
5.124
5.174
4.62
4.94
4.98
5.3
Chainage in km
The report should contain general comments on the credibility of the backcalculated moduli, overlay, etc.
as compared with the measured deflections. The results of trial pits, etc. should also be used as a means of
93
assessing the trustworthiness of calculated moduli, etc. In locations where the derived parameters do not
appear to concur with physical and visual information further investigation should be carried out.
This sub-chapter includes practical advice on alternative treatments, necessary preparation, and effect of
maintenance choice on adjacent environment. The report should contain advice on the best approach to
maintenance treatment. In many cases there will be rehabilitation work required on sections of the test
length prior to an overall treatment. This work will often take the form of removing and replacing cracked
areas, sealing cracks etc. In many urban environments the thickness of any surface treatment will be
restricted by kerbs and street furniture. In these cases it may be necessary to fully mill existing material
and replace it with new material. It should be pointed out that the material used in overlay layers should be
consistent with recommended nominal layer thickness values for the material.
The effect of treatments on road drainage must also be taken into account. Care must be taken to ensure
that drainage channels are not blocked by surface treatments.
4.8.14 References
4.8.15 Appendices
Detailed results such as deflection data files, should be contained in appendices, apart from the main body
of the report. In the case of large files it may be more economical to make this information available on
disk or any other electronic data information system.
4.9 References
[4-1] Manual for FWD Testing in the Long-Term Pavement Performance Program; SHRP-P-661;
PCS/Law Engineering and Braun Intertec Pavement, Inc; Strategic Highway Research Program.
National Research Council, Washington, DC, 1993.
[4-2] Design Manual for Roads and Bridges - Volume 7: Pavement Design and Maintenance - Non-
destructive Assessment Methods, HD 29/94, The Department of Transport, The Scottish Office
Industry Department, The Welsh Office Y Smyddfa Gymreig, The Department of the Environment
for Northern Ireland, Crown 1994.
[4-3] Stubstad, R., Lukanen, E., Richter, C.A., and Baltzer, S., Calculation of AC layer temperatures from
FWD field data, Proceedings from the Fifth International Conference on the Bearing Capacity of
Roads and Airfields. Trondheim, Norway 6-8 July 1998, Volume 3, pp. 919-928.
[4-4] Specifications for a harmonised European calibration station for improved Falling Weight
Deflectometers measurement of road quality. Final Report European Project SMT4-CT98-5518.
Consortium SpecifiQ. KOAC•WMD. Apeldoorn, November 2001.
[4-5] Van Gurp, C., Characterization of seasonal influences on asphalt pavements with the use of falling
weight deflectometers. Thesis Delft University of Technology. Delft, 1995. ISBN 90-9008036-8.
[4-6] FEHRL Report No. 1996/1. Harmonisation of the Use of Falling Weight Deflectometer on
Pavements, Part 1, Harmonisation of FWD measurements and data processing for flexible road
pavement evaluation. Transport Research Laboratory. Crowthorne, UK. 1996. ISSN 1362-6019.
94
[4-7] Jameson, G.W., Use of FWD to program works based on HDM III; COST 336 workshop on FWD at
Network Level, Lisbon, 4-5 June, 1997.
[4-8] Brunton, J.M., Armitage, R.J., Brown, S.F., Seven Years Experience of Pavement Evaluation, 7th
International Conference on Asphalt Pavements. Vol 3. ISAP. Nottingham. 1992. pp. 17-30.
[4-9] Deflection profile - not a pitfall anymore. Record 17. CROW. Ede, Netherlands. May 1998.
[4-10] Evdorides and Snaith, M., A Knowledge Based Analysis Process for Road Pavement Condition
Assessment, Proceedings, Transport, The Institution of Civil Engineers, 117. United Kingdom.
August 1996, pp 202-210.
[4-11] Interim Report on European Pavement Design Methods. PR/CE/23/97. Unpublished COST 333
Report. 1997.
[4-12] Meier, R.W., Alexander, D.R., and Freeman, R. A Forward Approach to Backcalculation Using
Artificial Neural Networks. Paper No. 970235, TRB 76th Annual Meeting, National Research
Council. Washington, DC. January 12-16, 1997.
[4-13] Guidelines on the Depth of Overlay to be used on Regional and County Roads. Draft NRA
Document. National Road Administration of Ireland. Not Published.
[4-14] Dynatest: ELMOD User Manual.
[4-15] Francken, L, and Clauwaert, C., Characterisation and structural assessment of bound materials for
flexible road structures. Proceedings 6th International Conference on the Structural Design of
Asphalt Pavements, Vol.1. University of Michigan, Ann Arbor, MI. 1987. pp 130-144.
[4-16] Brown, S.F. Determination of Young’s modulus for bituminous materials in pavement design.
Highway Research Record 431. Highway Research Board, National Research Council, Washington
DC, 1973. pp 38-49.
[4-17] Jansson, H., A Simple Structural Index Based on FWD Measurement. Proceedings 4th International
Conference on the Bearing Capacity of Roads and Airfields, Minneapolis, MN. 1994. pp 457-466.
[4-18] COST333 - Development of New Bituminous Pavement Design Method. Final Report, European
Cooperation in the field of Scientific and Technical Research, European Commission Directorate
General Transport, Brussels, 1999.
[4-19] The AASHO Road Test. Special Reports 61A-61E. HRB, National Research Council. Washington
DC, 1961.
[4-20] Richtlinien für die Standardisierung des Straßenoberbaus von Verkehrsflächen - RStO 86,
Forschungsgesellschaft für Straßen und Verkehrswesen, Ausgabe 1986, Fassung 1989.
[4-21] Blab, Ronald, Die Fahrspurverteilung als Einflußgröße bei der Bemessung des Straßenoberbaus,
Mitteilungen des Institutes für Straßenbau und Straßenerhaltung, TU Wien, Heft 5, Vienna, 1995.
[4-22] Impact of heavy freight vehicles, OECD Road Transport Research 1983, OECD Scientific Expert
groups.
[4-23] De Beer, M., Groenendijk, J. and Fisher, C., Three dimensional contact stresses under the
LINTRACK wide base single tyres, measured with the Vehicle-Road Surface Pressure Transducer
Array (VRSPTA) system in South Africa, TU Delft, 1996.
[4-24] Buseck, H., Über die Ergebnisse einer rechnerischen Untersuchung der Straßenbeanspruchung durch
die Fahrzeuge des Schwerverkehrs, BASt, Bergisch Gladbach, Germany, 1988.
[4-25] Shell Pavement Design Manual, Shell Petroleum Company Limited. London, 1978.
[4-26] Paterson, W.D.O., Road deterioration and maintenance effects. Highway Design and Maintenance
Effects, Standard Series. The World Bank, Washington DC, 1987.
[4-27] RAW Standard Conditions of Contract for Works of Civil Engineering Construction. CROW, Ede,
The Netherlands, 2000.
[4-28] Gerritsen, A.H. and Koole, R.C., Seven years' experience with the structural aspects of the Shell
Pavement Design Manual. Proceedings 6th International Conference Structural Design of Asphalt
Pavements. University of Michigan. Ann Arbor, MI, 1987. pp. 94-106.
95
[4-29] Brown, S.F., Tam, W.S. and Brunton, J.M., Structural evaluation and overlay design; analysis and
implementation. Proceedings Sixth International Conference Structural Design of Asphalt
Pavements, Vol. 1. University of Michigan, Ann Arbor, MI. 1987. pp. 1013-1028.
[4-30] Design Manual for Roads and Bridges. UK Department of Transport. Volume 7: Pavement Design
and Maintenance; Section 2: Pavement Maintenance Assessment; Part 3 HA 53/93: Use of the
Falling Weight Deflectometer on Trunk Roads.
[4-31] PARIS. Performance Analysis of Road Infrastructure. EC Project Reference PL95-317.
Rijkswaterstaat, Dienst Weg- en Waterbouwkunde. Delft, 1999.
96
CHAPTER 5
USE OF FWD AT NETWORK LEVEL
Preface
This chapter describes how and when FWDs should be used at network level, and how they can provide
estimates for the timing of maintenance and strengthening requirements to road agencies. This chapter was
prepared by Task Group 2 of COST Action 336.
The information was achieved from a literature survey and a workshop held at Laboratório Nacional de
Engenharia Civil (LNEC) in Lisbon, Portugal in June 1997 and from the experience in the COST 336
member countries. Details of the literature survey and the workshop are reported separately in Annex E to
this report.
The information presented in this chapter should be useful to the member countries to increase the benefit
of their FWDs by using it for both project and network analyses.
97
Table of contents
98
5.1 General
There is an increasing interest in the use of FWD for network level studies to collect more precise
structural values for Pavement Management Systems and finally to optimise road maintenance. Network
versus project level studies and statistical versus systematic monitoring are addressed and used as
framework for this chapter. The different elements in FWD monitoring are briefly described. Only very
few road administrations use FWD for network level studies as daily routine, but more countries are
underway. The procedures described in this chapter shall support and encourage the COST member
countries and others to extend and harmonise their use of FWD for both network and project level studies.
The network level considers the road system usually as a set of roads in the form of clusters or links.
These sets are composed of aggregated classes of roads pre-divided by traffic volume, climate, functional
road classes, condition, geometry, etc. rarely an itinerary. The size of a network is typically of the order of
100 km and over. As a consequence, monitoring structures at network level can thus be defined by the
following objective:
Network level monitoring is a process bound to provide information needed for budgeting and
allocation, and selection and evaluation.
Budgeting and allocation are based on technical/economic analysis and on quantifying needs. Selection
and prioritisation are based on identifying locations, on setting priorities and on scheduling. This
information will be accounted for in Pavement Management Systems.
In brief:
Network Level analysis consists of evaluation of a group of pavements enabling selection of
candidate projects, project scheduling and budget estimates
99
It is of course to be remembered that the project level phase of activities lags behind the network level
management phase. The delay will be typically one year at minimum.
When sampling for project monitoring purposes, homogeneous subsections need to be identified.
Advantage can be drawn from previous monitoring at network level if the latter was conducted in a
continuous mode. Care should be taken, however, that a significant change of road condition does not
occur between the two series of monitoring and also with the actual timing of road works. Comparison of
the two levels can best be perceived through the description presented in Table 5-1.
The network monitoring based on a statistical sampling programme at network level depends on the:
• financing terms that determine the periodicity of monitoring (typically three to five years)
• way in which the network is classified or grouped (by road type and structure, function, traffic level,
overall global condition assessment, administrative structure and geographical location)
Depending on these conditions, sampling modes can be chosen so as to obtain data that are useful for
network level broad economic analysis and planning. Statistical representative sample sizes are usually
small, in the order of 5% of lane length per network category and can make use of, for example, 1 km long
sections randomly distributed over each cluster or link of the network.
For network programming, monitoring should preferably be conducted on a continuous basis. Neverthe-
less, if a statistical sampling approach is used, monitoring deflection should ensure a sufficient degree of
sensitivity to distinguish, in a meaningful way, between the mean values of sections. On the basis of the
desired confidence limit, it is possible to determine the minimum number of randomly located samples to
be measured in each road section.
100
5.4.2 Systematic (continuous) monitoring
Structural monitoring must have the ability to identify homogeneous zones (this might require a
compromise on accuracy versus continuity of measurement) and to provide statistical reduced values of
deflection (associated with weak, average, strong resistant zones) suitable for cost-benefit analysis of
reinforcements. Procedures are used to determine those homogeneous zones such as:
• Moving average deflection and classes (Belgium, Germany)
• Cusum technique (Netherlands)
• Statistical randomness tests (France).
Further information on these examples can be found in Annex E, chapter 4. More general description of
delineation into homogeneous subsections is presented in chapter 4 of this publication.
At network level, peak deflections are of primary concern. The radius of curvature at peak deflection
follows this or some other index associated with the curvature of the deflection bowl. Nevertheless, as it
will be presented in sub-chapter 5.5.1, these and other requirements made on deflection data collected at
FWD network level surveys depend on intended use.
The results of various studies (COST actions) show that deflections measurements in one way or another
form an integral part of PMS. Furthermore it is known that FWDs are most extensively used throughout
Europe and worldwide as deflection recording device. Most recommendations in this guide will remain
valid for FWD and other bearing capacity devices, even with envisaged development of new high-
performance, high-speed devices for determining bearing capacity.
The requirements made on deflection data collected at FWD network surveys depend on intended use.
Whereas an overall economic analysis (Budgeting) does not call for detailed measurements, setting
priorities for maintenance (Prioritisation) may require a set of data close to that achieved in project level
surveys.
101
A number of parameters was identified. For each of these parameters a table has been drawn up with three
levels of details depending on the type of activity envisaged. Comments and considerations are given in
each table to describe the context, limits or requirements of each option.
• Level 1 is the absolute minimum to have any satisfaction (Budgeting)
• Level 2 is anything in between (Selection and Allocation)
• Level 3 is for the most detailed level, equal or close to project level (Prioritisation)
The readers and users should find their way in deciding for one of the levels to their own satisfaction and
should not use this guide as an instruction or manual. Precise instructions and manuals may be found in
the project level guide (see chapter 4) and the calibration guide (see chapter 6).
The various parameters associated with FWD measurements are grouped as follows:
• Parameters linked to the environment (structure, traffic, temperatures...)
• Parameters linked to the measuring method (number of sensors, load applied...),
• Parameters linked to handling of data (post processing, sectioning...).
All these parameters contribute to a greater objectivity of the pavement management systems. The quality
of the system's response is highly dependent upon the quality of the information collected.
Traffic Load
Average daily traffic or number of
Level 1
commercial / heavy goods vehicles
5.5.2.2 Environment
The environment such as solid rock or soft clay, low land or mountains, stable or changing climate affects
the measurement. At Level 1, the network to be analysed should be limited to areas where these
environmental factors are similar. For Level 3, seasonal variations and local conditions should be
considered. The environment issue can be summarised as follows:
102
Environment
Temperature recording
Level 1 Air temperature
For multi lane dual carriageways, measurements should be made, as a minimum, in the heaviest loaded
lane. Other lanes can be measured on special occasions to satisfy particular requirements. The comparison
103
of the heaviest loaded lane with other less loaded lanes can give additional information. When there is a
significant difference in construction or traffic loading between the directions, both directions should be
measured.
The choice of lanes issue can be summarised as follows:
Choice of lanes
Level 1 Heaviest trafficked lane, in one direction only
104
actual test point instead of omitting the actual test point. For concrete slab pavements, the number and the
position of test points are selected according to the methodology of analysis.
The test point spacing issue can be summarised as follows:
Level 2 8
Level 3 12
105
Continuous monitoring is required where programming and prioritisation is required. This mode of
monitoring is necessary in order to identify homogeneous subsections, which are requested to be known
for the purpose of programming and prioritisation.
The type of monitoring issue can be summarised as follows:
Type of monitoring
Level 1 Statistical sampling or continuous at high spacing
Level 3 Continuous
Frequency of testing
Level 1 Only sections required by pavement condition
106
Make and type of FWD
Level 1 Establish correlation if make and/or type are different
Number of deflection
sensors in analysis One sensor
Level 1
Centre deflection
Two sensors
Level 2 Centre deflection and offset for calculation of
surface curvature index
Six sensors
Level 3 Centre deflection and five other offsets allowing
backcalculation of layer stiffness moduli
107
Position of deflection sensors
Level 1 Centre
Level 2 Required
Level 3 Beneficial
108
Layer thickness and type of material
Level 1 Type characterisation only
Output data
Level 1 Centre deflection
Structural Number,
Level 2 Surface Curvature Index or
equivalent indicators
Table 5-2 shows the major parameters used in the PMS presented at the workshop, more details are
presented in Annex E.4.
109
Table 5-2 Major parameters in use at network level.
PMS Test point Centre SCI Bearing Residual life Overlay
Country spacing deflection and/or Capacity from need
(m) SN Ratio E-Moduli
Croatia 4 X X
Denmark 200 X X X
Finland 50 - 500 X
France 4 X X X
Hong Kong 50 X X
Ireland 50 X X
Netherlands X X X
Philippines 500 X X
Spain 200 X
Switzerland 25 - 50 X X X
UK 4 X X X
USA 800 X X
* white numbers: Deflectograph used as monitoring device
The parameters range from the simple use of centre deflection only over centre deflections together with
traffic load and asphalt temperature, to the use of deflection differences as structural curvature index (SCI)
or simple formulas which use the whole deflection bowl to calculate Structural Numbers (SN) or other
structural indicators. Some systems even use surface or layer moduli or residual structural life from project
level post processing.
For the 50 kN load (indicative for 100 kN axle load) a very general rule of thumb is that centre deflections
d0 below 100-150 µm are indicative of very good structural condition and values over 1000-1500 µm
indicate very poor structural condition. For these rough estimates the average daily temperature during
measurements should be within 10°C to 25°C.
5.7 Examples
This section provides good practical examples of use of the FWD at the network level and today's
Pavement Management Systems. More detailed descriptions are given in Annex E.
5.7.1 Spain
Spain has a Pavement Management System that is used for state road network and for a number of
regional road networks. It gives maintenance actions and budgets, both entire network estimates or broken
down per road segments (defined by pavement age, traffic volume and pavement type). No reference was
given in the presentation or in the copies about a distribution of a central budget over regions or sub-
networks. The curviameter is used at the state network level. For systematic surveys and at regional and
local government networks, the FWD is usually used for the same purpose. On project level Lacroix and
FWD are used for flexible and semi-rigid pavements and FWD for rigid pavements. If FWD is used for
input, measurements are required at every 200 m. This can be per lane or direction, probably depending
upon road width and such. Survey period for deflections is four years. It uses empirical deterioration
models and threshold values. For cracks it uses HDM.
110
5.7.2 Finland
Finland has a separate Highway Investment Programming System for planning purposes. For structural
assessment it uses KUAB FWD data, measured at 10 to 20 points per section of 1 to 10 km, but not further
apart than 500 m. The test points are chosen systematically. The measuring cycle is three to five years,
depending on the variability of the former measurement. It should be noted that the same data are also
intended for the district level Pavement Management System and even for project level. Main derived
structural parameter is the spring Bearing Capacity Ratio, determined as 160/d0 (deflection in µm) and a
target value depending on design standards and cumulative axle loads. It is only good at network level. A
problem is that the measurement is done in summer; this requires a correction factor (0.4, 0.6 or 0.8)
depending on observed frost behaviour. A statistical Markov process gives, per type of pavement per
traffic volume class, an optimal long time funding level and an optimal short term funding level, of course
observing other kinds of distress (rut depth, IRI, surface defects) and considering budgetary constraints
and economic indicators. No prediction model for the deterioration of the BCR is used, but the probability
of a change in condition class represents the deterioration. It is planned to change to SCI instead of 160/d0.
5.7.3 Denmark
Denmark has been using PMS since 1988. Deflections are generated by FWD measurements. For the
minor road network they use the centre deflection only for pre-evaluation of the possible need for more
detailed pavement evaluation. The combination of deflection and ADT determines whether the bearing
capacity is rated good, fair or poor. The rather small major network is measured and investigated in detail,
one could say on project level. Construction details are obtained from trial pits at each 300 m and are
updated with contract overlay thickness. Asphalt thickness is considered as not so important. It was argued
that some kind of "structural thickness" would solve a lot of problems. It was asked whether GPR would
give this. Deflections are measured at 200 m distance in both directions, staggered 100 m. Moduli are
backcalculated in a simplified way (by using equivalent thickness approach). These data and other
strength data are fed into the road database. Residual life and overlay need (from chosen design period)
are directly computed for each test point. As representative residual life for a section the 25 % value of
individual point values is taken; for overlay thickness the 75 % is taken. Future development of residual
life and overlay need is calculated from a (verified) theoretical degradation model. Tests are carried out on
sections with expected residual life of five years or less and on sections overlaid last year. The network
level information (necessary budgets; consequences of restricted budgets and such) for the major road
network comes from an aggregation of project level information.
The system applied by ARRB Traffic Research from Australia uses d0, d900 and d1500 deflections to
estimate SNC from deflections for use in HDM III system. Temperature correction requires asphalt
thickness information. Surveys are done every three years with a spacing of 50 m (Hong Kong) to 500 m
(Philippines). Actual network level application (global analysis, planning, programming, prioritisation) is
unclear. See sub-chapter E.4.3 of Annex E for details.
5.7.5 France
France carried out a systematic sample (2 km per 10 km) Lacroix measurement of the highway network in
the mid-sixties. Distance between two measurements was 4 m. From this, and traffic data, followed a
global analysis of the reinforcement needs (by classification into four maintenance needs) and a works
prioritisation for a period of five years. Recently, a new highway PMS has been introduced which is used
111
both for network and project level. Again, no sub-networks are discerned. The system uses central
deflection and radius of curvature, if available, to confirm visual condition in the network analysis.
Characteristic value are taken, being mean plus two times the standard deviation. Threshold values are
given for d0 and SCI depending on traffic class and maintenance. When a pavement is confirmed to be
structurally weak, it can be subjected to complementary studies, but the decision of reinforcement can also
be taken directly.
England does a full coverage of the network with the UK Deflectograph at 3.5 m intervals in a three years
cycle. Using normalised peak deflection (85 % per 100 m length), temperature, pavement type and traffic,
a residual life and overlay thickness is determined for each section. This is based upon empirical
deterioration models. Global analyses and planning are again an aggregation of these data, which can also
be used for programming, prioritisation and detection of further investigation needs.
In the United States, 4 out of 50 states do rough evaluation of bearing capacity from two FWD measure-
ments per mile. Most assessment is based upon visual observation of cracking.
5.7.8 Netherlands
The common system in the Netherlands does not use deflections. It uses visually observed distress that can
be entered in dimensionless deterioration models to arrive at a residual life and prioritisation. Some
provinces use deflection based systems. An elaborate system has been proposed by SHRP-NL. It gives a
panel rating based classification of the structural condition (10 classes from very poor to excellent) from
tables where the traffic intensity, the SCI from between the wheel paths, the ratio SCI in wheel path and
SCI between wheelpaths, and the visually observed degree of cracking is entered. It can be considered
applicable for programming and prioritisation.
5.7.9 Croatia
The project described by representative from Croatia used SNC calculated from d0 from Lacroix in the
HDM model. This model predicts damage development. It is again a section analysis that can be
aggregated to network. The optimum maintenance measure and strategy/timing is determined for each
section. User costs were included. Net present value of savings (compared to "do nothing") was optimised.
5.7.10 Ireland
Ireland is developing some criteria for classification of pavement bearing capacity (d1 and d1-d2 in five
classes: strong to poor), subgrade bearing capacity (d9: six classes; very stiff to very weak) and overlay
thickness for cost estimation (d1 and Traffic → thickness). They use 200 m sections with measuring
distances of 25 to 50 m. There is a National Road Needs Database under development that will also use
ARAN and SCRIM data, as a basis for future maintenance strategy decisions.
112
5.7.11 Switzerland
The Swiss study consisted of a detailed (project level) assessment (including backcalculation of moduli)
of the residual life of 140 km of road in one canton, as a first step for creating a road database. FWD
measurements were taken at 25 m intervals in each direction. Thicknesses were GPR measured in 0.5 m
intervals at 40 km/h for assessment of mean layer thicknesses per 25 m.
5.7.12 Summary
In brief can be concluded that it is hard to derive information for FWD requirements for planning
purposes from the present practice as reported by the participants. There are several reasons for this:
• Countries that perform network level analyses, often do not explicitly discern the four levels of network
analysis proposed by COST 336, and/or do not perform these analyses separately but more or less
integrally.
• A number of countries perform systematic measurements of the complete road network, often with other
deflection measuring devices than FWD, obtaining such deflection data that these can also be used for
programming and prioritisation (and in some cases even for project level decisions). Even when the
systems that interpret these data for global analysis/planning are separate from those used for
programming/prioritisation, it is not easy to determine which data would have been sufficient for
planning only. Often however, there is only one system, for compiling the global analysis and planning
on the basis of an aggregation of programming and prioritisation results, making it virtually impossible
to see a difference in requirements.
• This sensitivity of the requirements to the context of the assessment strategy, also means that one should
be very careful with comparing the tabular data presented in Appendix 4 of the Short Report of the
seminar.
113
For network level information, the maximum deflection is sufficient (at project level it is necessary to
know the shape of the deflection bowl). From the questionnaires, the average spacing of test points appear
to be close to 130 m with values ranging from 20 to 500 m, more often from 50 to 200 m.
The main difference between FWD and rolling deflection measuring equipment results from the fact that,
to have the same daily capacity, the distance between two consecutive test measurement for the FWD has
to be from 50 to 200m which is too long to make it suitable to define homogeneous sections as it is the
case for the rolling deflection measuring equipment with a sampling distance typically 3-5m.
Unfortunately the answers to the questionnaire do not provide suggestions for the maximum acceptable
sampling spacing.
The outcome of the COST 325 questionnaire stresses the need that development of new equipment should
address technical and economic issues as well as traffic safety.
For the network level measurements it is sufficient to measure the air and pavement surface temperatures.
The use of Ground Penetrating Radar could complete information on pavement structure (layer thickness).
5.9 Workshop
The FWD seminar was organised as a workshop at the LNEC premises in Lisbon Portugal, 4-5 June 1997
and supported by a separate financial grant from the European Union. The detailed use of FWD and other
deflections measuring systems at Network level in nine member countries was presented and discussed. In
addition a special EC grant had made it possible to permit presentations of systems used in USA and
systems developed by the Australian Road research Board (ARRB) for use in the Far East. The United
States, FHWA (Federal Highway Administration) explained that only four of the 52 states use FWD
equipment at network level. FHWA encourage the states to use structural parameters as part of their PMS
systems and expressed that Europe is far ahead of USA in this area. From Australia, ARRB (Australian
Road Research Board) Transport Research told that Falling Weight Deflectometers are not yet used for
network level in Australia. The systems presented are developed by ARRB and were used by road
administrations in the Philippines and in Hong Kong.
114
5.10 Network information gathering report
From April 1996 until 1999 information gathering was performed parallel to the organisation and
completion of the workshop. The first draft has been available since April 1998 for download from COST
336 homepage and an updated version will be available at the end of this action. The final information was
added in April 1999 after announcement at the FWD Users mailing list.
Questioning the IRRD-OECD and TRB-TRIS databases started the literature survey. The literature survey
was started in January 1997 by questioning the IRRD-OECD and TRB-TRIS databases. The key words
that were used were: Deflection, Deflectograph, Bearing Capacity and Network.
A second questioning updated the list in May 1999. More over an electronic mail was send to the FWD
user group to call for paper about the subject. At least, some papers were added directly from the members
of the COST action. In total 41 titles were analysed whereas only 11 different papers dealt with a method,
that was actually specific to the network level. Only papers with abstracts in English language were
selected. The list of this papers is provided in Annex E. and they are number from [1] to [41].
From the COST 324 (Long-Term Performance of Road Pavements) final report valuable information are
achieved on inter-calibration of deflection measurements. Deflections were compared amongst FWD,
Benkelman Beam, Curviameter and Lacroix Deflectograph.
The COST 325 (Road Monitoring Equipment) final report, chapter 5.2 presents the results from a
questionnaire. Some answers on the use of FWD at network level are found in 5.2.1 (Aim of Bearing
Capacity Data Collection), in 5.2.2. (Methods for Evaluating the Bearing Capacity of Roads) and in 5.2.3
(Measurement with Benkelman or FWD).
The Workshop in Lisbon June 1997 was a major part of the information gathering exercise. Supplemen-
tary COST 336 information includes the Task Group 1 questionnaire (see Annex D), the Short Workshop
report and the COST 336 home pages The FWD owners list which was earlier appendix 3 of the
Information Gathering report is copied into Annex G.
115
CHAPTER 6
CALIBRATION OF FALLING WEIGHT DEFLECTOMETERS
Preface
The research described in this report was carried out by all members of Task Group 3 ‘Calibration of Falling
Weight Deflectometers (FWDs)’ of COST Action 336 ‘Falling Weight Deflectometer’ of the European
Commission. The main task of Task Group 3 was to prepare adequate procedures for FWD calibration, and
to specify the requirements to be set to an FWD calibration station. For this purpose the state of the art of
FWD calibration was established and presented via a COST 336 Study Contract. The data and procedures
collected and derived in that study, and the FWD calibration procedures and protocols used by the various
manufacturers and CROW in the Netherlands served as basic input for the preparation of the FWD
calibration procedures to be issued under the aegis of COST 336.
The work was continued in the project 'SpecifiQ' (Specifications for a harmonised European calibration
station for improved FWD measurement of road quality) funded by the European Commission under the
4FP scheme of 'Standards, Measurements and Testing' under contract number SMT4-CT98-5518 from
1 March 1999 to 1 March 2001.
This chapter is based on the chapter originally launched by COST 336 in their first release but updated
where necessary on the basis of the outcome of the SpecifiQ project. The numbering and layout of the set
of FWD calibration procedures was adapted to the format used by the European Committee for Normali-
sation (CEN).
116
Table of Contents
117
6.1 General
6.1.1 History
For any structural (re)design procedure it is absolute necessary that the registered data are accurate and
unbiased. Inaccurate and biased data lead to incorrect conclusions about the structural condition of road
and airfield pavements. This in turn can imply that incorrect estimates are made of the remaining life.
Improperly calibrated FWDs, errors in testing and data collection, inaccuracies in normalisation and
adjustment approaches for magnitude and duration of FWD loads, are all, among other parameters factors
of influence in this process. A systematic error of 5% in deflection coincides with a similar error in critical
asphalt strain, but with an error of 25% in terms of pavement life.
The above given example explicitly shows that proper FWD calibration is an absolute need to avoid wasting
scarce pavement maintenance and rehabilitation resources. One of the means to overcome this problem is by
appropriate calibration procedures. Not much attention was given to the issue of FWD calibration in the early
years of FWD. All major manufacturers had their own FWD calibration methods that satisfied the users'
needs at that time. However, with the employment of various local FWD designs, the feeling of anxiety arose
that inconsistent data were gathered.
In the late eighties, when the basic components of the FWD were in the development stage, manufacturers
were not yet convinced about their accuracy. Therefore, quite some effort was spent on calibration of the
FWD test. At that time however, customers showed little interest in calibration and hardly any interest at all
in the quality of the calibration. For this reason the development of the calibration procedures took a 10 to 15
years break, and even then the use of the existing procedures was not extensive. Lately, the customers'
interest increased significantly. This is due to the growing popularity of the FWD as deflection testing device
and the attention given nowadays to quality control and quality assurance. Instrument and equipment
calibration form part of these activities.
Also independent organisations, such as the US Strategic Highway Research Program and the Netherlands
CROW Information and Technology Centre for Transport and Infrastructure initiated studies for the
development of appropriate FWD calibration procedures. This led to the issue of several guidelines for FWD
calibration. Experience with the available calibration procedures revealed that separate development per
country or continent would eventually lead to procedures that would depart too much from each other. To
guarantee an acceptable approach for Europe, it was considered necessary, if not obligatory, to inventory all
existing approaches and to review them for development of a uniform set of protocols.
To obtain and develop an acceptable set of FWD calibration procedures and protocols, a multi-stage
approach was used. This approach encompassed the following stages:
• Collecting available FWD calibration procedures worldwide
• Drafting of a checklist for review purposes
• Review of the collected information
• Development of missing FWD calibration procedures
• Set-up of FWD calibration system
• Drafting of FWD calibration protocols
118
• Listing of FWD calibration station requirements
The following sub-chapters will present the various stages of the development of the FWD calibration
system. In some cases, only a summary of the activities deployed is given, because the full content of that
subject is presented in the annexes.
Quite specific terminology is sometimes used in FWD calibration. This terminology and all definitions
applied are explained in sub-chapter 6.2.
Sub-chapter 6.3 summarises the state-of-the-art study. In this state-of-the-art survey all currently available
FWD calibration procedures were reviewed on their usefulness, applicability and completeness. This
review was performed on the basis of a checklist. This checklist was supposed to contain all necessary
calibration procedures and components.
Sub-chapter 6.4 describes the development of FWD calibration procedures that were considered to be vital
elements of the calibration system, but appeared to be missing or not documented in the available
information. Special attention should be paid to the investigation of the feasibility of servo-accelerometers
for in-situ reference calibration for deflection sensors not dismounted from their holders on the raise/lower
bar.
Combining the findings presented in the sub-chapters 6.3 and 6.4 provided the elements of the pan-
European FWD calibration system. The system developed essentially consists of two levels of calibration
to reduce the calibration activities to a minimum. The first level consists of rather simple and fast
calibration and verification actions to be performed by each FWD user at their home base or at a specific
NDT project. The second level encompasses all calibration procedures usually to be performed by the
independent FWD calibration station. At this level, strict impartiality is required to guarantee that the
calibration findings will be accepted by all parties involved in FWD testing. This level mostly requires
skilled manpower and more expensive instrumentation.
Sub-chapter 6.5 will present the details of the calibration system. This sub-chapter also presents the
calibration scheme with flow charts that disclose frequency of calibration and interaction between
procedures quite clearly.
Protocols have been written for each separate calibration procedure. These protocols describe in detail the
steps to be followed to calibrate or to verify accuracy of an FWD component or entire FWD unit. The
protocols can be found in Annex F to this report. Each protocol, in the lay-out of a pre-draft version of a
European Standard, addresses scope, referenced documents, apparatus, preparation (when applicable),
procedure, analysis, symbols, equations and report. Whereas the individual protocols list the type of
apparatus needed for that specific procedure, a description of the FWD calibration station requirements is
given in sub-chapter 6.6.
119
and 'reproducibility'. All these terms will be described below. For all other terms one is referred to the
glossary of terms annexed to this report (see Annex I). The last section of this sub-chapter presents the
various parties involved in FWD calibration.
When the instrument under test is compared to readings of reference instrumentation or to readings of
similar instruments or equipment, the results of the instruments may agree within specified limits or not.
When the test results are within the specified limits, then the instrument under test is accepted for its
designed use. If the results of the instruments deviate more than allowed from each other, the source of the
problem must be identified, and corrective activities have to be performed. After this step, the instrument
may be compared again against reference instrumentation or readings of same types of instruments. This
might be repeated until the results agree within specified limits. In other words, in any calibration action
the gain factors or calibration factors of the instrument under test are adjusted to match the readings of the
instrument under test to those of reference instrumentation or to the mean of a set of readings of similar
type of instrument.
In the approach of verification, no corrective activities are performed or only fine-adjustment within the
specified maximum error band is applied. If e.g. we check the mounting offsets of the deflections sensors
along the raise/lower bar and we find out that some sensors are slightly incorrectly mounted, than the
shifting of the sensors to the correct offsets is allowed as part of the verification procedure.
Reference calibration is the technique where the response of the instrument or device under test is matched
to the output of a reference instrument or device. This reference instrumentation may not form part of the
FWD or instrument under test. For instance, matching the readings of the FWD load cell to the readings of
a reference load platform is defined as reference calibration. If the results of the two instruments deviate
more than allowed from each other, corrective activities have to be taken.
Relative calibration is the technique used to fine-adjust within allowed tolerances the response of each
similar type of instrument under test, so that equivalent measurements are obtained. A direct result of this
procedure is usually the determination of a set of multipliers necessary to keep the measurements derived
from the instruments equivalent. Placing FWD deflection sensors in a vertical stand and comparing the
response of each deflection sensor with e.g. the mean reading is an example of relative calibration.
Repeatability is defined as the capacity of an FWD to produce consistent results on a specific test site in a
sequence of multiple drops, provided that the pavement is in sound condition, that the support of the
subgrade is not too poor and that temperature or other weather factors do not change significantly during
the test. Distinction can be made between short-term repeatability (usually abbreviated to repeatability)
and long-term repeatability. Long-term repeatability is defined as the capacity of an FWD to produce
consistent results on a specific test site for each day of testing under identical climatic conditions. Testing
of long-term repeatability requires time intervals of weeks to months. Readings are considered to be
repeatable when the variation tolerance is complied with.
Test results are defined as reproducible when various FWD instruments or FWD devices are capable of
reproducing readings at a specific test site under identical testing conditions, even when instruments and
120
devices are operated by different crews using different types or makes of instrument and device. Lack of
reproducibility usually asks for adequate reference or relative calibration actions.
Basically two parties can be distinguished in FWD calibration programmes, i.e. the FWD user and the
independent calibration station. In some cases the FWD manufacturer may play a role in the process. The
following sub-chapters will reveal the role of each party in the calibration scheme.
The main task of the calibration station is to perform those calibration activities that require impartiality.
Additionally, the calibration station should perform calibration activities that supersede the capacity of the
FWD user, simply because analysis of the test data is too complicated or due to use of expensive and/or
complicated instrumentation and data acquisition.
121
• Editing collected information on the basis of aspects and criteria listed in the check list and table of
contents
• Review and comment edited information
• Extracting salient aspects of existing procedures and listing omissions
• Drafting recommendations for future research in near and far future
• Preparation of report.
Written and verbal information was obtained from the following resources:
• Phønix Pavement Consultants at Vejen, Denmark (currently Carl Bro Pavement Consultants at Kolding,
Denmark)
• CROW Information and Technology Centre for Transport and Infrastructure at Ede, The Netherlands
• Federal Highway Administration, Strategic Highway Research Program at Washington, DC, USA
• Foundation Mechanics Inc. at El Segundo, CA, USA
• Österreichisches Forschungs- und Prüfzentrum Arsenal Ges.m.b.H. at Vienna, Austria
• Dynatest International A/S at Glostrup, Denmark
• KUAB Konsult & Utveckling AB at Rättvik, Sweden
• Viagroup SA (currenly Infralab SA) at Romanel/Lausanne, Switzerland
• KOAC•WMD Dutch Road Research Laboratories BV (currently KOAC•NPC) at Apeldoorn, The
Netherlands
A checklist was set up that had to serve as table of contents of the state-of-the-art reporting and main
structure for the system of FWD calibration procedures to be developed. Briefly, the following calibration
aspects were considered to be of vital importance:
• Calibration of dismounted load cell
• Calibration of mounted load cell
• Reference calibration of dismounted deflection sensor
• Reference verification of deflection sensor mounted to the raise/lower bar
• Relative calibration of deflection sensors
• Short-term repeatability of load
• Short-term repeatability of deflections
• Long-term repeatability of deflections
• Calibration of pavement temperature probe
• Calibration of infrared sensor
• Calibration of oedometer
• Verification of deflection sensor offsets
For each aspect of the gathered information, the following steps were performed to assess the collected
data:
• General description of the calibration or verification procedure
• Highlighting of details of the calibration or verification procedure (if applicable)
• Investigation of traceability to standards
• Review with emphasis on assessment of completeness, applicability and omissions
At the end of the survey, recommendations for future (short-term, medium-term and long-term) research
were drafted. The main findings of the state-of-the-art survey are presented in the following sub-chapters,
broken down per main FWD component.
122
6.3.1 Load cell
The collected information presented valuable information for drafting protocols for dynamic calibration of
FWD load cells with the use of a reference load platform. However, the material gathered showed that
hardly any institute or manufacturer performed a calibration test on dismounted load cells. Most
organisations relied on the data supplied by the load cell manufacturer. Not all inventoried procedures stated
how the actual FWD load calibration data were processed. The most extensive and best-documented
information was retrieved from SHRP and CROW.
Some dynamic load cell calibration procedures use test pits whereas others use portable units. Both may
serve the objective of FWD load cell calibration. The advantage of the portable unit is that it may be carried
to any test site in the field.
Two approaches of reference calibration of deflection sensors could be discerned. One approach was that of
tabletop calibration with vibration tables; the other that of the SHRP procedure with the concrete block and
the aluminium beam (for details, see Part 7 of Annex F). The first approach has advantages over the second
approach. The vibration tables allow investigation of the effect of various combinations of deflection pulse
duration and peak value of deflection. This cannot be achieved via the SHRP approach as in this approach the
FWD is used as load generator, implicating that the FWD dynamics and the local pavement structure and
subgrade govern the deflection pulse duration. On the other hand, the data of the SHRP approach is simpler
to evaluate.
All inventoried relative calibration procedures for deflection sensors seemed to have their basis in the US
SHRP calibration procedure. In the SHRP approach, deflection sensors are rotated in the stand, whereas in
the CROW approach, deflection sensors are not rotated in the stand. Simplification is the main idea behind
this second approach. Both procedures were well documented and did not require special skills or equipment.
Choice between the two procedures depends on what objectives need to be met in the relative calibration.
Since the main objectives of the relative calibration procedure were that it should be a fast tool for evaluation
of accuracy of deflection sensors, and it should be easy-to-use by FWD users, preference was given to the
structure of the CROW approach.
CROW provided tests for verification of repeatability of FWD load and deflections. The procedures were
easy-to-use and could be performed on a daily basis when necessary. The repeatability test shows whether
the FWD is capable of producing consistent data at a given test site. All procedures were copied into the
FWD calibration system described in this report.
CROW also provided a test to determine what they define as 'FWD field calibration factor'. In this test, a
group of FWDs (in The Netherlands usually more than ten) measure deflections at a wide variety of test
sites. On the basis of the results the 'field calibration factor' is determined. Measured deflections need to be
multiplied by this factor to convert them to 'standard' deflections. CROW claims to obtain satisfying results
with the approach. In all the years that they had used it, the number of FWDs producing non-reproducible
results diminished year by year. However, they recognised that the reproducibility experiment is less suited
for use with a group of FWDs containing a mixture of devices producing long load pulses (e.g. KUAB) and
devices producing short load pulses (e.g. Dynatest, Phønix, Carl Bro, JILS). These pulses work out
differently on various types of subgrade. The procedure gives insight in differences among FWDs. Because
of the inaccuracies generated by incorrect positioning at each test site and variation of temperature between
first and last visiting FWD, the procedure was not really regarded as a calibration action. The 'field
calibration factor' was considered more to be a kind of harmonisation factor, reason for copying the CROW
approach with some amendments and having the procedure labelled as optional rather as obligatory.
123
No procedure was found to calibrate or verify accuracy of deflection sensors while they were mounted in
their holders on the raise/lower bar. This issue was regarded to need further investigation. For this reason, a
COST Study Contract was used to investigate the feasibility of servo-accelerometers for in-situ calibration of
deflection sensors mounted in their holders on the raise/lower bar. This subject will be addressed in sub-
chapter 6.4.2.
Not many temperature probe calibration procedures appeared to be available, and certainly not so well
documented as the procedures for calibration of load cell and deflection sensor. From the information
inventoried, it seemed feasible to calibrate the temperature probe at two stable temperatures, one low (0°C
to 5°C) and one high (35°C – 45°C). The calibration factor is determined on the basis of readings provided
by the FWD temperature probe under test and those provided by a high accuracy reference temperature-
measuring instrument.
The same procedure may be applied to calibration of infrared temperature sensors. Experience has
indicated that unless performed carefully, calibration of infrared sensors may often lead to poorer, less
accurate results than when the nominal sensitivity as supplied by the manufacturer of the sensor is used. For
this reason no calibration procedure for the infrared sensor was written in COST Action 336.
6.3.4 Oedometer
The approaches for calibration of the oedometer of the FWD or the towing vehicle were not very well
documented and certainly did not comply with the level of accuracy and reliability of the FWD components.
In the reviewed approaches, standard tape measures were used as reference. Elongation of the tape under
tensile stress and inaccuracies in setting out the reference section, may lead to systematic errors of 0.2
percent. Since this degree of accuracy is sufficient for the purpose of distance measuring, the available
procedures were more or less copied.
Only rough information was available for the simple verification procedure for setting deflection sensors
to the correct offsets along the raise/lower bar. In all cases, standard tape measures and folding rulers were
used as reference. Since they were considered accurate enough to measure distances over a maximum
distance of 2.5 m, the current approaches were used as input into the calibration system.
124
action. The following sub-chapters describe how the two categories of problems were resolved in the
framework of COST 336.
Sub-chapter 6.3 showed to what extent the current information was directly usable for the COST FWD
calibration system. For most calibration aspects, procedures appeared to be available either in detailed or in
rough simple wording. Detailed information was available for the following procedures:
• Short-term repeatability verification of load and deflections
• Long-term repeatability verification of deflections
• Relative calibration of FWD deflection sensors
• Dynamic reference calibration of FWD load cell
• Laboratory reference calibration of dismounted FWD deflection sensors
• In-situ reference calibration of dismounted FWD deflection sensors
• In-situ FWD harmonisation procedure or correlation trial
The last procedure contains the test and analysis procedures used during the FWD comparative study day.
All above-mentioned procedures were copied from their original sources with only marginal amendments.
The amendments were mainly necessary to get the procedures and protocols all written in the same style and
format. In some cases specifications were changed and adapted to the European requirements.
For the following procedures, only rough or incomplete descriptions were available. Procedures were set up
and protocols were written to have these procedures written in the same style and format as those of the
above-mentioned calibration procedures.
• Verification of FWD deflection sensor positions along the raise/lower bar
• Reference calibration of FWD oedometer
• Static reference calibration of FWD load cell
• Reference calibration of FWD temperature probe
For the verification of FWD deflection sensor positions along the raise/lower bar, a simple protocol was
written in which standard tape measures or folding rulers and cardboard or similar means are used for setting
the sensors to the required offsets. Actually, a more or less similar procedure was used for setting up the
protocol of calibration of the FWD oedometer.
A detailed description was available for dynamic reference calibration of the FWD load cell. Since static
reference calibration is not so complicated and has much in common with the dynamic version of the
procedure, drafting of the protocol was performed quite smoothly.
In the static reference calibration of the FWD load cell, the readings of the load cell under test are matched to
those of the reference load cell. For this purpose, the load is stepwise increased until the nominal range of the
load cell is reached. The procedure to be used for the FWD temperature probe is actually identical. In this
case, also readings of two instruments are stepwise compared and matched. It is obvious that the basic
structure of the protocol of static calibration of the FWD load cell was also used for the protocol for
calibration of the FWD temperature probe.
The review and assessment of the current FWD calibration procedures revealed already that reference and
relative calibration of the load-measuring unit of FWDs seemed to be covered well by the existing
125
procedures. The same applied for some protocols for the deflection sensors. Doubt still existed whether
laboratory calibrated deflection sensors would measure the 'real' deflection in the field. Mounting of the
deflection sensors, release of the FWD mass, passing traffic, weight of the FWD trailer and the nearby
influence of the wheels of the trailer all could have impact on the accuracy of the recorded deflection signal.
FWD comparative studies conducted in The Netherlands had disclosed that for some FWD equipment these
disturbing factors heavily influenced the deflection output data.
The problem of calibration described above might be resolved by in-situ reference calibration of the signal of
the deflection sensor under test to the output of a calibrated accurate reference transducer. The major problem
that arises is how to align the reference instrument with the deflection sensor under the test without
introducing measurement errors due to movement of the pavement generated by the FWD load impact.
Another problem to be resolved was guaranteeing that the reference transducer would not move prior to
recording the essential data of the test.
Review of potentially feasible and still affordable instruments led to an investigation of the feasibility of
servo-accelerometers for in-situ reference calibration of deflection sensors. It should be mentioned that this
might only been seen as a first exploratory attempt. The study was restricted to testing on laboratory scale [6-
1]. Although originally planned, the in-situ trials could not be performed any more within the COST336 time
frame.
The laboratory experiment resulted into excellent agreement between the test results from the deflection
sensor and the servo-accelerometer, both in terms of peak values (see Figure 6-1) as in deflection histories.
This result led to drafting of a calibration protocol for reference verification of FWD deflection sensors
mounted to the raise/lower bar. Since this procedure is only based on the findings of the laboratory
experiment, this protocol is labelled as optional, just to gain experience with it.
126
1,4
1,2
0,8
0,6
0,4
0,2
0,0
-0,2
-0,4
-0,6
0 100 200 300 400 500 600 700
D eflection Accelerom eter (µm )
The following sub-chapters address the structure and periodicity of the FWD calibration programme. At
the end of each calibration or verification, a brief summary of periodicity, frequency and applicable part of
the pre-draft standard is supplied. Flow charts attempt to clarify all calibration actions and their interac-
tion. The flow charts also show that in a few cases, users may choose from two procedures.
Each flow chart starts with an indication of the periodicity. The circle in each flow chart lists the flow
chart number, whereas boxes with the text ‘Part x’ refer to that part x of the pre-draft standard (see Annex
F). The contents of sub-chapter 6.5 is presented in brief format in Part 1 of the pre-draft standard.
This level of calibration describes procedures to be performed by each FWD user on a periodic basis. It
contains simple-to-use procedures that provide warrantees of proper and accurate operation of the FWD.
This level consists of the following six procedures:
• Verification of FWD deflection sensor positions (pre-draft standard Part 2)
• FWD short-term repeatability verification (pre-draft standard Part 3)
• FWD long-term repeatability verification (pre-draft standard Part 4)
• Relative calibration of FWD deflection sensors (pre-draft standard Part 5)
• Reference calibration of FWD temperature probe (pre-draft standard Part 11)
• Reference calibration of FWD oedometer (pre-draft standard Part 12)
127
All procedures will be elaborated briefly. Full details of testing and analysis of the test data are given in
the corresponding parts of the pre-draft Standard (see Annex F).
This procedure should be conducted at least once per month and each time when the deflection sensors
have to be moved or have been moved along the raise/lower bar. FWD testing on airfields one day and
FWD testing on low-volume roads the other day with different deflection sensor mountings, are typical
moments for this calibration action. The details of the procedure are described in Part 2 of the pre-draft
standard in Annex F. This procedure may be used by the FWD user, the FWD manufacturer and the
independent calibration station.
1 2 3
no no
OK ? Adjust no OK ? Adjust
OK ? 5
yes
yes
End
Part 4
End
OK ? 8
End
4 5 6
no no no Check
Check
OK ? Adjust OK ? OK ? Adjust or
Adjust
Repair
yes yes yes
128
6.5.1.2 FWD short-term repeatability verification
The objective of this procedure is to verify whether the FWD under test is capable of producing consistent
results on a specific test site. In this procedure the short-term repeatability of an FWD is verified by using
a series of multiple drops without lifting the loading plate. The deflections are all normalised to the mean
of the load imparted. The standard deviation of the load and normalised deflections should agree with the
specified limits. When the results do not meet the requirements, the test should be repeated. Cases of
persistent non-compliance invalidate data collected by the instrument under test. Use flow chart #2
presented in Figure 6-2.
This procedure should be performed at least once per month or more frequently as considered necessary
by the FWD user. The details of the procedure are described in Part 3 of the pre-draft standard in Annex
F. This procedure may be used by the FWD user, the FWD manufacturer and the independent Calibration
Station.
This procedure should be applied at least once per month or more frequently as considered necessary by
the FWD user. The details of the procedure are described in Part 4 of the pre-draft standard in Annex F.
This procedure may be used by the FWD user, the FWD manufacturer and the independent calibration
station.
The objective of the test is to verify similarity of the response of each of the deflection sensors. If one or
more sensors generate deviating results, the deflection sensor calibration factors should be fine adjusted. If
large adjustments are required, the sensors should be subject to closer investigation. No rotation of the
order of deflection sensors in the sensor stand is applied in this procedure, mainly to simplify use of the
approach for each FWD user as much as possible. Optionally, deflection sensors may be shifted to another
level in the stand in case the FWD user wants to repeat the test. Use flow chart #5 presented in Figure 6-2.
This procedure should be conducted once per period of six months or more frequently as considered
necessary by the FWD user. When the results of application of the procedure give reason to further
investigation and/or repair or servicing of components, this procedure should be repeated too after the
equipment has been returned from repair or servicing. The details of the procedure are described in Part 5
129
of the pre-draft standard in Annex F. This procedure may be used by the FWD user, the FWD manufac-
turer and the independent calibration station.
This procedure should be conducted once per year or more frequently as considered necessary by the
FWD user. When the results of application of the procedure give reason to further investigation and/or
repair or servicing of components, this procedure should be repeated too after the instrument has been
returned from repair or servicing. The procedure does not need to be performed when the FWD tempera-
ture probe carries a valid calibration certificate. The details of the procedure are described in Part 11 of
the pre-draft standard in Annex F. This procedure may be used by the FWD user, the FWD manufacturer
and the independent calibration station.
This procedure should be conducted once per period of six months or more frequently as considered
necessary by the FWD user. This calibration procedure should also be conducted after change of tyre of
the car, van or trailer to which the oedometer is mounted. When the results of application of the procedure
give reason to further investigation and/or repair or servicing of components, this procedure should be
repeated too after the instrument has been returned from repair or servicing. The details of the procedure
are described in Part 12 of the pre-draft standard in Annex F. This procedure may be used by the FWD
user, the FWD manufacturer and the independent calibration station.
130
6.5.2 Calibration station level procedures
This level of calibration describes procedures to be performed by the FWD calibration station on a
periodic basis. In some cases the FWD manufacturer may take over the calibration or verification
procedure when he is adequately equipped and experienced to do so, and preferably also certified for this
test. This level contains the following procedures:
• Static reference calibration of FWD load cell (pre-draft standard Part 9)
• Dynamic reference calibration of FWD load cell (pre-draft standard Part 8)
• Laboratory reference calibration of dismounted FWD deflection sensors (pre-draft standard Part 6)
• In-situ reference calibration of dismounted FWD deflection sensors (pre-draft standard Part 7)
• FWD correlation trial (pre-draft standard Part 10)
In the static calibration procedure, the reference load cell and the FWD load cell under test are mounted in
a test rig, making sure that both load cells are properly aligned (see Figure 6-4).
The pressure in the (hydraulic) press is slowly increased until the nominal range of the load cell under test
has been reached. Readings of the reference load cell and the FWD load cell are recorded at various
intervals spread over the whole range of testing. The FWD load cell calibration factor is computed on the
basis of the two sets of readings. Use flow chart #7 presented in Figure 6-2.
This procedure should be conducted once per two years and when the dynamic reference calibration
cannot be performed. The procedure may also be applied when a new FWD load cell is mounted in the
FWD unit. The details of the procedure are described in Part 9 of the pre-draft standard in Annex F. This
procedure may be used by the FWD manufacturer and the independent calibration station.
131
6.5.2.2 Dynamic reference calibration of FWD load cell
Dynamic reference calibration of the FWD load cell is applied to ensure that the unit measures the peak
value of the load pulse accurately. Various drop heights are applied to detect any non-linearity in the
output of the FWD load cell. The reference load cell constitutes a load platform (see Figure 6-5). This
platform is positioned beneath the FWD loading plate, making sure that the reference load cell and the
FWD loading plate are properly aligned. Then series of replicate drops from different heights are used to
match the readings of the load cell under test with those of the reference unit. Use flow chart #7 presented
in Figure 6-6.
This procedure must be conducted once per two years or more frequently as considered necessary by the
FWD user. When the results of application of the procedure give reason to further investigation and/or
repair or servicing of components, this procedure must be repeated too after the equipment has been
returned from repair or servicing. The details of the procedure are described in Part 8 of the pre-draft
standard in Annex F. This procedure may be used by the FWD manufacturer and the independent
calibration station.
132
In the second case, the calibration factors for adjusting deflection sensor output to reference instrumenta-
tion output are determined. In this process as in the last part of the first case, the readings around the
routinely used pulse duration form the basis. Use flow chart #8 presented in Figure 6-7.
7 10
either either
or or
Part 8
Part 8 short
Part 9
normal
no
OK ? 7
no Check
OK ?
Repair
yes
yes
Part 10
Part 3
no
no OK ? 7 9
OK ?
yes
yes
End
End
Figure 6-6 Flow chart calibration station and manufacturer level calibration
This procedure should be conducted at least once for any new FWD model or release of an update of a
combination of deflection sensor and signal processing unit in case the effect of varying deflection pulse
duration on the accuracy of peak value of deflection is determined. This procedure should be conducted
at least once per two years in case the deflection sensors need to be calibrated. The details of the
procedure are described in Part 6 normal version of the pre-draft standard in Annex F. This procedure
may be used by the FWD manufacturer and the independent calibration station.
133
This procedure should be conducted once per two years in case the reference deflection sensor
calibration factors must be determined. The details of the procedure are described in Part 6 short
version of the pre-draft standard in Annex F. This procedure may be used by the FWD manufacturer and
the independent calibration station.
8 9
Part 6
either normal
or
no Check
OK ?
Part 6 Repair
Part 7
short
yes
Part 7
no Check
OK ?
Repair
yes
no
OK ?
Part 5
yes
Part 5
no
OK ?
yes no
OK ?
End yes
End
134
1
This procedure is especially valuable to investigate whether release of the FWD falling mass and
vibrations generated by this release might affect accuracy of reading of the FWD deflection sensor. Since
the test rig is positioned not too far from the FWD loading plate, vibrations from the FWD load impact
will inevitably reach the test rig and have this rig vertically deflected or rotated. However, with proper
design of the test pavement structure, the test rig and well chosen stiffness and damping characteristics of
the rig, vertical movement of the reference sensor and the deflection sensor under test will be detected
later than measuring the peak value of deflection. Use flow chart #8 presented in Figure 6-7.
This procedure should be conducted once per two years or more frequently as considered necessary by
the FWD user. When the results of application of the procedure give reason to further investigation and/or
repair or servicing of components, this procedure should be repeated too after the equipment has been
returned from repair or servicing. The details of the procedure are described in Part 7 of the pre-draft
standard in Annex F. This procedure may be used by the FWD manufacturer and the independent
calibration station.
This procedure is primarily intended for use in FWD groups with membership of FWDs with limited
differences in load pulse durations. Use of widely different load pulse durations will result in deflections
for which differences in peak values may be influenced by differences in pavement structure and structural
support of subgrade. More complex conversion techniques are needed under those circumstances to
determine appropriate FWD correlation factors for all participating FWD equipment.
135
Figure 6-9 Snapshot CROW FWD correlation trial 1999
Apart from the reproducibility test, the additional objective of the procedure is to verify whether the FWD
under test is capable of producing consistent results on a specific test site. In this procedure the short-term
repeatability of an FWD is verified by using a slightly for the purpose adapted version of the test
described in sub-chapter 6.5.1.2. The procedure used is not completely identical, due to the lack of
opportunity to analyse repeatability data during the test sequence. In this case, a set of three test stations is
visited. The FWD complies with the requirements when full compliance is achieved for at least two test
stations. Also a short version of the dynamic verification of the FWD load may form part of this
calibration procedure.
The procedure was written with the following in mind. Even when all FWD load cells and deflection
sensors have been calibrated satisfactorily, reproducibility among FWDs may not be achieved. This may
be due to differences in thicknesses and properties of the rubber pads under the loading plate, properties of
the rubber buffers, etc, all leading to differences in load pulse shape and load pulse rise time. For this
reason, use of FWD harmonisation factors might enhance reproducibility among FWDs. The data of the
1999 FWD correlation trial held in The Netherlands [6-2] were used as basis for the determination of the
precision of the correlation activity. Sub-chapter 6.6 presents the results of this action.
The procedure is obligatory in The Netherlands since 1993 and is in use in the United Kingdom since
1999. The CROW correlation trial is visited by many FWDs from abroad.
Inaccuracy of positioning FWD loading plates at the exact test stations, and variation of temperature,
weather and structural support during the period of testing may affect test results for the worse. Use flow
chart #10 presented in Figure 6-6.
This procedure should be conducted once per two years. The details of the procedure are described in
Part 10 of the pre-draft standard in Annex F. This procedure may be used by the independent calibration
station only.
136
the end result. This sub-chapter presents the determination of the reliability of the pavement life
assessment based on the procedure published in [6-3]. The contribution of FWD calibration receives
special attention.
The determination of the structural life according to the approach described in [6-3] is based on the Miner
concept. The Miner number and the residual structural pavement life are mainly dependent on the:
• Traffic loading
• Critical design strain in the asphalt layer
• Fatigue characteristics of the asphalt layers
In some cases, especially at thin asphalt pavements, the vertical compressive strain at the top of the
subgrade might be more critical than the asphalt strain. Usually however, the asphalt strain is the most
important (re)design parameter. The critical asphalt strain is controlled by the:
• Homogeneity of the pavement section under investigation or more specifically by the dispersion in test
results
• Uncertainty in the process of backcalculation of layer stiffness moduli
• Uncertainty in the measurement of the pavement temperature
• Repeatability of the FWD test
• Reproducibility of the FWD test
In summary, the total inaccuracy in the calculation of the Miner number can be computed as follows:
Publication [6-3] contains indicative values for most of the inaccuracies mentioned in the equations (6.1)
and (6.2). In some cases, homogeneity of the pavement section and the accuracy of the actual test result
have some effect on the end accuracy.
137
No reliable data were available to estimate the inaccuracy of the FWD test result. The FWD correlation
trial of 1999 held in The Netherlands was used to derive indicators for repeatability and reproducibility.
Thirteen FWD of various makes and models participated at this test [6-2]. Data of only twelve units were
analysed since one the FWDs broke down during the test. Repeatability and reproducibility were
determined according to ISO 5725 [6-4].
6.6.2 Repeatability
Repeatability is defined as the closeness of agreement between mutually independent test results obtained
under repeatability conditions, i.e. that results are obtained with the same method on identical test material
(identical test station on the road) by the same operator using the same equipment within short intervals of
time. Repeatability is characterised by the lower case r, indicating the value below which the absolute
difference between two single test results obtained under repeatability conditions may expected to lie with
a probability of 95 %.
In the FWD correlation trial all FWDs visited a number of approximately 30 test stations twice. Test
stations varied in degree of structural support and type of subgrade. The subgrade ranged from stiff sand
to soft peat. At each station multiple drops were imparted without lifting the loading plate between drops.
This means that data was collected at identical test stations. Figure 6-10 presents the results of the
analysis. From the graph it is crystal clear that repeatability is very dependent on the peak value of
deflection. The trend line shows that the repeatability of a single FWD amounts 1.1 %. This implies that
when a deflection of 200 µm was measured in the first drop, the second drop will generate, with a
probability of 95 %, a deflection between 198 µm and 202 µm.
5.0
4.5
y = 0.0108x
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0 100 200 300 400 500
Deflection (µm)
138
6.6.3 Reproducibility
Reproducibility is defined as the closeness of agreement between test results obtained under reproducibil-
ity conditions, i.e. that results are obtained with the same method on identical test material (identical test
station on the road) by different FWDs with different operators. Reproducibility is characterised by the
capital R, indicating the value below which the absolute difference between two single test results
obtained under reproducibility conditions may expected to lie with a probability of 95 %.
The objective of the FWD correlation trial is to reduce differences in peak values of deflection among
FWDs as much as possible (see pre-draft Standard Part 10 in Annex F). On the basis of the test results,
deflection data of the FWDs under analyses are matched to those of a reference group. This matching
results in a correlation factor. This factor needs to be multiplied with the recorded deflections to obtain the
so-called reference deflections. Obviously no full compliance will be achieved. To determine the
reproducibility among FWDs all deflections recorded by the participating FWDs were multiplied by the
correlation factor. So ideally, all deflection bowls should be identical per test station.
The FWD correlation trial does not fully meet with the reproducibility conditions listed above. FWDs
travel form one station to another. Although the stations are marked by a white painted circle, the FWDs
are usually not capable of positioning the loading plate exactly in the centre of the circle. This means that
there is some dispersion in the position of the FWDs per test station. In other words, determination of the
reproducibility just based on the test data would not only describe the 'actual' reproducibility but also the
inaccuracy in positioning the FWD. The inaccuracy of FWD test results is composed of the following
factors:
The variables in equation (6.3) are actually other ways of expressing repeatability and reproducibility than
used in ISO 5725. Equation (6.4) shows an example of the transfer function:
Deflection data files from other repeatability and reproducibility experiments were analysed to separate
the overall reproducibility from the actual reproducibility [6-5]. These files consisted of various
repeatability tests:
• Repeatability test where the FWD remained stationary during the test without lifting the loading plate
between drops
• Repeatability test where the FWD remained stationary during the test with lifting the loading plate
between drops
• Repeatability test where the FWD drove around the block between series of multiple drops
139
These tests provided data to estimate the effect of incorrect positioning on the reproducibility. Figure 6-11
shows that again the repeatability is dependent on the peak value of deflection, i.e. 4.1 % (see trend line).
Figure 6-12 displays that the relationship between the overall reproducibility (combination of actual
reproducibility and inaccuracy due to incorrect positioning at test station) is dependent on the peak value
of deflection. The relationship contains much more scatter than in the repeatability graphs. A conservative
trend line was drawn through the data points to estimate the rate of reproducibility. Figure 6-12 shows that
the overall reproducibility equals 13.9 %.
60
Reproducibility (µm)
y = 0.1387x
50 def@300
def@600
40
def@900
30
def@1200
20 def@1500
def@1800
10
Lineair
(max)
0
0 100 200 300 400 500
Deflection (µm)
Figure 6-12 Overall reproducibility of test result inclusive incorrect positioning of FWD
140
The actual reproducibility value R can be computed by using equations (6.3) and (6.4) and the data
determined for repeatability with and without repositioning, and the overall reproducibility. The actual
reproducibility value R equals 13.3 % This implies that when a deflection of 200 µm was measured by
any FWD, any other FWD would generate, with a probability of 95 %, a deflection between 173 µm and
227 µm.
The reproducibility value for positioning the same FWD with the same operator on a specific test station
amounts 3.9 %. This implies that when a deflection of 200 µm was measured in a test, the same FWD
would generate, with a probability of 95 %, a deflection between 192 µm and 208 µm in the next round of
visiting the same test station at least within a very short time interval.
The results of the precision analysis have been entered into the appropriate places of the pre-draft
Standard Part 10 specifying the FWD correlation trial (see Annex F).
The FWD calibration station requirements can be classified in several categories. The first category
contains the obvious instrumentation. Not all instrumentation is listed in this category because this could
lead to specification of each bolt and nut.
• Reference electronic thermometer with resolution of 0.1°C and accuracy of ±0.5°C over the range from
0°C to +50°C
• Straight edge of 1.2 m for measuring cross-fall, gradient and rut depth
• 3 m long measuring tape or folding ruler with a relative accuracy of at least 0.1 percent
• 25 m long measuring tape or running wheel, both with a relative accuracy of at least 0.1 percent
• Cardboard or soft board or equivalent system to which the deflection sensors may leave holes or prints
when lowering the deflection sensor bar
• Electric drilling machine capable of drilling core holes down to a depth of 150 mm
• Digital clock indicating hours, minutes and seconds
• Spirit level
• FWD deflection sensor stand
This category contains instrumentation for recording, holding and analysing test data for dynamic and
static reference calibration of the FWD load cell.
The following instrumentation is required for the dynamic version of the load cell calibration procedure:
• Stiff test pad or test floor if test is performed indoors
• Reference load cell platform with minimum diameter of 300 mm and maximum diameter of 450 mm.
The platform must consist of either three load cells or a single custom-made wide based load cell
141
sandwiched between two plates to constitute a stable platform. The upper plate must be made of light
metal. The lower plate may be made of stainless steel.
• In case of three reference load cells, the load cells must have been matched
• The load cell platform must comply with the following specifications for static conditions:
- Error band encompassing non-linearity, repeatability and hysteresis: ±0.2 kN on the full scale range
if the maximum load of the FWD under test is less than 150 kN, otherwise ±0.4 kN on the full scale
range.
- User temperature range of 0°C to +40°C
- Adequate signal processing equipment and data acquisition electronics with peak holding feature
The following instrumentation is required for the static version of the load cell calibration procedure:
• Rigid test rig for mounting FWD load cell and reference load cell in series
• Hydraulic press or similar loading system capable to produce load beyond maximum load level of FWD
load cell
• Reference load cell. This load cell must comply with the following specifications for static conditions:
- Error band encompassing non-linearity, repeatability and hysteresis: ±0.2 kN on the full scale range
if the maximum load of the FWD under test is less than 150 kN, otherwise ±0.4 kN on the full scale
range.
- User temperature range of 0°C to +40°C
- Adequate signal processing equipment and data acquisition electronics with peak holding feature
The next set of instrumentation is required for recording, holding and analysing test data for laboratory
reference calibration of dismounted FWD deflection sensors:
• Concrete inertial block or stable solid floor, both with background vibrations of less than 1 µm. The
deflection of the floor should be on the order of 400 µm to 600 µm. Walls should not be too close to the
test floor to allow indoor manoeuvring of the FWD on and over the test floor.
• Vibration testing system (shaker + power amplifier). The system must be capable to generate single
deflection impulses of different amplitude up to 2 mm with a shape and duration range matching FWD
generated deflection impulses
• Reference displacement transducer or reference servo-accelerometer
• Programmable function generator or equivalent capable of generating single shock pulses
• Clamps to mount deflection sensors to the coil of the shaker. Light metals should be used to reduce
inertial effects as much as possible
Basic sketches of the dimensions of the inertial block are presented in pre-draft Standard - Part 4. Figure
6-13 displays the concrete inertial block operated by Carl Bro Pavement Consultants.
The next category contains instrumentation for recording, holding and analysing test data for in-situ
reference calibration of (dis)mounted FWD deflection sensors:
• Test rig for mounting deflection sensor under test, reference instrumentation
• Reference displacement transducer or reference servo-accelerometer
• Clamps to mount deflection sensors to the test rig
142
Figure 6-13 Concrete inertial block for reference deflection sensor calibration
For the calibration of temperature probes, the following equipment and instrumentation is needed:
• Bath with contents between 5 dm3 to 10 dm3 filled for more than 50 percent with glycol or other fluid
• Stirring device
• Protective cover at container with two holes allowing the temperature recording tips of the temperature
probe under test and reference thermometer to be lowered in the fluid. Cork stubbles or equivalent
material should be used to reduce heat loss through the holes
• Heater/cooler system with sufficient capacity for increasing the temperature of the fluid
For some calibration procedures it is necessary for the calibration station to have available a test pad or
test floor that is more or less supported independently from the rest of the floor and the surrounding. The
test pad or test floor should comply to the following specifications:
• Slab size approximately 4 m x 5 m
• Concrete floor resting on gravel or crushed stone base course resting on a soft layer with stiffness of
around 50 MPa resting on the local subgrade. The aggregate system and subgrade should generate a
deflection level between 400 µm and 600 µm using a FWD peak load of 50 kN. This might be feasible
by using the following dimensions:
- 125 mm fibre reinforced concrete
- 150 mm gravel or crushed stone
- 500 mm sand or sand/clay mix with dynamic stiffness modulus of 50 MPa
• 1 m free space at long sides of test pad
• Free entry to FWD and van at one of the short sides
• 2.50 m free space at opposite short side (see Figure 6-14)
• No walls within 3 m from the slab
143
1.00 m
5.00 m
1.00 m
Concrete block
1.00 m x 1.50 m 4.00 m
Test pad
For a detailed description and photo of the concrete block, see Figures 6-8 and 6-13.
6.8 References
[6-1] Sørensen, A. and Van Gurp, C., Determination of feasibility of servo-accelerometers for in-situ
deflection sensor calibration verification. EC contract R98/98 SIN 001686-B6721104, Report
e98386. Dynatest and KOAC•WMD, Apeldoorn, 1999.
[6-2] 1999 Relative calibration of European Falling Weight Deflectometers. CROW Research Report 00-
03. CROW, Ede, The Netherlands, March 2000.
[6-3] Deflection profile - not a pitfall anymore. Record 17. CROW, Ede, The Netherlands, May 1998.
[6-4] Precision of test methods - Determination of repeatability and reproducibility for a standard test
method by interlaboratory tests. International Standard ISO 5725 - 2nd Edition. September 1986.
[6-5] Reproducibility of a single Falling Weight Deflectometer. Report E96333. Wegmeetdienst.
Apeldoorn. January 1997.
144
CHAPTER 7
BENEFITS TO DIFFERENT USERS
7.1 General
COST Action 336 provided four main deliverables. These are:
• Guidelines for the use of FWD at project level
• A proposal for applicable procedures for the use of FWDs at network level
• Common requirements for calibration of FWD devices and components
• Requirements for FWD calibration stations
These deliverables will have an ample number of benefits, and for most of these, several benefiting parties
can be mentioned.
7.2 Benefits
The most important benefits that are anticipated on the basis of the direct results of COST Action 336 are:
• Better reproducibility of pavement evaluation results, both within each country and amongst
countries, with less dependency upon consultant and equipment
• More optimal project and network level maintenance decisions due to sharing of knowledge during
COST Action 336, leading to reduced maintenance expenditures, less traffic hindrance and lower
environmental impacts
• Possibility for consultants and FWD devices to operate throughout Europe
• Readily available work descriptions for road authorities when contracting pavement consultants for
pavement testing, monitoring and evaluation
• More accurate test data due to better calibration procedures
• Avoiding duplication in research and development efforts enabling common use of results
• Enabling mutual exchange of data to be used for validation and fine-tuning of pavement deterioration
models and evaluation procedures
• Uniformity in FWD hardware requirements
• Preparatory basis for common standards
• Preparatory basis for the development of performance pavement specifications
• Possible contribution to common legislation as to traffic load limitations as a function of structural
capacity
7.3 Users
The deliverables of COST Action 336 are applicable to several parties. Some of the following parties can
be seen as direct users of the deliverables:
• Road managing authorities usually employing staff pavement engineers but often also commissioning
private consultants for project or network analysis of pavements or pavement networks
• Road design and maintenance engineers, either private consultants or road authority employees, both
involved in the project level analysis of FWD measurements and the associated bearing capacity
assessment and maintenance design actions
145
• Pavement Management System engineers, either private consultants or road authority employees
involved in network level decisions on required national maintenance budgets, budget allocation over
individual regions, and selection and prioritisation of pavement sections requiring project level analysis
• Industry involved in manufacturing FWD equipment and/or data processing systems
• Educational and research institutes involved with training of pavement engineers and/or active in the
process of research in the structural behaviour of pavements and with developing improved procedures
for evaluation of pavement behaviour
• Decision makers on national and European level requesting information on the bearing capacity on
network level as a basis for financial decisions but also for decisions on the subject of e.g. traffic load
legislation
Some other parties will not use the deliverables of COST Action 336 directly, but will benefit in one way
or another by the Action, and can therefore be regarded as indirect users:
• European citizens, most of which paying taxes for pavement maintenance, education and research, but
usually also active as road users, meaning that they will experience pavement safety, riding comfort, and
traffic hindrance due to maintenance
• Environment affected by several effects of pavement condition and pavement maintenance, such as
noise, energy consumption, use of materials, production of waste materials etc.
Table 7-1 presents an overview of the principal benefits for the various categories of users.
146
Table 7-1 Main benefits for different users
Education organisations
Decision makers on
Research institutes
European citizens
nance engineers
PMS engineers
Environment
Industry
Better reproducibility of pavement
evaluation results 3 3 3
More optimal project and network
level maintenance decisions 3 3 3 3
Possibility for consultants and FWDs
to operate throughout Europe 3 3
Readily available work descriptions 3 3 3
More accurate test data 3 3 3
Avoiding duplication of research and
development efforts 3 3 3 3
Enabling mutual exchange of data 3 3 3
Uniformity in FWD hardware
requirements 3 3
Preparatory basis for common
standards 3 3 3 3
Preparatory basis for the development
of performance specifications 3 3 3 3
Possible contribution to common
legislation on traffic load limitation 3
147
CHAPTER 8
CONCLUSIONS AND RECOMMENDATIONS
This chapter summarises briefly the outcome of the work of the different task groups of COST Action 336
'Falling Weight Deflectometer'. Furthermore, some general comments relating to the outcome of the
Action as a whole and aspects for future research are pointed out.
8.1 Conclusions
• The FWD is widely used throughout Europe for the evaluation of structural conditions of road
pavements at project level with a wide range of different approaches
• Guidelines for the whole set of FWD measurements and analysis procedures are provided
• A comprehensive set of FWD calibration procedures has been developed for increasing accuracy,
repeatability, reproducibility and exchangeability of deflection data
• Approaches are provided to facilitate periodic check-up of a variety of calibration aspects at FWD User
level
• Guidelines are provided for the installation of an FWD calibration station
• The calibration protocols drafted form excellent raw material for the set up for European standards for
FWD calibration
• Application of the calibration procedures will improve the reliability and consistency of FWD results,
which should enable more meaningful exchange of research results amongst countries
• Guidelines for the use of FWD at network level support candidate users to choose between three
different levels of details for each of the fifteen selected, most important FWD parameters covering
structural aspects of the actual network and fitting PMS requirements
• The guidelines for the use of FWD at network level are an attempt to guide the users in providing them
with sufficient information to make their own decisions and improved assessments
• COST Action 336 resulted in a better network for exchanging information and it brought people and
information together
• The FWD Users mailing list and the list of FWD owners help to identify organisations, companies and
institutes that can be contacted for additional information
• COST Action 336 resulted in an upgrade of knowledge in countries where FWDs are not used at present
• The work performed in COST Action 336 is a major advance in the reduction of duplication of research
effort in the field of using FWDs for the evaluation of the structural condition of road pavements
8.2 Recommendations
• Guidelines need to be refined and developed into a form of national or international standard or official
guide to fully realise the benefits of the FWD measurement and analysis procedure
• The results of various studies (COST Actions) show that deflection measurements in one way or another
should form an integral part of a Pavement Maintenance Management System
• Research effort should be increased to study the effect of FWD load pulse duration on deflection
readings, especially in situations with soft subgrade
• Analysis procedures should be developed that convert FWD deflection output into standard structural
condition data that are independent of the FWD capturing the raw data
• More emphasis should be laid in the development of project level processing tools in which the whole
deflection history is used as input parameter. Current practice satisfies with the peak readings only.
148
• More effort should be put into specification of the required accuracy of deflections measured by a high-
speed deflection measurement system
• Studies should be initiated to investigate the influence of better pavement and material characterisation
by using stress-dependent material models (i.e. finite elements method)
• A study should be performed to examine the influence of fluctuations in the saturation level of the soil
and ground water table on the pavement behaviour
• The activities of COST Action 336 were for several reasons mainly focused on flexible pavements. It is
recommended to start another COST Action on the (harmonisation of) design and evaluation of rigid
pavements for roads and airfields
• The calculation of residual life and the determination of overlay thickness have to be handled carefully.
Without having the background basis of fatigue behaviour of asphalt layers in-situ and without having a
transfer function 'laboratory behaviour – in-situ behaviour', the determination of residual life from FWD
data is not convenient.
149
European Cooperation in the Field of Scientific and Technical Research
COST 336
Use of Falling Weight Deflectometers in Pavement Evaluation
Final Report of the Action
Annexes
2nd Edition – April 2005
European Commission
Directorate General Transport
ANNEX A
TECHNICAL ANNEX TO THE MEMORANDUM OF UNDERSTANDING
TECHNICAL SUB-COMMITTEE
COST 336
COST/308/95
Annex A-1
At its 59th meeting, on 3rd May 1995, The Technical Committee on Transport set up a sub-committee
responsible for defining the scientific content of COST 336.
Chairman Technical Sub-committee: Mr. Jan Jansen (Danish Road Institute, Denmark)
Secretary Technical Sub-committee: Mr. Andrew Stimpson (European Commission)
Annex A-2
GENERAL DESCRIPTION OF THE PROJECT
A Background
This Action deals with a device used for measuring the bearing capacity of a road pavement, and also with
the post processing procedures required for actual assessment of the strengthening overlay need evaluated
at road project level. This type of device is called a Falling Weight Deflectometer, FWD.
The FWD is already used in many EU countries and, in the US/SHRP programme, it is chosen as the
authorised equipment for bearing capacity assessment of pavements. Use of FWD testing is a growing
requirement in modern pavement maintenance management, and it has already been implemented as a
routine procedure in some European countries, and in others as equipment used in research or special
investigations. Three types of apparatus are most commonly used (two Danish and one Swedish), and a
few non-production machines also exist. This situation, which gives a variety in testing results and
procedures, must be harmonised to make exchange of results and experience possible.
Different procedures and usage have been investigated by a FEHRL FWD group for three years with the
aim of creating guidelines for making measurements with FWDs. This work has resulted in a common
FEHRL-FWD publication. Participants in COST 336 maintain close links with work being undertaken
within associated international projects, such as those operated within the OECD and US/SHRP
programmes. Interchange between such activities and COST 337 will be on a selective basis, as with each
of the Actions initiated by FEHRL.
The COST framework has been chosen as the most appropriate coordination funding mechanism in this
area for the following reasons:
• It is desirable that as many of the nominated COST countries benefit from the work as possible; not only
those within the EU at the present time.
• In order to create an effective common code of good practice, it is desirable to have agreement between
the technical representatives of national governments.
The main objective of the Action is the development of a European common code of good practice for use
of Falling Weight Deflectometers in pavement evaluation. This will encompass the following activities:
• Expand FEHRL harmonisation proposal to incorporate strengthening evaluation on the basis of FWD-
tests
• Establish common requirements for calibration of measurements and machines
Annex A-3
• Describe the potential for use of FWDs in evaluation at network level
• Establish a preparatory basis for possible European standardisation in the field
The situation today is such that usage and experience in pavement evaluation, by FWD means, is very
different among the European countries which utilise this equipment, and needs in this area should be
addressed. Further to this, there is a need for the identification of common requirements for calibration
and correlation of the various machines on the market.
The Action will also contribute to the achievement of a number of wider objectives. Examples are as
follows:
• Lack of dependence on individual FWD types
• A more harmonised market for organisations involved in bearing capacity testing by means of FWD
• Extended knowledge of FWD testing on roads with flexible and rigid pavements
B.3 Benefits
The FWD has been developed in Europe, and is now in worldwide use. Much of the expertise and
knowledge on FWD operation is available in European countries, and taking advantage of this position
will address certain shortcomings in the US/SHRP approach, which have already been identified as a
result of European experience. Previously, many countries had carried out parallel research in the field,
and formulated national standards and procedures will effectively be taken into account within the COST
Action.
The Action will allow less experienced countries (including CEE and NIS) to enhance their capabilities in
this area, and will promote the exchange of services within the field of pavement strengthening evaluation
in the EU single market. Thus, a formulated Common Code of Good Practice will be a prerequisite.
Finally, the Action will contribute to the pooling of the abilities of several COST countries with
experience in the proposed research field.
C Scientific programme
The Action is a continuation of the successful FEHRL FWD Activity group which drafted the FEHRL-
publication No 1: 'Harmonisation of FWD Measurements and Data Processing for Flexible Road
Pavement Evaluation at Project Level'. This was the result of three FEHRL FWD Seminars held in The
Netherlands, France and Denmark with contributions from 14 European countries.
The Action comprises four tasks:
This task will give a complementary description of the post processing of FWD data at project level to the
following items which had not been covered until now:
• Calculation of in-situ layer stiffness
• Correction of layer stiffness to standard conditions
• Calculation of critical stresses and strains
• Estimation of residual structural lives and the required thickness of strengthening overlays
Annex A-4
The sub-tasks necessary to achieve this are as follows:
• Collection of existing information
• Determination of elements to be documented
• Evaluation of the existing post processing models
• Specification of a post processing model
• Definition of the term 'residual life'
• Calculation of in-situ E modulus
• Normalisation of pavement parameters
• Assessment of critical stresses and strains
• Calculation of residual structural life
• Evaluation of the required thickness of strengthening overlays
• Generation of guidelines for post processing procedures
Having completed this task, the benefit will be a harmonised description of how FWD measurements are
post processed, enabling effective exchange of FWD consultancy services within the Single Market. It
will also give participating countries a better method for pavement condition evaluation and maintenance
of roads. Hence, maintenance costs will become more effective when appropriate decisions can be made
and this will lead to saving of much money for the EU countries.
This task will extend the work on harmonisation of FWD measurement already carried out by the FEHRL
FWD Activity group for executing measurements at project level. It will involve the description of how
and when FWDs should be used at network level, and will provide estimates for the timing of mainte-
nance and strengthening requirements to road agencies.
The sub-tasks necessary to achieve this are as follows:
• Collection and assessment of existing information
• Drafting of guidelines for FWD use at network level
• Organisation of a seminar on use of FWDs at network level
• Dissemination of information to seminar participants
• Generation of a task report
Having completed this task, the benefit will be the specification of circumstances under which FWD
measurement can be used effectively for network level evaluation.
This task will focus on quality assurance of FWD measurements, from a calibration point of view, in three
areas:
• Calibration of the complete FWD measurement system
• Description of an instrumented calibration station for calibration checks on real pavements
• Normalisation of data from different FWD types
This work will be carried out under the following task headings:
• Creation of an inventory of current calibration procedures
• Assessment of existing calibration approaches
• Investigation of the effect of load pulse shape and width on deflection peak value
• Investigation of FWDs with different load pulse widths
• Evaluation of load pulse and peak value investigations
Annex A-5
• Assessment of procedures for normalising FWD values
• Assessment of the calibration approach and activities
• Description of requirements for an FWD calibration station
• Verification and validation of the prescribed procedures
• Preparation of the calibration protocol
• Documentation and reporting of results
Having completed this task, the deliverable will be a set of guidelines for achieving the level of precision
required for FWD testing.
This task is the concluding and production stage of the proposed work programme. It will provide the
following deliverables:
• Guidelines for the post-processing of FWD data at action level
• Proposal for applicable procedures for use of FWDs at network level
• Calibration protocols for FWD machines
• Requirements for FWD calibration stations
D.1 Organisation
The work will be organised and supervised by the Management Committee, which will probably meet, on
average, three times per annum. The Management Committee will report to the Technical Committee on
Transport. Working groups will be set up and assigned responsibility for completing tasks 1, 2 and 3,
together with their share of task 4. Each working group will typically consist of 5-7 people. This
organisation structure is shown in Figure A-1.
Management
Commitee
It is anticipated that a small number of short-term scientific missions would be desirable to support the
work in tasks 1 and 3. Also, a seminar in the latter stages of year 1 would be a highly desirable part of the
work of the task 2 working group. Budget availability at the time will be taken into account when
formulating these, and any other, requests for financial support.
Annex A-6
D.2 Timetable
The duration of the COST Action 336 will be three years and the timetable for the tasks and sub-tasks is
shown in the following tables.
Annex A-7
E Economic dimension
The following COST countries have actively participated in the preparation of the Action:
Austria Belgium
Denmark Finland
France Ireland
Netherlands Portugal
Slovenia Spain
Switzerland United Kingdom
It can therefore be seen that this is a very strongly supported Action in a much specialised technical area.
On the basis of national estimates provided by the representatives of these countries, and taking into
account the coordination costs of the European Commission, the overall cost of the activities to be carried
out under the Action has been estimated, at 1995 prices, at roughly € 1,9 million. There are likely to be
15-20 man-years of effort involved. This estimate is valid under the assumption that all the countries
mentioned above, but no other countries, will participate in the Action. Any departure from this scenario
will change the total cost accordingly.
Annex A-8
ANNEX B
COMMITTEE MEMBERSHIP
Annex B-1
B.2 Meetings
The Technical Sub-Committee met as follows:
1st meeting 13 September 1995 Brussels
2nd meeting 5 December 1995 Vienna
3rd meeting 29 March 1996 Lisbon
4th meeting 28 June 1996 Brussels
The first meeting of the COST 336 Management Committee took place in Lausanne, on 7/8 November
1996. The internal rules of procedure were approved and the chairman and vice-chairman were elected as
follows:
Chairman Mr. Egbert Beuving (CROW, Netherlands)
Vice-chairmen Mr. Ole Fog, (DRI, Denmark)
Mr. Brian Ferne (TRL, United Kingdom)
Annex B-2
Gudrun Golkowski BASt GERMANY
Andreas Loizos University of Athens GREECE
Nikolaos Michas Greek Road Federation GREECE
Maria Sakki Greek Road Federation GREECE
László Gáspár KIR HUNGARY
Haraldur Sigursteinsson Public Road Administration ICELAND
Frank Clancy NRA IRELAND
Arthur van Dommelen DWW NETHERLANDS
Christ van Gurp KOAC•WMD NETHERLANDS
Helge Mork NTNU NORWAY
João Rocha de Almeida UNL-FCT PORTUGAL
Maria de Lurdes Antunes LNEC PORTUGAL
Stefan Hǎrǎtǎu Iptana Search ROMANIA
Aleš Hoçevar Druzba za Drzavne SLOVENIA
Guillermo Albrecht Geocisa SPAIN
Jorge Serrano CEDEX SPAIN
Olle Tholén KUAB SWEDEN
Leif Wiman VTI SWEDEN
Gerald Cuennet SWITZERLAND
Christophe Rohr Viagroup S.A. SWITZERLAND
Jean-Daniel Zufferey Viagroup S.A. SWITZERLAND
The COST Action 336 Management Committee wishes to acknowledge the persons mentioned below for
positive contribution to the Action and their occasional attendance of the Committee Meetings:
Jan M. Jansen DRI DENMARK
Heinrich Werner BASt GERMANY
Tom Scarpas Delft University of Technology NETHERLANDS
H. Haraldsson Public Road Administration ICELAND
Tom Jermyn NRA IRELAND
Enrico Salvatori SIPROMA sr.l ITALY
Fransicsco Sinis CEDEX SPAIN
Manuel Romana Universidad Politechnica de Madrid SPAIN
Annex B-3
Dr. Helge Mork, NTNU (N)
Annex B-4
ANNEX C
CURRENT PRACTICE OF POST-PROCESSING
OF FWD DATA IN EUROPE
Results of a Questionnaire
30 April 1999
Annex C-1
Preface
In 1996 COST Action 336 'Falling Weight Deflectometer' officially started and was a continuation of the
FEHRL-FWD group. The goal of this COST-336 Action is to develop a European common code of good
practice for the use of Falling Weight Deflectometers in pavement evaluation.
This involves:
• Development of a harmonisation proposal for the evaluation of flexible pavements at project level using
FWD-tests (This is an update and an expansion of the existing FEHRL document)
• Assessing of the potential for using of FWDs in evaluation at network level
• Establishing of common requirements for calibration of measurements and machines
• Establishing of a preparatory basis for possible European standardisation in the field of the use of Falling
Weight Deflectometers in pavement evaluation
One of the objectives of Task Group 1 was to review current practice in Europe. The results of this work
are presented in this annex.
Annex C-2
C.1 Introduction
Task Group 1 'Post-Processing of FWD Data' of COST 336 organised a questionnaire covering the
following items:
• Calculation of in-situ layer stiffness
• Normalisation of layer stiffness to standard conditions
• Calculation of critical stresses and strains
• Estimation of residual structural lives and required thickness of strengthening overlays
The questionnaire is intended to provide background information to aid the group in its main task of
producing a successor to the FEHRL Report No 1996/1 on the Harmonisation of the use of the Falling
Weight Deflectometer on Pavements Part 1. This report will cover the above listed subjects in detail as
well as providing an update to the measurement and data processing subjects covered in the FEHRL
Report.
A wide range of computer programs is used for backcalculation of stiffness moduli. Most common (24x)
is an elastic multi-layer program to model the response of the pavement under surface loading. A
computer program based on the finite element method is used by six respondents. A stress dependent
stiffness modulus for the subgrade and sometimes for a granular road-base is used by 14 respondents, the
method of equivalent thickness by 16 and 13 use a fixed-bottom or stiff-layer at depth approach to the
modelling. Just one respondent uses a visco-elastic program and another uses a method with the capability
of modelling unequal horizontal and vertical stiffnesses.
Besides several backcalculation computer programs developed in-house, the following programs were also
mentioned more than once: ELMOD (7x), PHOENIX (4x), MODULUS (2x), MISS (2x). In total 22
different programs were mentioned.
Annex C-3
If known, the following names of the multi-layer computer program used for forward analysis within the
backcalculation program were mentioned more than once: BISAR (5x), WES5 (4x), CHEVRON (3x), and
CIRCLY (2x). In total 10 different programs were mentioned.
In five cases the maximum number of pavement layers used for modelling the pavement depended on the
number of FWD-sensors used. The criteria used were either that the number of layers should be less than
or equal to the number of sensors or the number of sensors minus one. The maximum number of layers
used in modelling the pavement is most often 4 (14 out of 26 responses). Others use 2 layers (1x), 3 layers
(2x), 5 layers (5x) and four with six or more layers or no limit.
A stiffness ratio for the stiffness of some layers (e.g. the stiffness of a granular base-course equals three
times the stiffness of the subgrade) is used by ten respondents and sometimes by two others. Two
respondents specifically mentioned that this method was used to define the seed moduli. Four used the
Shell relationship between sub-base and subgrade stiffness, two used a simple factor and the remainder
defined the method as 'user selected'.
26 out of 30 respondents always use a perfect bond between the pavement layers.
The stiffness modulus is backcalculated by 21 respondents on all available deflection bowls (i.e. for each
test point measured on a road section). Only a few use the mean (theoretical) bowl (4x), somewhat more
(14x) use the nearest match to the mean bowl for a road section. Eight respondents use a representative
deflection bowl (of a road section). In the latter case an 85-percentile bowl is used most often (5x), one
respondent uses the 90- and 95-percentile and one claims to use the average minus one times the standard
deviation, which is similar to the 85-percentile.
Seven respondents use the mean layer thickness of the road section for backcalculation; 12 use the
measured layer thickness at the point to be backcalculated and nine use either method.
Annex C-4
C.3.4 Accuracy and tolerance of calculated deflections
One of the two following formulae is used by 25 out of 27 respondents to calculate the accuracy of the
back-calculated deflection bowl:
1 n δ c ,i − δ m , i
n
∑
i =1 δ m,1
100% ≤ 2%
(C.1)
or
2
1 n ⎛ δ c,i − δ m ,i ⎞
n
∑ ⎜
⎜ δ
⎟ .100% ≤ 2%
⎟
i =1 ⎝ m ,1 ⎠
(C.2)
A tolerance of 2% is most often used (12x). In other replies a tolerance between 2% to 5% is mentioned
(5x), 5% - 10% (2x) and in 4 cases the tolerance depends on the user input.
Eight of the respondents use weighting factors in the formula depending on the distance between the
sensor and the load centre (when the fit of one sensor is regarded as more important than another one). A
few respondents use absolute rather than proportional or percentage differences to judge the accuracy of
fit.
The horizontal strain at the bottom of the (bituminous bound) layer due to a standard load is used as the
critical strain in pavement evaluation by 27 of the 31 respondents to this question. The maximum of the
longitudinal or the transverse strain due to a standard axle load is used by the majority (20x). Five use the
longitudinal strain and three the transversal strain.
Annex C-5
The vertical stress/strain at the top of the (granular) road base is used by 14 respondents as a critical
stress/strain parameter in pavement evaluation. The horizontal stress/strain at the bottom of the road base
however is used by only one respondent for the evaluation of granular road bases. For bituminous or
cement-bound road bases 25 and 23 respondents respectively use this horizontal stress/strain for
evaluation purposes. For the stress/strain at the bottom of the road base the maximum of the longitudinal
and transverse stress/strain (due to the standard axle load) is used by 21 respondents; four use the
longitudinal and two use the transversal stress/strain.
The vertical stress/strain at the top of the sub-base is also used by 16 respondents as a critical stress/strain
parameter, in the case of a subgrade, this is used by 25 as a critical stress/strain parameter in pavement
evaluation.
A majority (25x) of the respondents obtain layer thicknesses by coring and a lesser number (16x) by trial
pits. Ground penetrating radar (GPR) is always used by 7 respondents and sometimes by 4. All except one
of these always use GPR in combination with cores.
Just over half the respondents use deflection bowls as an indication of where cores should be taken to
obtain the layer thicknesses. Six answered that the location chosen for coring is the nearest match to the
mean deflection bowl and 14 that this location is representative of the measured relative bearing capacity
level of the subsection.
Cores are most often drilled in the nearside wheel track (26x) or in nine cases always and 21 sometimes
one core is taken from the nearside wheelpath and one between the wheelpaths.
The recording of distress patterns in conjunction with FWD surveys is fairly common. The most
commonly recorded types are structural damage/alligator cracking (fatigue) recorded by 26, longitudinal
cracking by 25, sealed and unsealed cracks by 23 or patching applied to obviate structural distress by 24.
A special form to register these types of distress is used by over half (17x) of the respondents.
The moisture/groundwater table level in the road embankment is determined by three routinely and 14
sometimes. The drainage situation itself is considered routinely by 17 respondents and sometimes by nine.
This assessment is generally on a visual basis and is recorded in the measurement file but one respondent
commented that sometimes a check on ground water level would be made.
Annex C-6
C.7 Determination of material properties
In this case the bituminous bound layer is meant. The material properties of asphalt samples are
determined routinely by 12 and sometimes by 8 respondents. The investigation is most often based on
cores (18x). The core diameter is either 100 mm or 150 mm. In a few cases a different diameter is used
(60 mm) or a diameter of 200 mm (Germany) if the stiffness modulus is to be determined. This last
determination is carried out routinely by 9 and sometimes by 6. Routinely 23 perform a visual inspection
of cores, determination of type of layers and thicknesses and 21 look at any crack propagation in the cores.
Only three respondents routinely determine fatigue/life relationships from core material tests. Most, 19,
use a standard fatigue/life relationship for the asphalt material. Ten always and two sometimes use the
composition of the asphalt mix to determine the fatigue curve.
Cement bound, unbound and self-cementing road base (course) layers) are defined to be part of the road
base for the purpose of this questionnaire. Cores or trial pits are used to take samples from bound road
base layers for the determination of the type of road base material by 16 respondents routinely and five
sometimes. The material properties, especially grain size distribution, water content and sometimes
stiffness modulus, cube strength, compatibility etc. are determined by 12 and sometimes by four. A
standard fatigue/life relationship is used by 12 respondents and sometimes by two.
C.7.3 Sub-base
Cores are drilled or trial pits are routinely excavated for the determination of the type of sub-base material
by half the respondents (14x) routinely and three sometimes. The material properties are determined by
nine routinely and five sometimes. The main properties determined are the particle size distribution and
the water content.
C.7.4 Subgrade
The material properties of the subgrade are examined routinely by 15 and by four sometimes. As for the
sub-base the main properties determined are the particle size distribution and the water content. A standard
fatigue curve is used by 19 respondents.
Twenty-one routinely and 7 sometimes look at the type of environment adjacent to the road (shoulders,
trees, ditches etc.) when assessing the results.
Annex C-7
C.9 Traffic
Traffic loads are usually expressed in terms of equivalent standard axle loads (ESAL) (29 out of 31). For
calculating the number of ESAL’s, the 4th power law is used by the vast majority of the respondents but
one uses a 5th power law.
In the determination of cumulative standard axle loads the vast majority of the respondents (26x) base this
on traffic counts of total number of commercial vehicles (23x) or vehicle classification counts (3x). The
conversion of counts to axle loads is either on the basis of average vehicle equivalence factors (23x) or
determined by weigh-in-motion measurements (14x) and in some cases using both methods (11x).
Nineteen of the respondents take some account of the reliability of the traffic data sources in their
calculations.
The number of equivalent axle loads for the slow lane from the time the road was constructed until the
survey date is determined by 19. If there are more lanes in one direction, the values for these are also
calculated by 15 per direction. Thirty respondents determine the traffic growth rate or, when available, the
commercial vehicle growth rate. A correction factor for the width of the lane, due to lateral wander, is
used by 13 and by 3 sometimes and a correction factor for multi-lane carriageways is used by 22.
When modelling a standard axle load, most respondents use a load of 80 kN or 100 kN but three use
130 kN and one uses 115 kN. A 'dynamic (impact) factor' is used by ten respondents and of these ten,
eight use a value of 1.2. Seven respondents also use a 'Wide base single tyre adjustment factor' of up to
1.3 ESAL’s (6x) or 2-4 ESAL’s (3x).
Annex C-8
distribution of backcalculated fatigue life. Other methods mentioned to determine the structural residual
pavement life are based on Miner's Law, logN-stdev(logN) or defined as the remaining life under
predicted traffic. A safety margin in calculating the residual pavement life is used by four respondents,
two based on Miner's Law.
A calculation of an overlay thickness for every FWD-testing point is carried out by 17 respondents. A 75-
percentile of the normal distribution of the calculated overlay requirement is used by 7 as the required
overlay thickness.
The evaluation method used is a described/standardised one in Finland, France ('Erasmus') and the
Netherlands (CROW Record 17). The 'Design Manual for Roads and Bridges Vol. 7' in the United
Kingdom includes some guidance on the use and interpretation of FWD data, but is not a standardised
evaluation method.
Annex C-9
C.17 Acknowledgements
The assistance of all those who participated in this exercise particularly those who completed the
questionnaire is gratefully acknowledged. A list of the respondents' names and addresses are given in
Appendix C1. In particular the preliminary analysis of the results by Egbert Beuving and Hans Bakker is
gratefully acknowledged.
Annex C-10
Appendix C1
List of Respondents
Annex C-11
Name Position Organization Street/ number P. O. Box Zip code City
Helmut Nievelt General Manager Nievelt Labor GmbH Wiener Strasse 35 A 2000 Stockerau
Dr. Michael Fuchs ÖFPZ Arsenal GmbH Faradayg 3 A 1031 Wien
Niels Monson Vejle County Vejle
Ove Noer Ribe County Ribe
Per Ullidtz Associate Professor Technical University Building 115 DK 2800 Lyngby
of Denmark
Ole Fog Pavement Engineer Danish Road Institute Elisagaardsvej 7 235 4000 Roskilde
Jens. P. Pedersen Project engineer Phønix Pavement Fuglesangs alle 16 DK 6600 Vejen
Consultants
Anders Sørensen Director Dynatest Sydvestvej 136 2600 Glostrup
R&D
David Gershkoff Pavement Consultant TRL Old Wokingham RG45 6AU Crowthorne
Researcher Road
Steven B. Finnie Civil Engineer May Associates Lion Buildings ST14 8HZ Uttoxeter
"Market Place"
Dr. Francis Chan Principle Engineer SWK Pavement 9 Faraday Buliding N67 ZQP Nottingham
Engineering Highfield Science
Lauri Liimatta University Of Oulu Kasarmintie 4 191 FIN-90101 Oulu
Matti Ruuti Civil Engineer Finnra 33 FIN-00521 Helsinki
Antti Ruotoistenmäki Research Scientist VTT Lämpömiehenkuja 2 19031 FIN-02044 Espoo
VTT
Ph. Lepert Chef de Section LCPC Centre de Nantes BP 19 F 44340 Bouguenais
Becker / Wolf BASt Brüdestr. 53 51427 Bergisch
Gladbach
Dr. Andreas Loizos Ass. Professor National Technical 5, IroonPolytechniou Gr 15773 Athens
NTUA
University of Athens Str.-Zografou
Dr. Làszlò Gàspàr Head of Department KTI Rt (Institute for Temesvar v. 11-13 H-W6 Budapest
Transport Sciences)
Haraldur Public Roads Borgartùn 5-7 IS-105 Reykjavik
Sigursteinsson Administration
Frank Clancy Civil Engineer NRA of Ireland Pottery Road; Dun Dublin
Laoghaire
João Rocha de Associate Professor UNL/FCT Quinta da Torre P-2825 Monte da
Almeida Caparica
Maria de Lurdes Research Officer Laboratório Nacional Avenida do Brasil 101 P-1799 Lisbon
Antunes de Engenharia Civil
Stefan Hărătău Vice President Iptana-Search Caderea Bastiliei 71139 Bucharest
technical Nr 65
Francesco Sinis Pavement Centro de estudios y Autovia de Colmenar 28790 El Goloso
rehabilitation experimentacion de Obras km. 18,2 (MADRID)
division Publicas (CEDEX)
Annex C-12
Name Country Phone No. Fax No. E-mail Research Public Consultant
Road
Annex C-13
Appendix C2
Questionnaire
Egbert Beuving
CROW
P O Box 37
6710 BA Ede
The Netherlands
Annex C-14
Introduction
The information collected in this questionnaire will assist COST Action 336 'Use of Falling Weight
Deflectometers in Pavement Evaluation'. COST is the acronym for European Co-operation in the field of
Scientific and Technical Research, and is a framework for scientific and technical co-operation, allowing
the co-ordination of national research on a European level. COST Actions consist of basic and precom-
petitive research as well as activities of public utility.
COST Co-operation was set up in 1971. The organisation is based upon a flexible set of arrangements
enabling different national organisations, institutes, universities and industry to join forces and make
concerted efforts in a broad range of scientific and technical areas (e.g. Transport).
This COST Action 336 is a concerted European research action, which has the objective of developing a
European common code of good practice for the use of Falling Weight Deflectometers in pavement
evaluation.
This questionnaire of Task (Group) 1 of COST 336 is intended to expand the FEHRL harmonisation
proposal to incorporate strengthening evaluation on the basis of FWD-tests. This task group will give a
complementary description of the post-processing of FWD-data at project level and will cover the
following items, which have not yet been tackled in detail by the first FEHRL document:
• Calculation of in-situ layer stiffness
• Normalisation of layer stiffness to standard conditions
• Calculation of the critical stresses and strains
• Estimation of residual structural lives and required thickness of strengthening overlays
General
1. Most questions can be answered with yes or no. Strike out no when your answer is yes and strike out
yes when your answer is no.
3. When you have to fill in an answer, please write it on (and after) the: .........
4. When you want to give additional information, please use a blank sheet and mention first the number
of the question.
Annex C-15
Terminology
The following terminology will be used in this questionnaire
Annex C-16
Name: .................................................................................................…
Position: ................................................................................................
Organisation: .........................................................................................
Address: ................................................................................................
...............................................................................................................
...............................................................................................................
Phone: ...................................................................................................
Fax: .......................................................................................................
Email: ...................................................................................................
It would be helpful to the interpretation of the answers to know on what basis the answers were given; for
example are they the techniques used by a researcher, a government engineer or a consultant.
or
j) In case `you don’t back-calculate stiffness moduli’ what do you do: ...............
Annex C-17
1.1 Stiffness moduli backanalysis computer program
Do you use:
d) Do you use a stiffnesses-ratio for the stiffness of some layers (e.g. the stiffness
of a granular base course = 3 x the stiffness of the subgrade) yes/no
f) What is the minimum thickness of the top modelling layer (bound layer) : ..... mm
Annex C-18
What Poisson’s ratio do you use for:
i) bituminous bound pavement layers: .............................................................
j) cement bound (pavement) layers: ................................................................
k) unbound granular road base layers: ..............................................................
l) the sub-base layer: ........................................................................................
m) the subgrade layer: ......................................................................................
n) Do you always use a 100% bond between the pavement layers yes/no
a) all deflection bowls = for every test point measured of a road section yes/no
c) the nearest match of the mean deflection bowl (of a road section) yes/no
d) a representative deflection bowl (of a road section) e.g. 85% bowl yes/no
f) Which layer thicknesses do you use: A: the mean value(s) of the road section
or B: the thicknesses measure at the point to be backcalculated A/B
1 n δ c ,i − δ m , i
n
∑
i =1 δ m,1
100% ≤ 2%
or
2
1 n ⎛ δ c,i − δ m ,i ⎞
n
∑ ⎜
⎜ δ
⎟ .100% ≤ 2%
⎟
i =1 ⎝ m ,1 ⎠
Annex C-19
d) Do you use weighting-factors (in the formula) depending on the distance
between the censor and the load centre (so the fit of one sensor is
regarded to be more important than another one) yes/no
f) Do you select a design traffic speed for the road (to be evaluated) and correct
the asphalt moduli for this traffic speed (because the load duration of the FWD
used and truck traffic load duration do not have to be similar) yes/no
e) Do you use the Longitudinal strain (due to this standard (axle) load
or the Transversal strain or the Maximum (of these two) L/T/M
f) Do you use the vertical stress at the top of the (granular) road base, due to
a standard load as a critical stress parameter in pavement evaluation yes/no
Annex C-20
Do you use the horizontal stress or strain at the bottom of the road base, due to a standard load as a critical
stress parameter in pavement evaluation:
j) Do you use the Longitudinal stress/strain (due to this standard (axle) load
or the Transverse stress/strain or the Maximum (of these two) L/T/M
k) Do you use the vertical strain or stress at the top of the sub-base, due to a
standard load as a critical strain/stress parameter in pavement evaluation yes/no
l) Do you use the vertical strain or stress at the top of the subgrade, due to a
standard load as a critical strain/stress parameter in pavement evaluation yes/no
a) by coring yes/no
f) is this location the test point with the nearest match to the
mean deflection bowl yes/no
Annex C-21
Where do you drill those cores:
i) one between the wheelpaths and one in the nearside wheelpath yes/no
l) Which accuracy do you use for unbound road base and lower layers:
within .....% of the total thickness of these layer(s)
4.2 Visual condition survey of the pavement (distress identification) for pavement evaluation
purposes
Do you register:
Annex C-22
b) Is this based on cores yes/no
e) Do you look at crack propagation (if there are cracks) in the cores yes/no
f) Do you determine the stiffness modulus of asphalt (by a test on cores) yes/no
j) Do you determine the fatigue curve by using a composition of asphalt mix yes/no
(= in this case: cement bound-, unbound and self-cementing road base (course) layers)
a) Do you drill cores for taking samples and the determination of the
type of road base material yes/no
d) Do you use a standard fatigue curve for the road base yes/no
if yes:
5.3 Sub-base
a) Do you drill cores for taking samples and the determination of the
type of sub-base material yes/no
Annex C-23
5.4 Subgrade
7. Traffic
a) Do you express the traffic loads in equivalent standard
axle loads (e.g. 80 kN of 100 kN) yes/no
f) Do you use a `Dynamic (impact) factor’ (e.g. 1.2 x static load) yes/no
if yes:
Annex C-24
if yes:
i) Which factor do you use: ............................................................................
o) Do you account for the reliability of the traffic data source yes/no
q) since road construction until survey date for the slow lane yes/no
s) the traffic growth rate or when available the truck traffic growth rate yes/no
t) Do you use a correction factor for the width of the lane due to lateral wander yes/no
Annex C-25
e) Do you compare stiffness moduli calculated from deflections
measured between the wheelpaths with stiffness-moduli calculated
from deflections measured in a wheelpath (for a road section) yes/no/sometimes
..................................................................................................................
..................................................................................................................
..................................................................................................................
c) Do you account for crack propagation in the calculation of the pavement life yes/no
or
or
d) Do you use a safety margin in calculating the residual pavement life yes/no
if yes:
Annex C-26
12. Calculation of the required thickness of strengthening overlays
a) How do you calculate the overlay thickness, if required:
....................................................................................................................
....................................................................................................................
c) Is the evaluation method you use (in your country) a described/standardized one yes/no
if yes:
Annex C-27
ANNEX D
CURRENT FWD ANALYSIS PROGRAMS
Results of a Questionnaire
December 1999
Annex D-1
D.1 General
In 1996 COST Action 336 ‘Falling Weight Deflectometer’ officially started and was a continuation of the
FEHRL-FWD group. The goal of this COST-336 Action is to develop a European common code of good
practice for the use of Falling Weight Deflectometers in pavement evaluation.
As part of the work programme of Task Group 1 'Post-processing of FWD Measurements' a questionnaire
was setup and distributed in 1999. Seventeen candidate authors of FWD analysis programs were contacted
in November and December 1999. Fourteen replied with answers to the questions. The results have been
compiled and are presented in the following pages.
1 Program Name UECSlab PAVERS BAP Evercalc CARE CANUV BOUSDEF MODCOMP 5
2 Version Number 1.1 (1999) 1.0 (2000) 1 5.0 2.10 3.3 2.01 5.1
Name of Developer (Person or KOAC Pavement Consultants KOAC Pavement Jozef KOMACKA, Ján L. Irwin, Cornell
3 Organization) (on behalf of CROW) Consultants J.R. Almeida WSDOT RHED (DWW) CELKO Haiping Zhou University
7 Fax Number +31 555 543 3111 +31 555 543 3111 360 709-5588 00-31-15-2518555 00 421 89 7243351 (775) 825-7477 +1-607-255-4080
13 Screen Language (English) English English English English Dutch Slovak English English
microns or mils, user
14 Deflection Units (deflection in mm) um (SI units) um (SI units) micron microns and mils microns mm mils (1/1000 inch) specified
MPa or psi, user
15 Moduli Units (MPp) Gpa (SI units) Gpa (SI units) MPa MPa and ksi Mpa MPa psi (pound/sq. in.) specified
Free, soon will be
Indicative cost for supply and $ 15 plus postal available over the
16 delivery (Euro, Dollar) 50 US Not priced yet N/A 2500 Euro charges No cost Web.
Annex D-2
MODULUS
UMPED PEDD MICHBACK MFPDS DAPS EFROMD2 ELMOD 5.0 SIDMOD
Michigan Flexible
Pavement Design
UMPED PEDD MICHBACK System DAPS EFROMD2 Elmod 4 MODULUS 5.1 SIDMOD
Version 4.2- July
Version 3.1- October 1999 (earlier version
1999 PEDD1) 1 1 4.2 5.1 3.0
Henan Research
Mrak Sharrock and Binh Vuong (ARRB Dynatest Center for Road NDT
W. Uddin W. Uddin/EMA Ronald Harichandran Ronald Harichandran Geoff Rowe Transport Research) International, A/S Uzan,Scullion,et al. Technology
662-915-5363 662-915-5363 (517) 355-5107 (517) 355-5107 316 687 9800 (61) (3) 98811555 +45 7025 3355 86 0371 3887447
662-915-5523 662-232-8845 (517) 432-1827 (517) 432-1827 316 687 9064 (61) (3) 98878104 +45 7025 3356 409 845 9910 86 0371 3886043
www.egr.msu.edu/~h
http://www.olemiss.ed http://www.olemiss.ed arichan/software/mp www.egr.msu.edu/~h http://www.abatech.c
u/~cvuddin/ u/~cvuddin/ mb.shtml arichan/software om www.dynatest.com www.tamu.tti.edu
Yes, 07:00-16:00
from Denmark, 13:00-
22:00 from US
Upon agreement and (Florida), 16:00-01:00
Upon request only request only No No Yes No from US (California) No no
DOS, Windows(Ver
3); Win95/98/NT(Ver Dos, Windows, Dos Windows 9x,
DOS, Windows 4) DOS Window Windows Windows Window Windows NT Windows Windows
Windows3.x, Win95/98/NT4 NT 3.x, NT 4.x, Win
Win95/98 (Version 4) Win 95/98/NT 95, 98, NT Win 95 9x 95 and 98 Windows 95/98
Minimum Pentium
Minimum 486 and up and up (Version 4) 486 Pentium 486 386, 486, pentium 486 486 or better 586
English, Spanish
English (optional) English English English English English English English
mils (English); mils (English); microns (SI), mils
microns (Metric) microns (Metric) Either mm or inches Either mm or inches microns mm (English Standard) US mils mils or microns
psi (English); MPa psi (English); kg/sq Mpa (SI), KSI
(Metric) cm or MPa (Metric) Either Mpa or psi Either Mpa or psi Mpa MPa (English Standard) US ksi MPa or ksi
US $ 1500 to $4500
US$ 40 (overseas US$ 2500 (overseas depending on number
plus 25) plus 85) Free Web download To be determined $500 Eu$1000 of users Free
No extra charge No extra charge Free Web download To be determined Included Included US $0 $20
Free by e-mail;
training charges upon 20 hours classroom 2-day course through
Free by e-mail request Yes Yes Avaiable 1/2 day training TxDOT or TTI
Annex D-3
UCESLAB PAVERS BAP EVERCALC 5.0 CARE CANUV BOUSDEF MODCOMP 5
Input requirements and method
of operation
21 Analysis Method (static or dynamic) Static Static static Static static static Static Static
up to 12 geophones,
22 Maximum number of geophones 9 7 9 10 9 7 Seven (7) up to 8 load levels
Measurement format
23 (Dynatest.F20) n/a n/a Dynatest Edition+C61 20 Both F20 and F25 Kuab.DAT User input Any
Analyzes all test points or
statistically representative test User selected (all or individual
24 points n/a n/a test locations) sas desired all test points All test points User controlled
Forward calculation method (Multi Slab on Pasternak or Winkler Multi layered linear multi layered linear Multi layered linear Multilayer linear or
32 layered linear elastic) foundation near elastic (isotropic a Multilayered Multi layered linear elastic elastic elastic elastic nonlinear elastic
Forward analysis program Van Cauwelaert OPMEKO (results Boussinesq theory CHEVLAY 2
33 (WESDEF) Van Cauwelaert (improved WESDEF) BISTRO Weslea WESDEF nearly to BISTRO) and MET (corrected version)
35 Subgrade modeling
- semi-infinite (Y/N) n/a Y Y Yes Y yes Yes Yes
- stiff layer at depth (Y/N) n/a N Y Yes N no Can be fixed by user Yes
Annex D-4
MODULUS
UMPED PEDD MICHBACK MFPDS DAPS EFROMD2 ELMOD 5.0 SIDMOD
Yes, optional Not allowed Yes Yes N Yes Yes yes yes
Error Function =
{Sum [weight.(1-
estimated
deflection/actual
Minimum absolute Minimum absolute defelction)]^2/Sum[w RMS (either mils, Weighted Absolute
difference difference Root mean square Root mean square rms eight^2]}^0.5 microns, or %) Sum Modulus Tolerance
Odemark-Boussinesq
Multi layered linear Multi layered linear Multi layered linear Multi layered linear multi layered linera Multi layered linear (method of equivalent Multi layered linear
elastic elastic elastic elastic elastic elastic thicknesses) Multi Layer Elastic elastic
BISAR for flexible,
None but WESDEF FEM & Spline semi-
PAVRAN (based PAVRAN (based can be used to analysis method for
upon ELSYM5) upon ELSYM5) Enhanced Chevron Enhanced Chevron ELSYS CIRCLY calibrate MET results Weslea rigid
Deterministic Deterministic
equations and bowl equations and bowl Singular Value bowl matching or
matching matching Bowl matching Bowl matching Dicompostion bowl matching radius of curvature bowl matching bowl matching
Single Layer with
Subgrade modeling Subgrade modeling Rock Depth
semi-infinite semi-infinite Yes, optional Yes, optional Avaiable Y Y optional Y
stiff layer at depth stiff layer at depth Yes, optional Yes, optional Available Y Y yes Y
Can backcalculate
Option to create a Option to create a Nonlinear stress the depth of the stiff
rigid bottom rigid bottom Nonlinear elastic softening user input layer
K midslab, K joints, K
E; K at mid slab, corners, load transfer E for slab, E or K for
E only interior (%) elastic moduli only subgrade
Annex D-5
UCESLAB PAVERS BAP EVERCALC 5.0 CARE CANUV BOUSDEF MODCOMP 5
Output Possibilities
Formatted ASCII,
Output file format (formatted ASCII, All options possible; also All options possible; ASCII comma deliminated, Database for residual user specified amount
37 ASCII comma delimitated) printable also printable formatted ASCII formatted ASCII life analysis module ASCII, DBF ASCII of output
Layer stiffness moduli at test temp Does not correct for
38 (Y/N) n/a Y Y Yes, for asphalt layer Y yes temp Yes
Layer stiffness moduli at standard Does not correct for
39 temp (Y/N) n/a, rigid eral E-T mix relations in N Yes, for asphalt layer Y yes temp No
variable, depending
Is standard temp fixed or variable, if Variable load time on air temperature Does not correct for
40 fixed, what is it (20 C) n/a and temperature User specified acc. to SPDM fixed (11 C) temp ---
variable, acc. to
Temperature correction approach chosen stiffness Does not correct for
41 (none, fixed, variable) n/a Shell E-T none Fixed characteristic fixed temp ---
Goodness of fit (percent error 2 % cosidered good; percent error between RMS error, plus
between measured and predicted Percent error between 2-5 % as dubious, measured and interpretive
45 bowls) yes yes percent error measured and predicted more tha 5 % bad calculated bowl Percent error comments
Annex D-6
MODULUS
UMPED PEDD MICHBACK MFPDS DAPS EFROMD2 ELMOD 5.0 SIDMOD
AASHTO equations;
No; only modulus modulus
backcalculation backcalculation As above user defined Yes, Fixed fixed
Yes; AASHTO
equations; 3 types of
No overlays User defined User defined As above user defined No not finished
maximum % error maximum % error percent error between percent error between
10% cycle 1; 20 % 10% cycle 1; 20 % measured and RMS, measured vs. measured and
later later Yes Yes Reported predicted bowls calculated bowls Yes predicted bowls
Yes (for
backcalculation but
not remaining
No; one data file at a life/overlay
Yes time No No Avaiable Yes requirements No no
Part of
comprehensive
pavement design Can be used for
package that includes airports as well as
AASHTO design, highways, Includes
linear or nonlinear database of approx
finite element 170 aircraft with gear
analysis, configurations,
backcalculation, and ARRB TR will only weights for Can analyze multi-
automatic overlay or provide EFROMD2 to evaluation/design slab for rigid
AC thickness design no CIRCLY's user purposes. pavements
Annex D-7
ANNEX E
CURRENT PRACTICE OF FWD USE ON NETWORK LEVEL
February 2000
Annex E-1
Preface
In 1996 COST Action 336 ‘Falling Weight Deflectometer’ officially started and was a continuation of the
FEHRL-FWD group. The goal of this COST Action 336 is to develop a European common code of good
practice for the use of Falling Weight Deflectometers in pavement evaluation.
The Network level information was gathered from a literature survey and a workshop held at LNEC,
Lisbon in June 1997 and - last but not least - from the experience in the COST 336 member countries. The
literature survey and the workshop are reported separately in this report.
The workshop in Lisbon in June 1997 was organised by Dr. Maria de Lurdes Antunes and her staff at
LNEC. For membership of Task Group 2, see Annex B.
Annex E-2
E.1 Introduction
This report is one of the four deliverables of Task Group 2 as described in the technical annex to the
COST 336 Memorandum of Understanding (see Annex A). Information gathered started at April 1996. To
include the newest literature, this report was updated in May 1999 at the end of COST Action 336. Other
deliverables are the Workshop in Lisbon, June 1997 and the Workshop Management Report distributed in
August 1997. The ultimate deliverable is the combination of those three reports above in the guidelines on
the applicability of the Falling Weight Deflectometer at Network Level as presented in chapter 5.
Information gathering has been performed parallel to the organisation and completion of the workshop in
Lisbon. Information gathering included the following four main items:
• Literature survey
• Selected relevant information from other COST Actions
• Detailed summaries from the Lisbon Workshop
• Supplementary COST 336 information
The literature survey started by questioning the IRRD-OECD and TRB-TRIS databases in 1997. A second
questioning updated the list in May 1999. More over an electronic mail was send to the FWD User Group
to call for papers on the subject. At last, some papers were added directly from the members of the COST
Action. Eventually, 41 titles found were analysed. Only 11 different papers addressed in one way or
another with a method actually applying to network level. Only papers with abstracts in the English
language were selected.
Valuable information was gathered from COST Action 324 (Long Term Performance of Road Pavements)
regarding intercalibration of deflection measurements. Deflections are compared from FWD, Benkelman
Beam, Curviameter and Lacroix Deflectograph.
COST Action 325 (Road Monitoring Equipment) Final Report, Chapter 5.2 presents the results from a
questionnaire. Some answers on the use of FWD at Network Level were found in 5.2.1 (Aim of Bearing
Capacity Data Collection), in 5.2.2. (Methods for Evaluating the Bearing Capacity of Roads) and in 5.2.3
(Measurement with Benkelman or FWD).
The Workshop in Lisbon June 1997 was a major part of the information gathering exercise. The four
session secretary reports summarise the workshop presentations.
Supplementary COST Action 336 information included the Task Group 1 questionnaire, the Short
Workshop report, the COST 336 home pages and chapter 5 of the main report as the final deliverable of
Task Group 2.
E.2.1 General
As a part of the objectives of COST 336, Task Group 2 produced a draft of guidelines for FWD use at
network level. This document should mainly answer the following question: "Can the FWD be used, and
Annex E-3
how, to bring structural information useable in a pavement study at network level ?" From the beginning,
it was clear that, to reach this objective, it was necessary to successively elaborate the following points:
• What is a study at network level?
• Which structural information is excessive, necessary, recommended, simply interesting and without
interest in the context of such a study at network level?
• Under which technical and economical conditions is the FWD able to provide some of this information?
The first question was discussed during a meeting of Task Group 2, and as a conclusion, a classification of
the studies at network level was proposed. This proposal is summarised in the next section of this annex.
To complete this approach, and to answer the other questions, it was decided that two approaches would
be followed in parallel: a literature survey, and a workshop.
The network level considers the road system as a set of roads arranged in different classes depending on
the function, traffic, climate, etc., rarely as one continuous route (one administrative road for example).
The minimum size of a network can be of the order of 100 km. The network level of management deals
with the following issues:
• Top level (later called 'owner level') decision, mainly in the economical and financial fields related to the
principal options of maintenance and rehabilitation ; they should be based on results easy to understand
by the executive officers and the economists
• Intermediate level (latter called 'central agency level') decisions of strategic order for localisation, priority
and scheduling of maintenance and rehabilitation; this level is also in charge of the preparation of
executive budget.
In comparison, the project level is dealing with diagnosis of the problems with the pavements and design
of the solutions, on the sections where some maintenance works were programmed from the study at the
network level. The decision level for these studies is the local agency.
Network level studies aim at providing the proper decision level with information needed for:
• Budgeting, this means defining the broad budget estimate required to reach maintenance objectives. This
budget is depending on both the actual condition of the roads (from a broad in situ evaluation) and the
condition required by the owner (maintenance objective).
• Planning, which means allocating the budget per area, agency or class of roads
• Programming, e.g. selecting all the sections of roads which actually require maintenance work
• Prioritisation, which means organising the program in a list of priority from the more to the less urgent
Annex E-4
As far as structural maintenance is addressed, evaluation of bearing capacity may be necessary or at least
useful, but it is not sufficient. It should be associated with other indicators (visual distresses, rutting,
roughness, etc.). There are different ways to assess the bearing capacity. At first sight, it was stated that
broad economic analysis and planning could be based on a statistical monitoring of the bearing capacity,
whereas programming and prioritisation require a systematic deflection monitoring.
E.2.3.1 Sources
The literature survey started in January 1997 by questioning the IRRD-OECD and TRB-TRIS databases.
The key words that were used were :
• Deflection
• Deflectograph
• Bearing Capacity
• Network
A second questioning updated the list in May 1999. More over an electronic mail was sent to the FWD
User Group to call for papers on the subject. At last, some papers were added directly from the members
of COST 336. Eventually, 41 titles found were analysed with only 11 different papers dealing with a
method, actually specific to network level. Only papers with abstracts in English language were selected.
The list of 41 papers, number [1] to [41] is provided in E.2.4 (short references). The abstracts are available
in Appendix E-1 of this annex.
E.2.3.3 Comments
The following comments should be considered with care, as the literature survey reported in this
document is not comprehensive. Furthermore, some of the papers are dealing with monitoring procedures
or testing equipment and provide no specific information on the use of deflection at network level,
although this is mentioned in the abstract.
Annex E-5
First, one can notice that, of the 41 references, only twelve papers ([2], [7], [13-15], [23-24], [29], [31],
[33-34] and [40]) address a method that is actually specific to the network level. In the other references, no
information is provided on the interpretation of data, or the described procedure is nothing but a
generalisation of a procedure that is currently used at project level. This is consistent with the fact that, in
most of the papers, the measurement procedure is systematic and the interpretation is deterministic.
Amongst these twelve papers, five i.e. [2], [7], [15], [23] and [29] are reporting measurements performed
with a Deflectograph, six others ([13-14], [24], [31], [34], [40]) are reporting FWD measurements. The
last one concerns the ASTM standard, which does not refer to a specific equipment.
As a last observation, it is often difficult to classify the actual objectives of the papers in the four
categories proposed by Task Group 2. This is not surprising since the proposed classification is still not
commonly accepted.
[1] Oliver J.E., Use of deflections to manage the structural maintenance requirements at network level in
England, 1995
[2] Sauterey R., Autret P., Guide de l'auscultation des chaussées souples, LCPC, 1977.
[3] Perrone E., Dossey T., Hudson W.R., Network level deflection data collection for rigid pavement,
1994.
[4] Beld, H. van de, Fuchs G., Bloem J., Falling weight deflection measurement as an integral tool in a
pavement management system, 1994.
[5] Briggs R.C., Harder B.T., Use of the falling weight deflectometer by the Texas state department of
highways and public transportation, 1991.
[6] Khedaywi T.S., Al-Suleiman T., Katkhuda E., Current methods of pavement evaluation in Jordan,
1991.
[7] Vervenne P.J., Janssens R., Valorization of road maintenance techniques by using standardized items
recorded in a price data base, 1990.
[8] Maennistoe V., Tapio R., Estimation of road bearing capacity with other automatically measured
condition variables, 1990.
[9] Kennedy C.K., Butler I.C., Road assessment survey systems, 1990.
[10] Parker D., A pavement management research program for Oregon highways. Final report, 1989.
[11] Schmidt B., Experiences in using Falling Weight Deflectometers as routine, 1989.
[12] Boromisza T., Gáspár L., Design of pavement condition preservation in the Hungarian pavement
management system, 1989.
[13] Scullion T., Incorporating a structural strength index into the Texas pavement evaluation system.
Final report, 1988.
[14] Gáspár L., Compilation of first Hungarian network-level pavement management system, 1994.
[15] Boulet M., Les études d'auscultation des chaussées pour la programmation des travaux d'entretien et
de renforcement du réseau routier français, LCPC, 1983.
16] Jansen J.M., Strengthening overlay design as routine procedure, a crucial star in the PMS implemen-
tation, 1990.
[17] Howard K.R., Tongue F.T., Pavement management - Development of a pilot PMS, 1995.
[18] Department of Transport, National road maintenance condition survey. Report on the 1993 survey,
1994.
[19] Kennedy C.K., Engineering interpretation of highway condition information, 1992.
[20] Chapman R.J., Routine assessment of rural roads, 1992.
[21] Hawker L.G., Routine assessment of the English trunk road network, 1992.
Annex E-6
[22] Kennedy C.K., Butler I.C., Narayanan N., Phang S.K., The Singapore pavement management
system, 1992.
[23] Kennedy C.K., Butler I.C., Road assessment survey systems, 1990.
[24] Eijbersen M.J., Van Zwieten J., Application of FWD-Measurements at the network level, 4th
International Conference on Managing Pavement, Durban, 1998.
[25] Fuchs, G., Deflectiemetingen in wegbeheer (in Dutch with English summary) Proceedings
"Wegbouwkundige Werkdagen". CROW, Ede, The Netherlands, 1994.
[26] Thewessen, H.P.M., Van de Blaak, O.R., Verkorte FWD-analyse op wegen in de provincie Utrecht
(in Dutch with English summary) Proceedings "Wegbouwkundige Werkdagen". CROW, Ede, The
Netherlands, 1994.
[27] Van den Ban, R.C.J., Valgewichtmetingen voor netwerkniveau. (in Dutch with English summary)
Proceedings "Wegbouwkundige Werkdagen". CROW, Ede, The Netherlands, 1994.
[28] Sigursteinsson, H., FWD measurement at network level in Iceland; Unpublished paper.
[29] Sršen, M., HDM-III Model – Appropriate Pavement Maintenance and Rehabilitation Programs
Under Conditions Prevailing in Central European Countries. Third International conference on man-
aging Pavements ; San Antonio, TX, 1994.
[30] Lenngren, C.A., Rolling deflectometer data strategy dos and don’ts. 5th Int. Conference on the
Bearing Capacity of Roads and Airfields. Trondheim, 6-8 July 1998.
[31] Sameh Zaghloul, Nick Vitillo, and Wei He. Project Scoping Using FWD Testing - New Jersey
Experience. Annual Meeting Transportation Research Board. Washington DC, January 1998.
[32] Elfino, M., Downsizing from network level to project level using FWD – The VA experience on I-
85. FWD User's Group Meeting 1998.
[33] American Society for Testing and Materials (ASTM)
D4694-96 Standard Test Method for Deflections with a Falling Weight Type Impulse Load Device
D4695-96 Standard Guide for General Pavement Deflection Measurements
E1166-91 Standard Guide for Network Level Pavement Management
E1777-96 Standard Guide for prioritization of Data Needs for Pavement Management
E1889-97 Standard Guide for Pavement Management Implementation
E1778-97 Standard Terminology Relating to Pavement Distress
[34] Hărătău, S., FWD at Network level in Romania. Iptana search.
[35] Benatov, Pantoja, Testing pavements in Spain with an MT-15curvometer. 13th World Meeting of the
International Road Federation. Toronto, 16-20 June 1997.
[36] Cohen, Development and implementation of a pavement rehabilitation system for Israel. 3rd
International Conference on Road and Airfield Pavement Technology. Proceedings volume 2. pp.
1126-1134. April 1998.
[37] Gáspár, L., A data time series registration subsystem for Hungarian road data bank. 3rd International
Conference on Road and Airfield Pavement Technology. Proceedings volume 2, pp 1023-1028.
April 1998.
[38] Ogras, Development of performance prediction model for prject level PMS Application in Turkey.
13th World Meeting of the International Road Federation. Toronto, 16-20 June 1997.
[39] Hoçevar, A., Use of Falling Weight Deflectometer at Network Level. Zbornik referatov, 4. slovenski
kongres o cestah in prometu, Portorož 1998 ( ISBN 961-90496-5-9).
[40] Hoçevar, A., Slovenian Experience in Using the Falling Weight Deflectometer for Determining the
Structural Adequacy of Roads, Zbornik referatov, 4. slovenski kongres o cestah in prometu, Portorož
1998 ( ISBN 961-90496-5-9 ).
[41] Orr D, Kennedy C., Pavement condition assessment in Northern Ireland. Highways and Transporta-
tion. 1998/09.
Annex E-7
Table E-2 Key information on reviewed papers - part 1
Reference Country Network Equipment
National Local ~length (km) Other FWD
[1] England England 11 000 Deflectograph
[2] France France Deflectograph
[3] US-Texas "rigid pavements" 4 600 X
[4] Netherland Gelderland X
Texas
[5] US-Texas X
Highways
[6] Jordania Jordania
[7] Belgium Belgium Deflectograph
Benkelman
[8] Finland Finland China 1340 X
Beam
[9] England England Deflectograph
[10] US-Oregon US-Oregon
[11] Denmark Denmark 15 000 X
Benkelman
[12] Hungary Hungary 30 000
Beam
[13] US-Texas US-Texas X
[14] Hungary Hungary X
[15] France France 30 000 Deflectograph
[16] Denmark Denmark 15 000 X
Urban
[17] Hong-Kong 160 X
network
[18] England England Wales 45 000 Deflectograph
[19] England England Wales Deflectograph X
[20] England Somerset Deflectograph
[21] England England Deflectograph X
[22] Singapore Singapore 2 780 Deflectograph
[23] England England Deflectograph
[24] Netherlands Netherlands 620 X
[25] Netherlands Gelderland X
[26] Netherlands Utrecht 100 X
[27] Netherlands Zeeland X
[28] Iceland Iceland X
[29] Germany Set of network 55145 Deflectograph
[30] Sweden Sweden Road deflection tester
[31] USA New Jersey 16 000 X
[32] USA Richmond Distri 65 X
Static, Dynamic or Impulse
[33] USA
device
[34] Romania Romania 4053 X
[35] Spain Spain Curviameter
[36] Israel Israel
[37] Hungary Hungary
[38] Turkey Ankara
[39] Slovenia
[40] Slovenia Slovenia 4900 X
[41] North Ireland North Ireland 1184 Deflectograph
Annex E-8
Table E-3 Key information on reviewed papers - part 2
Reference Measurement procedure Measurement parameters Interpretation
On sample Systematic Step length Lane / track Network General project
[1] X 3.8 m Track 1
[2] 20% 3m Track X
[3] 13% 1 / section Lane 1
[4] 1
[5] ? 1
[6]
[7] ? X
[8] China
[9] X 3.8 m Track 2
[10] X 2
[11] X 300 m Lane 1
[12] X
[13] X X
[14] X
[15] 4 years 3m Track X
[16] X 300 m Lane 1
[17] X 50 m 2
[18] 3, 5 years 2
[19] X 2
[20] X 3,8 m Track 2
[21] X Depending on equipment 2
[22] X 3.8 m Track 2
[23] 3.8 m Track X
[24] X 100 m Track X
[25] 1
[26] 50 m 1
[27] 1
[28] X 100 Lane 1
[29] X 3,8 Track X
[30] X 1 m or more Track 2
[31] X 10 tests per mile Lane X 1
[32] X 1
[33] X 3-500 m Lane X
[34] X 200 Lane X
[35] X 45 Lane 2
[36] X 1
[37] X 1
[38] X 1
[39]
[40] X 100 Maximum Track X
[41] X 3,8 Track 1
Annex E-9
Table E-4 Key information on reviewed papers - part 3
Reference Type of interpretation Measurement are processed to calculate
Maximum
Statistic Deterministic Deflection Moduli Residual life Other
[1] X X
[2] X X
[3] SSI
[4] X X
[5] X X
[6]
[7]
[8] X China
[9]
[10]
[11] X X X Overlay thickness
[12] X X
[13] Approach 1 Approach 2 Approach 2 SSI in approach 1
[14] Markov BC "score"
[15] X X
[16] X X X Overlay thickness
[17] X SN
[18] X X
[19] X X
[20] X X
[21] X
[22] X
[23] X X Overlay thickness
[24] X SA Index
[25] X X
[26] X IDK
[27] X X IDK
[28] X X X
[29] X SNC
[30] X X X
[31] X X
[32] X X X
[33]
[34] X X X X
[35] X X
[36] X X PCI
[37] X
[38] X SAI
[39]
[40] X X X X X
[41] X X Overlay thickness
Annex E-10
Table E-5 Key information on reviewed papers - part 4
Reference Objective of the study Observations
Global
analysis Planning Program Prioritisation
[1] X X X
[2] X X
[3]
[4] X
[5] X
[6] Almost no information
[7] X Almost no information
Research on relationship between deflection
[8]
and other parameters
[9]
[10] "not proven valuable at the network level ,,,"
[11] X
[12] X X
[13] X
[14] X X
[15] X
[16] X
[17]
[18] X X? X?
[19] X X? X? General
[20] X X? X? General
[21] Information
[22] X
[23] General
[24] X X X X Detailed
[25] In Dutch
[26] In Dutch
[27] In Dutch
[28] X X X X
[29] X X X
[30] Prototype device
Sections triggered by the PMS (based on roughness
[31] X X X X
and distress) as candidate for rehabilitation
[32] X
[33] X X X X
[34] X X X
[35] X X
[36]
[37] X
[38] X X
[39]
[40] X X X X
[41] X X
Annex E-11
E.3 Information extracted from other COST Actions.
E.3.1 General
Bearing capacity issues at network level have also been formulated in the COST Actions 324, 325 and
333.
COST Action 324 on Long Term Performance of Road Pavements, acknowledges the importance of
structural adequacy as one of the major factors affecting pavement performance. The analysis of distress
data of test sections used to build performance prediction models incorporates information on structural
condition derived from deflection measurements. This Action was completed in February 1997.
COST Action 325 on Road Monitoring Equipment, has gathered information on current road condition
monitoring practices in Europe and has identified further needs for improvements particularly in the field
of measuring bearing capacity. It analyses among other sources of information the results of an inquiry
(by ways of a questionnaire) providing the statistics of answers to questions related to bearing capacity
data collection. It is to be noted that the measuring methods covered are the Benkelman beam, the FWD,
the Lacroix Deflectograph or similar and the Curviameter. This Action was completed in November 1996.
Both Actions were targeting the issue of network level monitoring although several of the practices
recommended (COST 324) or reported (COST325) are associated to project level procedures.
The information gathered from the two COST Actions is presented in the following manner. They are
grouped into two categories in the form of logically structured quotations. One category is related to the
general aims and requirements to be met in order to conduct measurements at network level. The other
addresses the reported measurement performance linked to present day's current procedures.
COST Action 333 on Development of New Bituminous Pavement Design Methods, reviewed the current
pavement design methods in a pan-European questionnaire. An interim version (English, French,
German), Glossary of terms used in pavement design including references to other road glossaries is
available from the COST 333 homepage.
Further information on COST Actions can be found at the Internet pages http://www.cordis.lu/cost-
transport/src/
Methods to evaluate the bearing capacity of roads at network level are in a majority of cases applied to the
whole network and more scarcely only to heavy loaded roads. The purpose for collecting bearing capacity
data can be ranked in order of importance as follows:
• to determine rehabilitation measures on sections of roads
• to determine a long term maintenance plan for roads
• to predict network evolution
• to determine a long term budget for road maintenance
• for research
• to follow-up the efficiency of a maintenance policy
It is considered that the most important applications of information on bearing capacity are to determine
rehabilitation measures on sections of roads primarily and to determine long term maintenance plans for
roads secondarily.
Annex E-12
The evaluation of bearing capacity is performed either on a continuous or discontinuous basis. In the latter
case, it can be selective, regular or random. The deflection measurements are mainly carried out in the
wheelpaths, and a few between the wheelpaths or in both lines.
from the combination of the results of the priorities set on the purpose and expression of bearing capacity
data it appears that the latter is mainly used for the programming of structural maintenance on sections of
roads (project level) or for maintenance plans (network level) with the calculation of thickness (quantity of
maintenance) and residual life of the pavement (priority of maintenance). To achieve this objective
additional information is also needed on:
• Pavement structure
• Pavement temperature
• Surface distress
• Traffic
In the case of processing of data, a majority of practices use deflection measurement data in combination
with surface distress data and determine also homogeneous (uniform) sections rather based on pavement
structure or deflection measurement than on traffic volume or pavement distress.
For structural network level information, the maximum deflection alone would be sufficient (at project
level it is necessary to capture the shape of the deflection bowl). From the questionnaires, the spatial
distance between test points is close to 130 m with values ranging from 20 m to 500 m, more often from
50 m to 200 m.
Annex E-13
E.3.3.3 Measurement capacity and cost
The capacity per day varies from 10 km to 20 km (mean value approximately 14 km per day). Measure-
ment costs are in the range of € 50 to € 140 per km (1995 price level) depending partially on the number
of test points.
The main difference between FWD and rolling deflection measuring equipment comes from the fact that,
to have the same daily capacity, the distance between two test points for the FWD has to be between 50 m
to 200 m, which is too long to make it suitable to define homogeneous sections. The rolling deflection
measuring equipment has no problem with this issue since it operates with a sampling distance of typically
3 m to 5 m. Unfortunately the answers to the questionnaire do not provide any indication on the maximum
acceptable sampling distance.
The outcome of the COST 325 questionnaire stresses the need that development of new equipment should
address technical and economic issues as well as traffic safety. Target performance at network level
monitoring was summarised as follows:
• Maximum sampling distance: 5 m - 20 m.
• Deflection sensor reading accuracy: equal to or less than 0.05 mm
• Load range: 30 kN - 130 kN
• Operating speeds: up to 60 km/h for rolling deflection measuring equipment
For the network level measurements it is sufficient to measure the air and pavement surface temperatures.
The use of Ground Penetrating Radar could complete information on pavement structure (layer thick-
nesses).
Annex E-14
Table E-6 Sessions at Lisbon workshop
Session Presentations by Secretary
#1 Broad economic analysis Guillermo Albrecht (E) Aleš Hoçevar (SI)
Leif G. Wiman (S)
#2 Planning Jani Saarinen (FIN) Brian W. Ferne (GB)
J.M. Jansen (DK) Arthur van Dommelen (NL)
#3 Programming Geoff W. Jameson (AUS) Frank Clancy (IRL)
Jean-Michel Simonin (F) Francisco Sinis (E)
Brian W. Ferne (GB)
#4 Prioritisation Frank Botelho (USA) Michel Gorski (B)
Bert Thewessen (NL) László Gáspár (H)
Mate Sršen (HR)
Frank Clancy (IRL)
Gerald Cuennet (CH)
Unevenness is another sign of deterioration or even poor construction, but in a PMS, it is generally related
to increased road-user costs.
Skid resistance on the other hand is very important and synonymous with road safety. Therefore, actions
to improve skid resistance, along with transverse unevenness (rutting), usually have a high priority and are
often considered before any other measure, especially if there is a risk for accidents.
Assessment of the structural adequacy of pavements however, has traditionally been linked to project level
analysis. The fact that COST 336 had been initiated to deal with, among other issues, the use of FWDs at
network level, and that an international workshop had been organised on this topic, is a sign that this is
evidently changing. This conclusion is also supported by the results of studies mentioned previously.
The costs of laying new overlays are very high. Therefore, including information on bearing capacity at
network level - along with other parameters - certainly improves the results of a PMS in the determination
and selection of optimum solutions. It provides road agencies with better estimates for the timing of
maintenance and strengthening requirements.
In countries, where inadequate funding in the past did not enable the appropriate (timely) development of
the network (not to mention maintenance), in terms of strengthening needs, necessary widening,
construction of bypasses and motorways, the increase in traffic volume and loading that followed, caused
a rapid deterioration of the road infrastructure. In certain areas, where load limitations during the thaw
period were not respected due to the fact that a route may have provided the only accessibility, overall
damages to the pavement were even greater. In all such cases, where parts of a network are structurally in
poor condition, funds can be used economically and effectively simply by approving programmes on the
basis of prioritisation, following an evaluation of bearing capacity based on deflection measurements.
On the other end of the scale, where the development of the network had progressed with the needs, and
the construction of motorways and other major roads carrying the greatest and heaviest traffic is almost
complete (while the remaining road network had been maintained reasonably well), the strategy is more
complex and more 'fine tuning' is possible due to the availability of comprehensive road data bases and
well established road management systems. Life-cycle costs, environmental concerns, possibilities to
optimise cost/benefit effects, by applying appropriate new overlays very close to the end of the structural
lifetime of a road section, are a few among many possible considerations.
Annex E-15
E.4.1.2 Considerations for executing FWD measurements at network level
Whether or not FWD measurements are conducted at network level and what type of methodology is
adopted, depends on the overall needs and strategy formulated by the central (and/or local) road
authority/agency. In this respect, the presentations and discussions during the workshop, have shown that
there are several important points that need to be considered, such as:
• Size of the road network
• Status of the road data base regarding the availability of information on bearing capacity
• Overall importance given to particular parameters (and therefore measurements) within a PMS and the
Overall costs for executing measurements and the time needed for fulfilling the task
• Requirements of such a system
• Requirements of the customer
• Historic reasons
Overall costs for executing measurements and the time needed for fulfilling the task
In some cases, costs associated with the mere quantity of measurements determine whether a statistical or
systematic sampling method is adopted, or whether FWD measurements at network level are executed at
all. In other cases, due to safety precautions, circumstances necessitate road or lane closure, which again
has significant cost implications. Traffic disruptions, time delays and safety implications thus also need to
be considered. On the other hand, availability of government employed staff for executing various
monitoring activities may in itself, together with needs for increased equipment utilisation, favour network
monitoring, to utilise the time available between project level studies, for a useful purpose.
Annex E-16
the extent, is effectively comparable to network level monitoring, necessitates from the organisation
carrying out the task, to consider minimum necessary requirements and an appropriate methodology. The
same is true for another example given on evaluating the bearing capacity of pavements anew in Bosnia
and Herzegovina, after the war, due to the transport of tanks and other military equipment on those roads
Historic reasons
Historic reasons may include a combination of design standards, an adequate budget over a longer period
of time and sound information on traffic levels and trends. In the United Kingdom and Germany, the
motorways were designed based on standards that were more conservative, due to the fact that the budget
available enabled this and existing traffic at the time was already high enough to merit a somewhat
different approach. Moreover, the roads received timely maintenance treatments that kept prolonging the
structural lifetime of the roads. In these cases, executing deflection measurements to obtain information on
the bearing capacity of such roads is unnecessary. This however, is neither true for the remaining road
network (i.e. secondary roads, etc.), nor for roads in many other countries.
Whatever strategy the road authority/agency adopts when considering (directly or indirectly) the above
points, it remains a fact that each country needs to obtain in some way information on the bearing capacity
of its pavements at network level, at least once. In some cases, the network approach using deflection-
measuring devices has been accomplished over twenty years ago (i.e. France), in other cases, such actions
are underway. In doing so and making use of other available information, the short, medium and long-
term broad budget needs can be assessed, and the distribution of funds between new construction,
maintenance and rehabilitation (strengthening) can be determined.
Annex E-17
E.4.2 Session 2: Planning
E.4.2.2 Spain
Spain has a Pavement Management System that is used for State Road Network and for a number of
Regional Road Networks. It gives maintenance actions and budgets, both in rough terms and by road
stretches (defined by pavement age, traffic volume and pavement type). No reference was given in the
presentation or in the copies about a distribution of a central budget over regions or sub-networks. On
network level at the state network, the Curviameter is used for systematic surveys. At regional and local
government networks, the FWD is usually used for the same purpose. On project level, Lacroix
Deflectograph and FWD are used for flexible and semi-rigid pavements and FWD for rigid pavements. If
the FWD is used for input, measurements are required at every 200 m. This can be made per lane or
direction, probably depending upon road width e.g. The survey period for deflections is four years. The
approach uses empirical deterioration models and threshold values. For cracks, HDM is used.
E.4.2.3 Finland
Finland has a separate Highway Investment Programming System for planning purposes. For structural
assessment it uses KUAB FWD data, measured at 10 to 20 points per section of 1 km to 10 km, but not
further apart than 500 m. The measuring points are chosen systematically. The measuring cycle is three to
five years, depending on the variability of the former measurement. It should be noted that the same data
are also intended for the District Level Pavement Management System and even for project level. Main
derived structural parameter is the spring Bearing Capacity Ratio, determined as 160/d0 and a target value
depending on design standards and cumulative axle loads. It is only good at network level. A problem is
that the measurement is done in summer; this requires a correction factor (0.4, 0.6 or 0.8) depending on
observed frost behaviour. A statistical Markov process gives, per type of pavement per traffic volume
class, an optimal long time funding level and an optimal short term funding level, of course observing
other kinds of distress (rut depth, IRI, surface defects) and considering budgetary constraints and
economic indicators. No prediction model for the deterioration of the BCR is used, but the probability of a
change in condition class represents the deterioration. It is planned to change to SCI instead of 160/d0.
E.4.2.4 Denmark
Denmark has been using PMS since 1988. Deflections are from FWD measurements. For the minor road
network it uses only the centre deflection for pre-evaluation of the possible need for more detailed
pavement evaluation. The combination of deflection and ADT determines whether the bearing capacity is
rated good, fair or poor. The rather small major network is measured and investigated in detail, one could
say on project level. Construction details are obtained from trial pits at every 300 m and are updated with
contract overlay thickness. Asphalt thickness is considered as not so important. It was argued that some
kind of 'structural thickness' would solve a lot of problems. It was asked whether GPR would give this.
Deflections are measured at 200 m distance in both directions, staggered 100 m. Moduli are backcalcu-
Annex E-18
lated in a simplified way (by using equivalent thickness approach). These data and other strength data are
fed into the road database. Residual life and overlay need (from chosen design period) are calculated
straight away for each measuring point. As representative residual life for a section the 25% value of
individual point values is taken; for overlay thickness the 75% is taken. Future development of residual
life and overlay requirement are calculated from a (verified) theoretical degradation model. Tests are
carried out on sections with expected residual life of 5 years or less and on sections overlaid last year. The
network level information (necessary budgets; consequences of restricted budgets and such) for the major
road network comes from an aggregation of project level information.
E.4.2.6 France
France carried out a systematic sample (2 km per 10 km) Lacroix measurement of the highway network in
the mid-sixties. Spatial distance between two test points was 4 m. From the deflection data and traffic data
followed an overall analysis of the reinforcement needs (by classification into four maintenance needs)
and a works prioritisation for a period of 5 years. Recently, a new Highway PMS has been introduced
which is used both for network and project level. Again, no sub-networks are discerned. The system uses
central deflection and radius of curvature, if available, to confirm visual condition in the network analysis.
Characteristic value are taken, being mean plus two times the standard deviation. Threshold values are
given for d0 and SCI depending on traffic class and maintenance. When a pavement is confirmed to be
structurally weak, it can be subjected to complementary studies, but the decision of reinforcement can also
be taken directly.
Annex E-19
E.4.2.10 Croatia
The project described by representative from Croatia, uses SNC calculated from d0 from Lacroix in a
HDM model. This model predicts damage development. It is again a section analysis that can be
aggregated to network. For each section the optimum maintenance measure and strategy / timing is
determined. User costs are included. Net present value of savings (compared to 'do nothing') were
optimised.
E.4.2.11 Ireland
Ireland is developing some criteria for classification of pavement bearing capacity (d1 and d1 - d2 : five
classes strong to poor), subgrade bearing capacity (d9: six classes ranging from very stiff to very weak)
and overlay thickness for cost estimation (d1 and traffic: thickness). They use 200 m sections with spatial
distances of 25 m to 50 m. There is a National Road Needs Database under development that will also use
ARAN and SCRIM data, as a basis for future maintenance strategy decisions.
E.4.2.12 Switzerland
The Swiss study was a detailed (project level) assessment (including backcalculation of stiffness moduli)
of the residual life of 140 km of road in one canton, as a first step to creating a road database. FWD
measurements were taken at 25 m intervals in each direction. Thicknesses were radar measured in 0.5 m
intervals at 40 km/h while average layer thicknesses were determined every 25 m.
Only data of four countries were presented at Session 3. Mr. Geoffrey Jameson, ARRB Transport
Research from Australia presented the use of HDM III for flexible roads in the Philippines and Hong
Kong. The use of deflection measurement in network analysis in France was outlined in the presentation
by Jean-Michel Simonin from LCPC. The UK presentation was held by Brian W. Ferne, TRL. The input
made by Mr. L.G. Hawker of the Highways Agency and Dr. C.K. Kennedy of WDM Ltd. into the
presentation preparation was cordially acknowledged.
Table E-7 lists the most important data of the four countries.
Annex E-20
Table E-7 Overview of main characteristics of Session 3
Road network in Philippines Hong Kong France England
Length (km) 9000 800 45000 100001
Equipment used FWD FWD Deflectograph UK Deflectograph
Test interval (m) 500 50 4 3.5
Min. section length ca. 1 km ca. 1km 2 km per 10 km varies; 4 km on
Motorways
Maintenance Based on roughness Based on roughness Systematic Systematic
section selection
Nr. of lanes tested All All
Threshold values
Return period 3 year 3 year 4 year up to 1993 3 - 5 year
Methodology HDM III HDM III Bearing capacity Bearing capacity
Strength parameters SNC SNC Deflection, Deflection adjusted
Layer thickness for temp./age
Relevant equations SNC = SN + SNSG SNC = SN + SNSG Overlay based on Residual life and
deflection and overlay based on
traffic deflection
Limitations Not used on Reservations about Used on flexible Used on flexible
concrete roads thin layer roads roads only roads only
Deflections used d0, d900, d1500 d0, d900, d1500 Peak deflection Peak deflection
1
Trunk road network
SNC = ∑ SN + SN SG
(E.1)
SN = 0.04 ⋅ a i h i
(E.2)
The influence of structural number on other pavement parameters such as roughness and crack initiation
was also discussed. Structural number has much more influence on heavily trafficked roads(one million
standard axles per year) than on lightly trafficked roads. A modelling relationship was set up to estimate
SN and CBR (hence SNC) from FWD deflections. The derived modelling relationships were :
Annex E-21
SN = 13.5 − 6.5 log(d 0 ) + 3.7 log(d 900 )
(E.4)
Thickness and layer moduli had been obtained for over 400 sites in Hong Kong. Structural number was
calculated for these sites and then compared with modelling relationships, which had been developed
using FWD deflections (d0, d900). Regression analysis was then carried out in order to check the validity at
the modelling relationship. The regression relationship that resulted was:
813 39
SN = 1.7 + −
d 0 − d1500 d 900 (E.6)
Units are SI and all deflections were corrected by linear regression to a stress level of 700 kPa (almost
similar to load level of 50 kN). Centre deflections were normalised for temperature based on air
temperature.
The relationship between structural coefficients and layer stiffness moduli was also investigated. One
interesting point to notice is, that a given layer stiffness modulus for asphalt material can result into a
structural coefficient three to four times higher that for a cemented material with the same layer stiffness
modulus. The FWD estimated CBR values were compared to values from DCP for a number of sites. The
results indicated that more work needs to be done in this area. The speaker suggested that the deflections
used to predict CBR might need to be reviewed.
E.4.3.2 France
Network analysis using the Deflectograph began in the early sixties. The network was rated on the basis of
deflection and traffic scales. Overlay thickness tables were constructed for both hydraulic (water) and
bituminous bound materials. Deflectograph deflections are carried out every four meters during testing.
Two kilometres per ten were tested systematically every four years up to 1993.
Since 1993 systematic deflection testing has ceased as a new PMS has been introduced. The PMS is
based on visual condition of the road using high speed road monitoring. Deflection testing is now used at
project level for detailed overlay design.
Annex E-22
E.4.3.4 Comments on the presentations in Session 3
From the above workshop presentations, the following set of comments could be drafted.
France
• The mean plus two standard deviations of the deflection value was used initially for delineating roads
into homogeneous sections. According to Mr. Simonin, the 97.5 percentile is now more often used as the
distribution is expected to be normal.
• The continuous 2 km per 10 km was chosen arbitrarily.
• The overlay thickness estimated on the basis of high-speed monitor readings in the new PMS, is used for
network analyses only. Detailed overlay design is carried out at project level.
United Kingdom
• A flow chart was produced that outlined the steps involved going from using deflection at network to
project level.
• In the UK both HRM and Deflectograph are used at network level. The roughness information from
HRM is not used to identify deflection locations, because there are still some problems in matching
deflection data with roughness data on the road. The cost of Deflectograph makes it economical and
worthwhile to continue to use it systematically at network level.
• Mr. Ferne envisaged that the implementation of a Pavement Management System (PMS) would look at
the life cycle costs of design strategies.
Prioritisation is defined as organisation of the program in a list of priority from the more to the less urgent
according to a criteria based on a predefined indicator. Session 4 contains two parts:
• Response to a series of topics related specifically to prioritisation (Prioritisation specifics)
• Summary of the information that should be collected together with the data processing and reductions to
be performed retrieved from the workshop reports
Basis of prioritisation
The basis of prioritisation is to rank maintenance sections selected in the programming phase in order of
importance or urgency. These sections are the sections that already needed structural reinforcement or
fitted still in an acceptable time schedule before reaching the strengthening trigger threshold. The ranking
scale is a function of the structural adequacy or inadequacy of the maintenance sections, described by an
Annex E-23
appropriate indicator. This indicator is compatible with the criteria adopted for priority decision-making.
The indicator can be time dependent or not. The nature of the indicator is a result of the type of analysis,
technical and/or economical, which is conducted in the PMS process for optimising maintenance strategies
(Life Cycle Cost Analysis, Remaining Service Life Analysis, Multicriteria Analysis).
Deflection measurements contribute with other road condition factors (functional, structural and
environmental) to the calculation of such indicators. But in some cases statistical expressions of deflection
are used directly for the classification of priorities.
Accuracy
Accuracy of prioritisation is dependent on two factors. The first is how well the section selected for
maintenance is identified and located. In particular, how are the homogeneous zones determined and what
are the minimum lengths accepted for overlays. The second is linked to the sensitivity of the indicator
used and its ability to discriminate between structural conditions. This necessarily implies that deflection
data provided for the network level are obtained from systematic monitoring. Systematic monitoring
means continuous or regular measurements with a maximum equal sampling spacing of 20 m to 50 m. The
latter can be produced if consecutive values remain relatively constant, that is if variations of the
measurements are kept within limits, for example remain within a given class of deflections. If this is not
the case, then spacing should be reduced to a maximum of 20 m in order to detect rates of changes and
thus allow sections to be determined with an accuracy of ±20 m. This spacing should be applied if the
minimum allowable lengths for sections to be reinforced are reduced typically to 100 m or 200 m sections,
which is not so uncommon.
These parameters will have to be translated into indicators that best express the concept of bearing
capacity. Indicators currently practised are in the order of importance:
• Technical
- Strengthening thickness
• Time dependent
- Residual life of pavements
Annex E-24
- Residual ESAL capacity for the pavement
• Economical
- Cost of maintenance solution (amortisation, present net benefit, rate of return)
Remark: some management systems have an objective function to optimise, which targets the
functional aspect of pavement (road user implication) and where evenness is of major importance
(for example the HDM3 of the World Bank). In such systems deflection plays a restricted role and
is being considered only as an explanatory variable among those used to predict the progression
of unevenness.
Low budgets for the network level study implies limited affordability for data collection and monitoring.
In the context of financial constraints, there will be no alternative other than to cut down on the quality
(accuracy) and amount of information collected. But this in turn should not be detrimental to the
evaluation and sensitivity of determining the chosen priority indicator. One could switch from continuous
measurements to a statistical sampling approach applied to homogeneous sections predetermined by an
assessment other than deflection measurement, for instance, unevenness or visual inspection of cracking.
If statistical sampling is used, then it is necessary to determine the minimum number of randomly located
samples to be measured in each predetermined homogeneous road section. The risk will nevertheless
always remain as to an over or under evaluation of structural adequacy following this procedure.
If the PMS priority is based on the 'worst first' approach, the indicator to use is either the residual life of
pavements, or the residual ESAL capacity for the pavement. If the PMS is driven by the life cycle cost
approach, then the indicator, cost of maintenance solution (expressed in terms of economic performance)
will necessarily be used. Both options will display in their priorities, residual road sections whose
maintenance will have to be postponed. In general, both options will generate priority lists in different
orders.
This questionnaire was answered by 28 participants in 13 COST member countries representing today’s
FWD practice. The answers give a detailed picture of the practical post processing both for project and for
network level.
Combined with one ore more systems presented at the workshop, they will serve as a useful guide for
Road Administrations when they implement or expand the use of FWD at Network level.
The questionnaire deals with all aspects of FWD post processing from backcalculation programs over
traffic to evaluation report and quality assurance.
Annex E-25
Table E-8 Information summary
Subject Details Author
Swiss guidelines G. Cuennet (CH)
HDM3 M. Sršen (HR)
Assessment strategy Priority based on classes of residual lives H. Thewessen (NL)
Road information system associated with
F. Clancy (IRL)
intervention levels
G. Cuennet (CH)
Full deflection bowl
H. Thewessen (NL)
Partial deflection bowl H. Thewessen (NL)
Curvature indexes H. Thewessen (NL)
Required measurements M. Sršen (HR)
Visual distress
H. Thewessen (NL)
Layer thickness (GPR) G. Cuennet (CH)
M. Sršen (HR)
Traffic data
H. Thewessen (NL)
Sensor spacing @ 300 mm F. Clancy (IRL)
Test point spacing G. Cuennet (CH)
Measurement procedures
(regularly 20 m - 50 m) F. Clancy (IRL)
Transverse spacing (lane, wheelpath) H. Thewessen (NL)
Normalisation for temperature, climatic / soil
G. Salt (NZL)
condition
Engineering units, deflections and residual G. Cuennet (CH)
Data processing lives as output H. Thewessen (NL)
HDM3
SNC as output
M. Sršen (HR)
New indicators as output F. Botelho (USA)
The COST 336 workshop on the use of FWD at network was held in Lisbon in June 1997. Systems from
nine member countries and two non-member countries were presented and discussed. The Short
Workshop Report summarises the workshop with the programme, list of participants and a table showing
the major parameters in the systems presented during the workshop.
Chapter 5 of this report summarises as the final deliverable all the achieved information during the
literature survey and the Lisbon workshop, into a useful network level guide for Road Administrations in
the use of FWD at network level.
The official home page of COST Action 336 can be found at this URL :
http://www.cordis.lu/cost-transport/src/cost-336.htm
Annex E-26
Appendix E1
Annex E-27
[1]
SB: TRB-TRIS
TI: USE OF DEFLECTIONS TO MANAGE THE STRUCTURAL MAINTENANCE REQUIRE-
MENTS AT NETWORK LEVEL IN ENGLAND.
AU: Oliver-JE
SO: Conference Title: Maintenance Management. Location: Orlando, Florida. Sponsored by: Transporta-
tion Research Board; American Association of State Hig-hway and Transportation Officials; and Federal
Highway Administration. Held: 19940718-19940721. CONFERENCE PROCEEDINGS 5. 1995. pp77-
86 (4 Fig., 1 Tab., 5 Ref.)
PB: Transportation Research Board, 2101 Constitution Avenue, NW, Washington, DC, 20418, USA
PY: 1995
IS: 1073-1652
RN: 0309061067
LA: English
AB: A major task for the manager of a maintenance program is to collate and present a robust case for
funds to ensure that the network is maintained at an appropriate level. In England a project was conducted
to collate and analyze deflection data collected, principally for project-level design, for the benefit of
network-level planning. The total national highway network in Great Britain is some 11,000 km long,
representing only 4% of total road length but carrying 30% of all traffic and 60% of heavy traffic.
Deflection data usually are collected under contracts let by some 90 maintenance agents. Between 1985
and 1991 some 80% of the length of flexible roads was surveyed, some of it more than once, and the
results analyzed. On the basis that the most cost-effective strategy for strengthening is to overlay the
surface at the critical point, the network has been shown to be in suboptimal condition. The project
collated deflection data and presented them in a consistent and easily understood format to illustrate
requirements for restoring the network to its optimal condition. Condition data were also analyzed
through a network model to investigate alter-native strengthening strategies over the medium and longer
terms. Collection and analysis of data in this way offers the opportunity to carry out valuable audits of the
range of projects put forward--and not put forward--by agents. It therefore improves the management of
maintenance at both network and project levels.
DE: CONFERENCES-; MAINTENANCE-MANAGEMENT; ENGLAND-; NETWORK-LEVEL;
PAVEMENT-DEFLECTION; DATA-COLLECTION; PAVEMENT-CONDITION; DATA-ANALYSIS
SC: MAINTENANCE,-GENERAL (H40); EQUIPMENT-AND-MAINTENANCE-METHODS (I61)
PA: Transportation Research Board Business Office
AN: 00681828
Annex E-28
English & Français [2]
SB: IRRD-OECD
TI: GUIDE TO THE SOUNDING OF FLEXIBLE PAVEMENTS. (GUIDE D' AUSCULTATION DES
CHAUSSEES SOUPLES.)
AU: AUTRET-P (LCPC); SAUTEREY-R (LCPC)
SO: COLLECTION DU LCPC. 1977. XVI+180P
PB: EYROLLES, BOULEVARD SAINT GERMAIN 61, PARIS, F-75005, FRANCE
PY: 1977
LA: FRENCH
AB: THIS BOOK DEALS WITH A METHODOLOGY FOR THE STRENGTHENING OF FLEXIBLE
PAVEMENTS. SPECIAL ATTENTION IS PAID TO THE IN SITU COLLECTION OF DATA, THE
COMPOSITION OF THE WORK TEAMS, EQUIPMENT FOR SOUNDING AND LABORATORY
TESTS, AND THE INTERPRETATION OF RESULTS AS THESE AFFECT DESIGN METHODS.
DE: FLEXIBLE-PAVEMENT; 2944-; SURVEILLANCE-; 9101-; SPECIFICATIONS-; 0147-;
APPARATUS-MEASURING; 6155-; DEFLECTION-; 5586-; DETERIORATION-; 5255-; SURFACE-;
6438-; PHOTOGRAPHY-; 6751-; METHOD-; 9102-; STRENGTHENING-PAVEMENT; 3096-;
PAVEMENT-DESIGN; 3055-; PROFILOMETER-; 6103-; SOUNDING-; 5720-
SC: PAVEMENT-DESIGN (22)
AN: 106063
UD: 199511
CN: LCPC16488E
SB: IRRD-OECD
TI: GUIDE D' AUSCULTATION DES CHAUSSEES SOUPLES.
AU: AUTRET-P (LCPC); SAUTEREY-R (LCPC)
SO: COLLECTION DU LCPC. 1977. XVI+180P
PB: EYROLLES, BOULEVARD SAINT GERMAIN 61, PARIS, F-75005, FRANCE
PY: 1977
LA: FRANCAIS
AB: L' OUVRAGE, REDIGE PAR UNE EQUIPE D' INGENIEURS DU LCPC ET DES PAR (POINT
D' APPUI RENFORCEMENT) DES LABORATOIRES REGIONAUX, AIDE A FAIRE CONNAITRE
LA METHODOLOGIE MISE AU POINT POUR LES ETUDES DE RENFORCEMENTS COORDON-
NES EN DECRIVANT EN PARTICULIER LE RECUEIL DES DONNEES EN PLACE, LA COMPO-
SITION DES EQUIPES, LE MATERIEL D' AUSCULTATION ET LES ESSAIS DE LABORATOIRE,
L' INTERPRETATION DES RESULTATS ET EN DONNANT DES INDICATIONS SUR LA
METHODE DE DIMENSIONNEMENT. (12592 - GEOTECH : M LAGABRIELLE ; 12591 -
CHAUSSEES : M LEFLAIVE ; 12589 - BIBL ; 12590 - BIBL).
DE: CHAUSSEE-SOUPLE; 2944-; AUSCULTATION-; 9101-; GUIDE-RECOMM; 0147-; APPAREIL-
DE-MESURE; 6155-; DEFLEXION-; 5586-; DEGRADATION-; 5255-; SURFACE-; 6438-; PHOTO-
GRAPHIE-; 6751-; METHODOLOGIE-; 9102-; RENFORCEMENT-CHAUSSEE; 3096-; DIMEN-
SIONNEMENT-DES-CHAUSSEES; 3055-; APPAREIL-DE-MESURE-DE-PROFIL; 6103-;
SONDAGE-; 5720-
SC: DIMENSIONNEMENT-DES-CHAUSSEES (22)
AN: 106063
UD: 199511
CN: LCPC16488F
Annex E-29
[3]
SB: TRB-TRIS
TI: NETWORK-LEVEL DEFLECTION DATA COLLECTION FOR RIGID PAVEMENTS. INTERIM
REPORT.
AU: Perrone-E; Dossey-T; Hudson-WR
CA: Texas University, Austin, Center for Transportation Research, 3208 Red River, Suite 200, Austin,
TX, 78705-2650, USA; Texas Department of Transportation, Office of Research and Technology
Transfer, P.O. Box 5051, Austin, TX, 78763-, USA
SO: 1994/07. pp57 (Figs., Tabs., 17 Ref., 1 App.)
NT: Research study title: Texas Pavement Management System.
PY: 1994
RN: Report Number: TX-94-1908-3; Report Number: Res Rept 1908-3; Report Number: CTR 7-1908;
Contract/Grant Number: Study 7-1908
LA: English
AB: The existing rigid pavement deflection data contained in the Texas Pavement Evaluation System
(PES) database are evaluated and found to be inadequate for any network-level study of the structural
behavior of rigid pavements. The PES data were evaluated by comparing them with existing data
contained in the University of Texas' Center for Transportation Research (CTR) rigid pavement data-base.
Having found the data to be questionable, no further analysis was performed. For the network evaluation
of the structural behavior of rigid pavements, recommendations are provided for future falling weight
deflectometer (FWD) data collection for rigid pavements at the network level. The optimum sample size,
the testing procedures, and a cost estimate for the data collection plan are given.
DE: RIGID-PAVEMENTS; PAVEMENT-DEFLECTION; DATA-COLLECTION; FALLING-
WEIGHT-DEFLECTOMETERS; SAMPLE-SIZE; TEST-METHODS; RECOMMENDATIONS-; COST-
ESTIMATES; PAVEMENT-MANAGEMENT-SYSTEMS
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); PAVEMENT-DESIGN (I22); PROPER-
TIES-OF-ROAD-SURFACES (I23)
PA: National Technical Information Service
AN: 00677785
UD: 199602
Annex E-30
[4]
SB: TRB-TRIS
TI: FALLING WEIGHT DEFLECTION MEASUREMENT AS AN INTEGRAL TOOL IN A PAVE-
MENT MANAGEMENT SYSTEM.
AU: Beld-HVD; Fuchs-G; Bloem-J
CA: Swedish Road and Transport Research Institute, Linkoping S-58195, Sweden
SO: Conference Title: Strategic Highway Research Program (SHRP) and Traffic Safety on Two
Continents, Proceedings of the Conference. Held: 19930922-19930924. Research Institute, Transportation
Research Board. Location: Hague, Netherlands. Sponsored by: Swedish Road and Transport. 1994.
pp128-138 (6 Figs.; 6 Refs.)
PY: 1994
RN: Report Number: VTI 1A; Part 4
LA: English
AB: In this paper a method is presented to translate the results of a falling weight deflection measurement
into a straightforward matrix in such a way that a practical implementation in a pavement management
system (PMS), on network-level, will be possible. In this way, a practical combination and translation
from data collected during a visual inspection and data collected from a falling weight deflection
measurement into a PMS has been made. The presented method takes into account the effect of the
category of road (traffic load). This paper ends with some conclusions concerning the use and benefit of a
falling weight deflection measurement in a PMS.
DE: FALLING-WEIGHT-DEFLECTOMETERS; PAVEMENT-MANAGEMENT-SYSTEMS;
TRAFFIC-LOADS; VISUAL-INSPECTION; PAVEMENT-CONDITION
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); PAVEMENT-DESIGN (I22)
PA: Swedish Road and Transport Research Institute
AN: 00667777
UD: 199602
Annex E-31
[5]
SB: TRB-TRIS
TI: USE OF THE FALLILNG WEIGHT DEFLECTOMETER BY THE TEXAS STATE DEPART-
MENT OF HIGHWAYS AND PUBLIC TRANSPORTATION. STRATEGIC HIGHWAY RESEARCH
PROGRAM PRODUCTS. PROCEEDINGS OF A SPECIALTY CONFERENCE SPONSORED BY
THE HIGHWAY DIVISION OF THE AMERICAN SOCIETY OF CIVIL ENGINEERS AND THE
FEDERAL HIGHWAY ADMINISTRATION. DENVER MARRIOTT CITY CENTER HOTEL,
DENVER, COLORADO, APRIL 8-10, 1991.
AU: Briggs-RC; Harder-BT (Editor)
CA: American Society of Civil Engineers, 345 East 47th Street, New York, NY, 10017-2398, USA
SO: 1991/04. pp23-25
PY: 1991
AB: The Texas State Department of Highways and Public Transportation has 13 falling weight
deflectometers to obtain deflection information for the Pavement Management System, determine
optimum rehabilitation and reconstruction strategies for highway segments, design new pavements and
overlays, establish load restrictions on light pavements, and to determine wheel load capacity of
pavements for super heavy permit moves. SDHPT has developed procedures to utilize FWD in the
determination of elastic moduli of paving materials for design and evaluation on both rigid and flexible
pavements. Special equipment is used with each FWD to perform load transfer surveys across joints and
cracks on rigid pavements. A microcomputer program called MODULUS was developed to backcalculate
pavement layer moduli using FWD or dynaflect deflections. Pavement design and overlay thickness
determination is also accomplished with a microcomputer pro-gram. The network level deflection test
surveys of its highway system performed by SDHPT is described.
DE: STRATEGIC-HIGHWAY-RESEARCH-PROGRAM; FALLING-WEIGHT-DEFLECTOMETERS;
TEXAS-; PAVEMENT-MANAGEMENT-SYSTEMS; REHABILITATION-; MODULUS-OF-
ELASTICITY; LOAD-TRANSFER; MICROCOMPUTERS-; COMPUTER-PROGRAMS; PAVEMENT-
DESIGN; DYNAFLECT-DEFLECTIONS; CONFERENCES-
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); PAVEMENT-DESIGN (I22); PROPER-
TIES-OF-ROAD-SURFACES (I23)
PA: American Society of Civil Engineers
AN: 00607944
UD: 199602
Annex E-32
[6]
SB: IRRD-OECD
TI: CURRENT METHODS OF PAVEMENT EVALUATION IN JORDAN.
AU: KHEDAYWI-TS (JORDAN UNIVERSITY OF SCIENCE AND TECHNOLOGY); AL-
SULEIMAN-T (JORDAN UNIVERSITY OF SCIENCE AND TECHNOLOGY); KATKHUDA-E
(ARAB CENTRE FOR ENGINEERING STUDIES)
SO: AUSTRALIAN ROAD RESEARCH. 1991/06. 21(2) pp6-15 (3 Refs.)
PB: VERMONT SOUTH, VICTORIA, AUSTRALIA
PY: 1991
IS: 0005-0164
LA: ENGLISH
AB: The objective of this paper is to investigate current methods of pavement evaluation in Jordan. The
results of five methods are presented and discussed. These methods are: (i) present serviceability rating
(PSR), (ii) laboratory testing, (iii) visual inspection, (iv) pavement condition index (PCI) and (v) non-
destructive deflection testing (NDT). The results indicated that most of the pavement damage results from
heavy loading and the high percentage of trucks using the roads. Also, the degree of patching and
cracking as a measure of pavement serviceability was found to be more effective than the use of patching
alone. It was recommended finally to separate these methods depending on the objective of the
evaluation, ie, if it is at the network level or the project level. These methods can be successfully used in
the existing Highway Maintenance Management System (HMMS) in Jordan to set maintenance priorities
and allocate maintenance funds under a situation of limited budget and resources.
Annex E-33
[7]
SB: IRRD-OECD
TI: VALORIZATION OF ROAD MAINTENANCE TECHNIQUES BY USING STANDARDIZED
ITEMS RECORDED IN A PRICE DATA BASE.
AU: VERVENNE-PJ (BELGIAN ROAD RESEARCH CENTRE); JANSSENS-R (BELGIAN ROAD
RESEARCH CENTRE)
SO: PROCEEDINGS, SIXTH CONFERENCE, ROAD ENGINEERING ASSOCIATION OF ASIA
AND AUSTRALASIA, 4 TO 10 MARCH, 1990, KUALA LUMPUR; VOLUME 1 (SESSION 1, PAPER
7). 1990. 1 9P (2 Refs.)
PB: ROAD ENGINEERING ASSOCIATION OF ASIA AND AUSTRALASIA, OFFICE OF THE
DIRECTOR- GENERAL OF PUBLIC WORKS, PUBLIC WORKS DEPARTMENT HEADQUAR-
TERS, JALAN MAHAMERU, KUALA LUMPUR, 50582, MALAYSIA
PY: 1990
LA: ENGLISH
AB: In the past few years, the Roads Administration of the Belgian Ministry of Public Works has
developed, in co-operation with the Belgian Road Research Centre, a management system for the
maintenance of the Belgian road network (the 'SOGER' system). This system utilizes several files of the
Road Data Bank (RDB), of which it is a satellite. It is based, among other things, on a catalogue of
maintenance techniques, which contains an inventory of items relating to pavement maintenance. For
each item, a description is given of the operations involved, the influence on the characteristic road
parameters, and the ser-vice life resulting from the maintenance techniques in relation to traffic. The
characteristic parameters are skid resistance, longitudinal evenness, transverse evenness, deflection and
visual condition (cracking, stripping, deformation and miscellaneous). On the other hand, the system
makes use of a catalogue of standardized road work items, which contains an inventory of all items
involved in road construction, maintenance and repairs. Each item has been given a code number, so that
it can be retrieved in the various files of the data bank (A). For the covering entry of the Conference, see
IRRD number 823206.
DE: CONFERENCE-; 8525-; DEFLECTION-; 5586-; MAINTENANCE-; 3847-; MANAGEMENT-;
0145-; BELGIUM-; 8008-; ROAD-NETWORK; 2743-; DATA-BASE; 8614-; INVENTORY-; 9035-;
PAVEMENT-; 2955-; SKIDDING-RESISTANCE; 3031-; LONGITUDINAL-PROFILE; 2825-;
EVENNESS-; 3071-; CRACKING-; 5211-; STRIPPING-BINDER; 5287-; ROAD-CONSTRUCTION;
3665-; REPAIR-; 3635-; INFORMATION-DOCUMENTATION; 8555-; COST-; 0176-; CALCULA-
TION-; 6464-; DECISION-PROCESS; 2248-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); ECONOMICS-AND-
ADMINISTRATION (10); PROPERTIES-OF-ROAD-SURFACES (23)
AN: 823223
UD: 199511
CN: 9007AR617E
Annex E-34
[8]
SB: IRRD-OECD
TI: ESTIMATION OF ROAD BEARING CAPACITY WITH OTHER AUTOMATICALLY MEAS-
URED CONDITION VARIABLES.
AU: MAENNISTOE-V (VLASYS LTD, FINLAND); TAPIO-R (FINNISH NATIONAL ROAD
ADMIN)
SO: THIRD INTERNATIONAL CONFERENCE ON BEARING CAPACITY OF ROADS AND
AIRFIELDS. PROCEEDINGS. THE NORWEGIAN INSTITUTE OF TECHNOLOGY, TRONDHEIM,
NORWAY, JULY 3-5 1990. VOLUME 2. 1990. pp985-94 (7 Refs.)
PB: TAPIR PUBLISHERS, GLOESHAUGEN, TRONDHEIM-NTH, N-17034, NORWAY
PY: 1990
RN: 82-519-1033-1
LA: ENGLISH
AB: This paper presents a statistical study to use road condition variables and other technical variables
from the Road Data Bank to estimate road bearing capacity. This study was motivated by the need to
evaluate the bearing capacity distribution without using slow and expensive deflection measurements.
Three data sets were analysed, two of which were from roads in the Lappi district of Finland and the third
of which was from Shandong Province, China. A stepwise discriminant analysis was used to select the
best set of quantitative variables for a discriminant model; the set of exploratory observations classified
about 40% to 50% of the observations into the right classes. For further usage of the results, posterior
classification probabilities can be calculated from the discriminant function. The results of this study are
encouraging, despite the rather high misclassification rate of the observations. In many cases, it seems
that the required number of bearing capacity measurements can be reduced and the measuring effort can
be concentrated on the most important sections of the road network. Several problems need further
research; especially the evaluation of the number of variables needed in the dataset for convenient use of
the results. For the covering abstract of the conference see IRRD 833507.
DE: CONFERENCE-; 8525-; DATA-BANK; 8614-; BEARING-CAPACITY; 3085-; PAVEMENT-;
2955-; MATHEMATICAL-MODEL; 6473-; PLANNING-; 0133-; MAINTENANCE-; 3847-; DISTRI-
BUTION-STAT; 6572-; DETERIORATION-; 5255-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); ECONOMICS-AND-
ADMINISTRATION (10)
AN: 833592
UD: 199511
CN: 9011TR173E
Annex E-35
[9]
SB: IRRD-OECD
TI: ROAD ASSESSMENT SURVEY SYSTEMS.
AU: KENNEDY-CK (WDM LTD, UK); BUTLER-IC (WDM LTD, SINGAPORE)
SO: HIGHWAYS AND TRANSPORTATION. 1990/02. 37(2) pp10-5 (18 Refs.)
PB: INSTITUTION OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY
STREET, LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1990
IS: 0265-6868
LA: ENGLISH
AB: In this paper, the authors are concerned with the Pavement Maintenance Management Unit and in
particular with Road Assessment Survey Systems within the full Highway Management System. Road
Assessment Survey Systems include the collection, storage and processing of measurements and their
comparison with threshold or intervention levels. Detailed descriptions are given of the tools available to
produce these measurements required for routine network monitoring, including the High Speed Road
Monitor (HRM), the Sideway Force Coefficient Routine Investigation Machine (SCRIM), the High Speed
Texture Meter (a single function version of the HRM) and the Pavement Deflection Data Logging
Machine (PDDLM) or "Deflectograph".
DE: MAINTENANCE-; 3847-; HIGHWAY-; 2755-; CONTINUOUS-; 9006-; SURVEILLANCE-; 9101-
; DATA-ACQUISITION; 8623-; EVALUATION-ASSESSMENT; 9020-; DEFLECTION-; 5586-;
MEASUREMENT-; 6136-; SURFACE-TEXTURE; 3053-; SKIDDING-RESISTANCE; 3031-; DATA-
BANK; 8614-
SC: ECONOMICS-AND-ADMINISTRATION (10); MAINTENANCE (60)
AN: 827410
UD: 199511
CN: 9003TR323E
Annex E-36
[10]
SB: TRB-TRIS
TI: A PAVEMENT MANAGEMENT RESEARCH PROGRAM FOR OREGON HIGHWAYS. FINAL
REPORT.
AU: Parker-D
CA: Oregon Department of Transportation, 140 Transportation Building, Salem, OR, 97310, USA;
Federal Highway Administration, 400 7th Street, SW, Washington, DC, 20590, USA
SO: 1989/12. pp43 (Figs., Tabs., 20 Ref., 1 App.)
PY: 1989
RN: Report Number: FHWA-OR-RD-90-08; Contract/Grant Number: 5253
AB: An extensive program was developed to measure pavement deflection, skid resistance, and
rideability throughout Oregon. The data from those "objective" measures were then evaluated for
correlations with observed pavement distress and traffic factors. It is concluded that "Dynaflect"
deflections and other "objective" measures of pavement performance can best be used on the project level.
The mechanized data gathering methods evaluated here have not proven valuable in network level
pavement management.
DE: PAVEMENT-MANAGEMENT; OREGON-; PAVEMENT-DEFLECTION; DYNAFLECT-
DEFLECTIONS; SKID-RESISTANCE; RIDEABILITY-; CORRELATION-; PAVEMENT-DISTRESS;
TRAFFIC-LOADS
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); PAVEMENT-DESIGN (I22); PROPER-
TIES-OF-ROAD-SURFACES (I23)
PA: National Technical Information Service (PB90-253238/AS)
AN: 00493602
UD: 199602
Annex E-37
[11]
SB: IRRD-OECD
TI: EXPERIENCES IN USING FALLING WEIGHT DEFLECTOMERS AS ROUTINE
AU: SCHMIDT-B (ROAD DIRECTORATE, DANISH ROAD INSTITUTE)
SO: STATENS VEJLABORATORIUM NOTAT. 1989. (220) 19P
PB: ROAD DIRECTORATE, DANISH ROAD INSTITUTE, ELISAGAARDSVEJ 5, POSTBOX 235,
ROSKILDE, 4000, DENMARK
PY: 1989
IS: 0109-5315
LA: ENGLISH
AB: This report describes the Use of the Dynamic Loading Facility and the Falling Weight Deflectometer
in the Danish Pavement Management (PM) System. Within the last decade, a great deal of work has been
undertaken in Denmark to develop a Pavement Management System suitable for the Danish state road
network. In the period from 1984 to 1987 the present PM system has been tested in three different
counties and improvements and further developments of the system have been based on the results from
these tests. In late 1987 the revised PM system was considered to be reliable enough to be implemented in
all counties in Denmark. The present paper describes the experiences obtained during 1988. This
involved the determination of the structural conditions and hence the remaining lifetime and overlay
design of a particular road based on measurements from the dynamic loading facility. Falling Weight
Deflectometer measurements, and the incorporated mechanistic approach used for pavement analysis are
described as is the method of Equivalent Layer Thicknesses. Although the Falling Weight Deflectometer
is regarded as non-destructive equipment to determine the bearing capacity of a road, it is necessary when
using a backcalculation procedure such as the Method of Equivalent Layer Thicknesses to be in
possession of exact information about the pavement layer thicknesses.
DE: PAVEMENT-DESIGN; 3055-; EVALUATION-ASSESSMENT; 9020-; DYNAMICS-; 5473-;
LOAD-; 5567-; BEARING-CAPACITY; 3085-; FLEXIBLE-PAVEMENT; 2944-; MODULUS-OF-
ELASTICITY; 5919-; NON-DESTRUCTIVE; 6216-; STRENGTHENING-PAVEMENT; 3096-;
DEFLECTION-; 5586-
SC: MAINTENANCE (60); EQUIPMENT-AND-MAINTENANCE-METHODS (61); PAVEMENT-
DESIGN (22)
AN: 859211
UD: 199511
CN: 9309SL032E
Annex E-38
[12]
SB: IRRD-OECD
TI: DESIGN OF PAVEMENT CONDITION PRESERVATION IN THE HUNGARIAN PAVEMENT
MANAGEMENT SYSTEM.
AU: BOROMISZA-T (INSTITUTE FOR TRANSPORT SCIENCES, BUDAPEST, HUNGARY)
SO: ACADEMIC CONFERENCE PROCEEDINGS OF INTERNATIONAL CONFERENCE AND
EXHIBITION ON ROAD TRANSPORT, BEIJING, CHINA, 7-12 MAY 1989. 1989/05. pp205-9 (4
Refs.)
PB: TECHNOLOGY EXCHANGE CENTER OF THE MINISTRY OF COMMUNICATIONS, 22
DALIUSHUBEICUN, BEIJING, 100081, CHINA (30$)
PY: 1989
LA: ENGLISH; CHINESE
LS: CHINESE
AB: THE HUNGARIAN PAVEMENT SYSTEM IS BEING DEVELOPED USING BOTH THEO-
RETICAL, LABORATORY AND FIELD DATA. THE PAVEMENT CONDITION PRESERVATION
DESIGN SECTION IS ALREADY COMPLETED. AS PART OF THE SYSTEMATIC EVALUATION
OF THE CONDITION OF THE ENTIRE STATE MANAGED HIGHWAY SYSTEM (30000 KM) THE
PAVEMENT STRUCTURE BEARING CAPACITY AND THE SURFACE UNEVENNESS ARE
MEASURED. THIS INFORMATION, COMBINED WITH THE ROAD, TRAFFIC AND PAVEMENT
TYPE, DETERMINES THE OPTIMAL MAINTENANCE MEASURE AND ITS URGENCY RATE.
IF PAVEMENT STRENGTHENING IS NECESSARY, IT SHOULD BE DESIGNED IN A SEPARATE
STEP. THE PAVEMENT STRENGTH IS INFLUENCED MAINLY BY THE BEARING CAPACITY
WHICH IS BASED ON DEFLECTION MEASUREMENTS. A METHOD COMBINING ODEMARK-
TYPE MECHANICAL DESIGN PROCEDURE AND DEFLECTION MEASUREMENTS IS PRE-
SENTED.(A) FOR THE COVERING ABSTRACT OF THE CONFERENCE SEE IRRD ABSTRACT
NO 823365.
DE: PAVEMENT-; 2955-; ROAD-NETWORK; 2743-; MAINTENANCE-; 3847-; DATA-BANK; 8614-
; ADMINISTRATION-; 0145-; OPTIMUM-; 9056-; PRIORITY-GEN; 0131-; BEARING-CAPACITY;
3085-; EVENNESS-; 3071-; DEFLECTION-; 5586-; PAVEMENT-DESIGN; 3055-; CONFERENCE-;
8525-
SC: ECONOMICS-AND-ADMINISTRATION (10); PAVEMENT-DESIGN (22); EQUIPMENT-AND-
MAINTENANCE-METHODS (61)
AN: 824570
UD: 199511
CN: 8910CD008E
Annex E-39
[13]
SB: TRB-TRIS
TI: INCORPORATING A STRUCTURAL STRENGTH INDEX INTO THE TEXAS PAVEMENT
EVALUATION SYSTEM. FINAL REPORT.
AU: Scullion-T
CA: Texas Transportation Institute, Texas A&M University, College Station, TX, 77843, USA; Texas
State Department of Highways & Public Transp, Transportation Planning Division, P.O. Box 5051,
Austin, TX, 78763, USA; Federal Highway Administration, 400 7th Street, SW, Washington, DC, 20590,
USA
SO: 1988/04. 8409-8804 pp76 (16 Fig., 26 Tab., 21 Ref., 1 App.)
PY: 1988
RN: Report Number: FHWA/TX-88/409-3F; Report Number: Res Rept 409-3F; Report Number: TTI: 2-
18-85-409-3F; Contract/Grant Number: Study No 2-18-85-409
AB: The current Pavement Evaluation System used in Texas rates the condition of pavements in terms of
visual distress and present serviceability index. This report discusses the addition of another dimension to
the rating system; that of a Structural Strength Index. The Falling Weight Deflectometer is to be used for
this purpose. In this report, an overview is given of the FWD and data analysis techniques, a discussion
on sample size is presented and two possible structural strength schemes are proposed. The first is a
simple statistically based scheme which ranks pavement strength in terms of key deflection bowl
parameters, and includes weighting factors for traffic level and rainfall. The second is a mechanistic
approach in which a remaining service life is calculated. These two approaches were pilot tested on data
collected in several Texas districts. It was recommended that the statistically based scheme be imple-
mented. Although the mechanistic scheme shows promise at the project level, several factors including
incomplete layer information and insufficient traffic data, currently limit its applicability at the network
level.
DE: PAVEMENT-EVALUATION; PAVEMENT-MANAGEMENT-SYSTEMS; PAVEMENT-
CONDITION; PAVEMENT-DISTRESS; VISUAL-INSPECTION; SERVICEABILITY-INDEX;
STRUCTURAL-STRENGTH-INDEX; FALLING-WEIGHT-DEFLECTOMETERS; DATA-
ANALYSIS; SAMPLE-SIZE; FLEXIBLE-PAVEMENT
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); PAVEMENT-DESIGN (I22); PROPER-
TIES-OF-ROAD-SURFACES (I23)
PA: National Technical Information Service (PB88-233689/AS)
AN: 00469493
UD: 199602
Annex E-40
[14]
SB: TRB-TRIS
TI: COMPILATION OF FIRST HUNGARIAN NETWORK-LEVEL PAVEMENT MANAGEMENT
SYSTEM.
AU: Gaspar-L Jr.
SO: Transportation Research Record. 1994. (1455) pp22-30 (4 Fig., 1 Tab., 3 Ref.)
NT: This paper appears in Transportation Research Record No. 1455, Pavement Management Systems.
PB: Transportation Research Board, 2101 Constitution Avenue, NW, Washington, DC, 20418, USA
PY: 1994
IS: 0361-1981
RN: 030906063X
LA: English
AB: The first Hungarian network-level pavement management system relies on Markov transition
probability matrices. A combined condition parameter is applied taking into consideration the bearing
capacity, the unevenness, and the surface quality scores. The matrix variables are pavement type, traffic
volume, and intervention variants. The system can be used to calculate the funds needed for highways at
various condition levels, for the regional distribution of given amounts of money at a minimum cost to the
national economy, and for the determination of the economic and technical consequences of subsequent
modifications in funds distribution. Several trial runs have proved the practicability of the system.
DE: PAVEMENT-MANAGEMENT-SYSTEMS; HUNGARY-; FUND-ALLOCATIONS
SC: PAVEMENT-DESIGN-AND-PERFORMANCE (H24); ADMINISTRATION (H11); PAVEMENT-
DESIGN (I22); ECONOMICS-AND-ADMINISTRATION (I10)
PA: Transportation Research Board Business Office
AN: 00676653
UD: 199602
Annex E-41
[15]
SB: IRRD-OECD
TI: LES ETUDES D' AUSCULTATION DES CHAUSSEES POUR LA PROGRAMMATION DES
TRAVAUX D' ENTRETIEN ET DE RENFORCEMENT DU RESEAU ROUTIER FRANCAIS.
AU: BOULET-M (LCPC)
SO: RAPPORTS DES LABORATOIRES - SERIE CONSTRUCTION ROUTIERE. 1983/07. (CR-1)
59P (17 Figs.; 111 Refs.)
PB: LABORATOIRE CENTRAL DES PONTS ET CHAUSSEES, BOULEVARD LEFEBVRE,58,
PARIS CEDEX 15, F-75732, FRANCE
PY: 1983
RN: 2-7208-3502-1
LA: FRANCAIS
AB: LE LCPC ET LES 16 LABORATOIRES REGIONAUX CONTRIBUERENT A L' APPLICATION
DE LA POLITIQUE DE RENFORCEMENT DU RESEAU ROUTIER NATIONAL MIS EN PLACE
APRES LES DEGATS DE L' HIVER 1963 ET DE LA POLITIQUE D' ENTRETIEN PREVENTIF DES
CHAUSSEES NEUVES ET RENFORCEES MISE EN PLACE EN 1972 EN APPORTANT DES
ELEMENTS DE RATIONALISATION DES AFFECTATIONS BUDGETAIRES : - TOUT D' ABORD,
EN PARTICIPANT A L' EVALUATION GLOBALE DU RESEAU A RENFORCER ET A LA
DETERMINATION DES PRIORITES (PHASE D' AUSCULTATION GLOBALE), - PUIS, EN
PROCEDANT AUX ETUDES DETAILLEES DES ITINERAIRES CHOISIS AU COURS DE LA
PHASE PRECEDENTE POUR DETERMINER LES SOLUTIONS A APPLIQUER ET PREPARER LES
DOSSIERS D' EXECUTION DE TRAVAUX (PHASE D' AUSCULTATION PATHOLOGIQUE), -
ENFIN, EN EFFECTUANT LA SURVEILLANCE SYSTEMATIQUE DES CHAUSSEES NEUVES OU
RENFORCEES AFIN D' ALERTER SUR L' EVOLUTION DE LEURS QUALITES ET PREVENIR
SUR LEURS BESOINS EN ENTRETIEN. POUR CHACUNE DE CES PHASES, DES MATERIELS
ET DES METHODES D' AUSCULTATION DES CHAUSSEES ONT ETE MIS AU POINT ; ILS
PERMETTENT D' EVALUER LA QUALITE DES CHAUSSEES TANT SOUS L' ANGLE STRUC-
TUREL (PORTANCE, RESISTANCE MECANIQUE, DEGRADATION), QUE SOUS L' ANGLE
SUPERFICIEL (UNI, ADHERENCE). CES MATERIELS ET METHODES, CONCUS DANS LE
CONTEXTE FRANCAIS, ONT ETE RESPECTIVEMENT DIFFUSES ET TRANSPOSES PAR LE
LCPC DANS UN GRAND NOMBRE DE PAYS DE TOUS LES CONTINENTS DONT LES ADMINI-
STRATIONS ONT ADOPTE TOUT OU PARTIE DU MODELE FRANCAIS DE GESTION
ROUTIERE. (A).
DE: RESEAU-ROUTIER; 2743-; FRANCE-; 8036-; PLANNING-; 0133-; ENTRETIEN-; 3847-;
RENFORCEMENT-CHAUSSEE; 3096-; AUSCULTATION-; 9101-; EVALUATION-; 9020-;
GESTION-; 0145-; APPAREIL-DE-MESURE; 6155-; METHODE-; 9102-; AFRIQUE-; 8001-
SC: MATERIELS-ET-METHODES-D'ENTRETIEN (61); DIMENSIONNEMENT-DES-CHAUSSEES
(22)
AN: 115182
UD: 199511
Annex E-42
[16]
SB: IRRD-OECD
TI: STRENGTHENING OVERLAY DESIGN AS ROUTINE PROCEDURE, A CRUCIAL STAR IN
THE PMS IMPLEMENTATION.
AU: JANSEN-JM (NATIONAL ROAD LAB, DENMARK)
SO: THIRD INTERNATIONAL CONFERENCE ON BEARING CAPACITY OF ROADS AND
AIRFIELDS. PROCEEDINGS. THE NORWEGIAN INSTITUTE OF TECHNOLOGY, TRONDHEIM,
NORWAY, JULY 3-5 1990. VOLUME 2. 1990. pp927-34 (2 Refs.)
PB: TAPIR PUBLISHERS, GLOESHAUGEN, TRONDHEIM-NTH, N-17034, NORWAY
PY: 1990
RN: 82-519-1033-1
LA: ENGLISH
AB: This paper describes the implementation of the Danish Pavement Management System (PMS) from
a pilot project to a nationwide system. This was only possible through extensive collaboration between
the Road Directorate, National Road Laboratory and the Technical Administrations of the countries. As
foreseen, the introduction of strengthening overlay design as a routine procedure became a crucial part of
the implementation. 15000 falling weight deflectometer (FWD) tests were performed, to collect the
required bearing capacity data. 4000 trial pits were excavated, to measure accurately the layer thicknesses
at the analysis test points. 4000 FWD test results were backcalculated for assessment of pavement E-
moduli and needs for strengthening. The successful acquisition of the bearing capacity data was ensured
by stepwise implementation of the PMS, starting with those parts of the Danish road network that most
urgently needed attention. Available resources did not allow implementation of the whole road network in
one step. Close reference lines, extensive information for all participants, and follow-up meetings during
the implementation gave good motivation and high quality of work. 10 rules for successful implementa-
tion are given. For the covering abstract of the conference see IRRD 833507.
DE: CONFERENCE-; 8525-; DENMARK-; 8028-; PAVEMENT-; 2955-; MAINTENANCE-; 3847-;
PLANNING-; 0133-; STRENGTHENING-PAVEMENT; 3096-; BEARING-CAPACITY; 3085-;
MEASUREMENT-; 6136-; REPAIR-; 3635-; POLICY-; 0143-
SC: ECONOMICS-AND-ADMINISTRATION (10); EQUIPMENT-AND-MAINTENANCE-
METHODS (61)
AN: 833588
UD: 199511
CN: 9011TR169E
Annex E-43
[17]
SB: IRRD-OECD
TI: PAVEMENT MANAGEMENT - DEVELOPMENT OF A PILOT PMS
AU: HOWARD-KR (SMEC, HONG KONG); TONGUE-FT (HIGHWAYS DEPT, HONG KONG)
SO: HIGHWAYS AND TRANSPORTATION. 1995/09. 42(9) pp25-7 (8 Refs.)
PB: INSTITUTION OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY
STREET, LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1995
IS: 0265-6868
LA: ENGLISH
AB: This article reviews the Pilot Pavement Management System (PMS), which was implemented by the
Hong Kong Highways Department in mid-1993 with the objective of studying the practicality of the
PMS's full-scale introduction. The Pilot PMS generates life-cycle cost analyses of pavements, using the
World Bank's Highways Design and Maintenance Standard and prediction model HDM-3, and optimises
maintenance investment decision making. The network for the Pilot PMS is in Shatin new town, which
has 147 roads of total length 160km, including 40km of high-speed roads, and 20km of cycle tracks. 90%
of its roads have bituminous surfaced pavements, and 10% concrete pavements. The Australian NAASRA
roughness meter, rather than a laser profilometer, was selected, mainly because of its lower cost. The
network was tested using a Dynatest Series 8000 falling weight deflectometer (FWD) with a test spacing
of 50m. The paper also describes the use of low-cost data collection systems, to facilitate the automatic
processing of data collected by the PMS. It discusses: (1) the analysis and presentation of the data; (2) the
updating of summary results to the pilot PMS database; and (3) graphical presentation on a network basis.
DE: PAVEMENT-MANAGEMENT-SYSTEM; 3037-; COST-; 0176-; SPECIFICATION-STANDARD;
0139-; MATHEMATICAL-MODEL; 6473-; FORECAST-; 0122-; OPTIMUM-; 9056-; MAINTE-
NANCE-; 3847-; ROAD-NETWORK; 2743-; TEST-METHOD; 6288-; LASER-; 6772-; PROFILOME-
TER-; 6103-; DEFLECTOGRAPH-; 6187-; DATA-ACQUISITION; 8623-; DATA-PROCESSING;
8655-; AUTOMATIC-; 3882-; HONG-KONG; 8131-
SC: CONSTRUCTION-OF-PAVEMENTS-AND-SURFACINGS (52)
AN: 874729
UD: 199511
CN: 9511TR093E
Annex E-44
[18]
SB: IRRD-OECD
TI: NATIONAL ROAD MAINTENANCE CONDITION SURVEY. REPORT ON THE 1993 SURVEY
CA: STANDING COMMITTEE ON HIGHWAY MANAGEMENT
SO: STATISTICS BULLETIN. 1994. ((94)31) 89P
PB: DEPARTMENT OF TRANSPORT, ROOM A601, ROMNEY HOUSE, 43 MARSHAM STREET,
LONDON, SW1P 3PY, UNITED KINGDOM
PY: 1994
RN: 1-85112-214-1
LA: ENGLISH
AB: This report presents the results of the 17th visual condition survey of roads in England and Wales. It
also presents the results of a deflectograph survey measuring the structural condition of local authority
roads. This survey began in 1992. The trends in road condition over the last ten years are illustrated. A
comparison is made between road condition in 1993 and the condition in both in 1992 and ten years ago.
DE: UNITED-KINGDOM; 8119-; ROAD-NETWORK; 2743-; DETERIORATION-; 5255-; DEFLEC-
TOGRAPH-; 6187-; DEFECT-TECH; 5238-; FOOTWAY-; 2720-; KERB-; 2917-; HIGHWAY-; 2755-;
CLASSIFICATION-; 8513-; LOCAL-AUTHORITY; 0151-
SC: MAINTENANCE (60)
AN: 864454
UD: 199511
CN: 9406TR145E
Annex E-45
[19]
SB: IRRD-OECD
TI: ENGINEERING INTERPRETATION OF HIGHWAY CONDITION INFORMATION
AU: KENNEDY-CK (WDM LTD, UK)
SO: ALAN BRANT NATIONAL WORKSHOP ON PAVEMENT ASSESSMENT AND REHABILI-
TATION PROGRAMME, 14TH APRIL 1992, HELD AT THE ROYAL SPA CENTRE, LEAMINGTON
SPA, WARWICKSHIRE. 1992. pp45-54 (12 Refs.)
PB: INSTITUTE OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY STREET,
LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1992
LA: ENGLISH
AB: This paper considers the measurements available from routine surveys, and the intervention criteria
by which they can be interpreted to identify potential road maintenance schemes. Mechanically-based
surveys are undertaken using: (1) the High-speed Road Monitor (HRM); (2) the Sideway-force Coeffi-
cient Routine Investigation Machine (SCRIM); and (3) the deflectograph. To target maintenance works to
areas of most need, it is necessary to define minimum standards of condition at which road operation is
considered acceptable. Such standards are de-fined for: (1) functional condition; (2) rate of change of
functional condition; (3) safety; and (4) structural condition. They have been developed from long-term
pavement studies, supported by analytical and economic studies at the Transport Research Laboratory
(TRL). The purposes of network level evaluation are to: (1) identify what proportion of the network needs
maintenance in a given year, so that budgets can be adjusted; (2) define the general location of mainte-
nance schemes at the network level; (3) predict the future condition of the network; and (4) prioritise
network maintenance schemes, so that detailed design can be undertaken. Project level detailed design is
used to select the most appropriate engineering solution, in terms of whole life cost or pavement
condition. For the covering abstract see IRRD 867382.
DE: CONFERENCE-; 8525-; SURVEILLANCE-; 9101-; ROAD-NETWORK; 2743-; PAVEMENT-;
2955-; DAMAGE-; 1614-; DETERIORATION-; 5255-; SKIDDING-RESISTANCE; 3031-; DEFLEC-
TOGRAPH-; 6187-; MAINTENANCE-; 3847-; ADMINISTRATION-; 0145-; PLANNING-; 0133-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); PROPERTIES-OF-ROAD-SURFACES
(23)
AN: 867386
UD: 199511
CN: 9412TR053E
Annex E-46
[20]
SB: IRRD-OECD
TI: ROUTINE ASSESSMENT OF RURAL ROADS
AU: CHAPMAN-RJ
SO: ALAN BRANT NATIONAL WORKSHOP ON PAVEMENT ASSESSMENT AND REHABILI-
TATION PROGRAMME, 14TH APRIL 1992, HELD AT THE ROYAL SPA CENTRE, LEAMINGTON
SPA, WARWICKSHIRE. 1992. pp23-34
PB: INSTITUTE OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY STREET,
LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1992
LA: ENGLISH
AB: This paper discusses how rural roads are routinely assessed in the county of Somerset, which
currently spends less per km on road maintenance than almost any other English county. Thus the best
possible value must be obtained from available funds, and there is no incentive to experiment with
systems and data-bases, that are unlikely to provide a quick return on investment. Assessment techniques
and instruments include: (1) the deflectograph; (2) road monitors, which are vehicles all able to assimilate
quickly large quantities of road condition data; (3) visual condition assessment surveys; (4) road
construction surveys; and (5) skid resistance surveys. As mechanical surveys are relatively ex-pensive, in
relation to declining road budgets, coarse surveys should be directed to planning detailed surveys, which
can in turn be used for budget analysis and maintenance design. Some typical survey costs are given for
different methods. Increasing demands for assessment surveys indicate the need for better use of survey
vehicles, combined with the development of an automated coarse survey vehicle to provide a network
analysis before more detailed project analysis. It is essential to obtain enough data on which to base the
allocation of scarce re-sources. For the covering abstract see IRRD 867382.
DE: RURAL-AREA; 0328-; MAINTENANCE-; 3847-; DEFLECTOGRAPH-; 6187-; DETERIORA-
TION-; 5255-; DAMAGE-; 1614-; SURVEILLANCE-; 9101-; PAVEMENT-; 2955-; COST-; 0176-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); PROPERTIES-OF-ROAD-SURFACES
(23)
AN: 867384
UD: 199511
CN: 9412TR051E
Annex E-47
[21]
SB: IRRD-OECD
TI: ROUTINE ASSESSMENT OF THE ENGLISH TRUNK ROAD NETWORK
AU: HAWKER-LG (DEPT TRANSPORT, UK); FERNE-BW (TRL, UK)
SO: ALAN BRANT NATIONAL WORKSHOP ON PAVEMENT ASSESSMENT AND REHABILI-
TATION PROGRAMME, 14TH APRIL 1992, HELD AT THE ROYAL SPA CENTRE, LEAMINGTON
SPA, WARWICKSHIRE. 1992. pp1-22 (18 Refs.)
PB: INSTITUTE OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY STREET,
LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1992
LA: ENGLISH
AB: This paper describes various instruments and techniques for monitoring the English trunk road
network. The Department of Transport (DOT) uses two visual condition survey techniques cyclically to
assess the pavement conditions of trunk roads: (1) CHART (Computerised Highway Assessment of
Ratings and Treatments), which uses computer programs to process visually surveyed road condition data;
and (2) VCSC (Visual Condition Survey of Concrete pavements). Deflectograph surveys have been used
widely during the last 30 years, and routinely every three years since 1984. The SCRIM (Sideways force
Coefficient Routine Investigation Machine) measures the wet skidding resistance of a pavement. In 1990,
the DTp introduced the HRM (High-speed Road Monitor), which can survey up to 500km of road per day.
It consists of a van and trailer containing recording equipment, inclinometers and laser sensors. Continu-
ous surface properties of a pavement that can be measured are: (1) skid resistance, using the SCRIM; (2)
macrotexture, using the Transport Research Laboratory (TRL) high-speed texture meter; (3) rutting and
transverse profile, using the HRM; and (4) longitudinal profile, using the TRM. Measurement methods are
also available for: (1) discontinuous surface properties such as cracking and edge and surface deteriora-
tion; (2) continuous in-depth properties, using deflectographs; (3) discontinuous in-depth properties; and
(4) location referencing, using radio tags and satellite navigation. For the covering abstract see IRRD
867382.
DE: MAINTENANCE-; 3847-; SURVEILLANCE-; 9101-; ROAD-NETWORK; 2743-; MAIN-ROAD;
2748-; DAMAGE-; 1614-; DETERIORATION-; 5255-; DEFLECTOGRAPH-; 6187-; SKIDDING-
RESISTANCE; 3031-; RUTTING-WHEEL; 3081-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); PROPERTIES-OF-ROAD-SURFACES
(23)
AN: 867383
UD: 199511
CN: 9412TR050E
Annex E-48
[22]
SB: IRRD-OECD
TI: THE SINGAPORE PAVEMENT MANAGEMENT SYSTEM
AU: KENNEDY-CK (WDM LTD, UNITED KINGDOM); BUTLER-IC (WDM LTD, UNITED
KINGDOM); NARAYANAN-N (SINGAPORE. PUBLIC WORKS DEPARTMENT); PHANG-SK
(SINGAPORE. PUBLIC WORKS DEPARTMENT)
SO: SEVENTH CONFERENCE OF THE ROAD ENGINEERING ASSOCIATION OF ASIA AND
AUSTRALASIA, PROCEEDINGS, 22 JUNE - 26 JUNE 1992, SINGAPORE; VOLUME 2. 1992.
pp856-64 (6 Refs.)
PB: ROAD ENGINEERING ASSOCIATION OF ASIA AND AUSTRALASIA (REAAA), OFFICE OF
THE DIRECTOR- GENERAL OF PUBLIC WORKS, PUBLIC WORKS DEPARTMENT HEAD-
QUARTERS, 1A JALAN SETIAPUSPA, MEDAN DEMANSARA, KUALA LUMPUR, 50490,
MALAYSIA
PY: 1992
RN: 9971-88-324-4
LA: ENGLISH
AB: A rational approach to road maintenance management has been achieved in Singapore through the
introduction of a network based Pavement Management System (PMS). Engineers from the Public Works
Department are currently using the PMS to plan and control maintenance activities in 2,780 kilometres of
road in a pre-dominantly heavily urbanised environment. This paper describes the main features of the
PMS. A relational database management system is used to organise the data and present it in a user-
friendly way. Great emphasis is placed on the use of objective pavement condition data collected in a
continuous manner throughout the road network using the Deflectograph, SCRIM and High-speed Road
Monitor. The pavement condition data is used in conjunction with performance models to predict the
future condition of the network in terms of structural, skidding resistance and riding quality. Automatic
procedures use the trended survey data together with user defined intervention levels to identify road and
junction maintenance schemes and to rank them into priority order. Iterative procedures enable the user to
rapidly undertake consequential budget analyses over a user defined budget period. For the covering
entry of this conference, see IRRD abstract number 843191.
DE: PAVEMENT-MANAGEMENT-SYSTEM; 3037-; SINGAPORE-; 8101-
SC: EQUIPMENT-AND-MAINTENANCE-METHODS (61); ECONOMICS-AND-
ADMINISTRATION (10)
AN: 843282
UD: 199511
CN: 9208AR484E
Annex E-49
[23]
SB: IRRD-OECD
TI: ROAD ASSESSMENT SURVEY SYSTEMS.
AU: KENNEDY-CK (WDM LTD, UK); BUTLER-IC (WDM LTD, SINGAPORE)
SO: HIGHWAYS AND TRANSPORTATION. 1990/02. 37(2) pp10-5 (18 Refs.)
PB: INSTITUTION OF HIGHWAYS AND TRANSPORTATION, 3 LYGON PLACE, EBURY
STREET, LONDON, SW1W 0JS, UNITED KINGDOM
PY: 1990
IS: 0265-6868
LA: ENGLISH
AB: In this paper, the authors are concerned with the Pavement Maintenance Management Unit and in
particular with Road Assessment Survey Systems within the full Highway Management System. Road
Assessment Survey Systems include the collection, storage and processing of measurements and their
comparison with threshold or intervention levels. Detailed descriptions are given of the tools available to
produce these measurements required for routine network monitoring, including the High Speed Road
Monitor (HRM), the Sideway Force Coefficient Routine Investigation Machine (SCRIM), the High Speed
Texture Meter (a single function version of the HRM) and the Pavement Deflection Data Logging
Machine (PDDLM) or "Deflectograph".
DE: MAINTENANCE-; 3847-; HIGHWAY-; 2755-; CONTINUOUS-; 9006-; SURVEILLANCE-; 9101-
; DATA-ACQUISITION; 8623-; EVALUATION-ASSESSMENT; 9020-; DEFLECTION-; 5586-;
MEASUREMENT-; 6136-; SURFACE-TEXTURE; 3053-; SKIDDING-RESISTANCE; 3031-; DATA-
BANK; 8614-
SC: ECONOMICS-AND-ADMINISTRATION (10); MAINTENANCE (60)
AN: 827410
UD: 199511
CN: 9003TR323E
Annex E-50
[24]
Abstract
This paper describes the development of two indicators for the assessment of the
structural adequacy and the structural condition of roads at the network level. The
basis for the development are the current Dutch design procedure for flexible
pavements and data from SHRP-NL test sections. For both indicators a minimum of
input parameters is required, which facilitates their use at the network level of
pavement management. The outcome of three verification studies shows that results
from the indicators correlate very well with the results from detailed project level
studies. Further, the verification studies indicate that the input parameters for the
indicators can indeed be gathered in network level monitoring of pavements.
[25]
DEFLECTIEMETINGEN IN WEGBEHEER
ir. G. Fuchs
DHV Milieu en Infrastructuur BV
Amersfoort
Summary
In this paper a method is presented to translate the results of a falling weight deflection measurement into
a straight forward matrix in such a way that a practical implementation in a pavement management system
(PMS), on network-level, will be possible. In this way a practical combination and translation from data
collected during a visual inspection and data collected from a falling weight deflection measurement into a
PMS has been made. The presented methods takes into account the effect of the category of road (traffic
load). This paper ends with some conclusions concerning the use and benefit of falling weight deflection
measurement in a PMS.
Annex E-51
[26]
Summary
This paper presents a method which quickly analyses the structural condition of road pavements based on
FWD-measurements. It is used to evaluate about 100 km of provincial road in the Province of Utrecht.
The method has been verified using an available visual condition survey and the results of ARAN
measurements. Differences in results have been further investigated by additional core drilling in order to
determine the cause of these differences like cracking in top layers only or full depth cracking.
[27]
Summary
In the province of Zeeland Falling Weight Deflection measurements were used for determination of the
life of road constructions and for determination of improvements with linear elastic multi layer systems.
from practical experience it seemed that this could be done in a more simple way from these measure-
ments with D0, D180 and IDK60. Temperature corrections for D0 seemed necessary. For this, relations
have been developed. For D180 the thickness of the upper layer seemed to influence the level of this
measurement because of dynamic effects. Here it has given a try correct it to a level compared with a sub-
base modulus without the influence of the upper layers.
Annex E-52
[28]
Forced measurement : A visual inspection of structural distress is performed at least once a year. If
distress is seen in some part of the road, the whole section is measured with a FWD.
Normal measurement will be at 100 m interval in both directions, the roads is then measured with 50 m
intervals.
All FWD results will be saved in database and can bee seen on maps on computer network. There we will
also be able to see if there is a loss of bearing capacity in tons from one measurement to an other.
Annex E-53
[29]
Mate Sršen ; “ HDM III Model – Appropriate Pavement Maintenance and Rehabilitation Programs Under
Conditions Prevailing in Central European Countries” Third International Conference on Managing
Pavements ; San
Antonio ; Texas ; USA ; May 21-24 1994 ; National Research Council ; TRB ;
Washington D.C. pp 246-256
Abstract
This paper considers possibilities for using the HDM - 3 (Highway Design and Maintenance) model to
optimize selection of appropriate pavement maintenance and rehabilitation programs under conditions
prevailing in Central European countries. The consideration was carried out on a part of the Bavarian road
network, and it is the first attempt to apply the HDM-3 model for optimizing road maintenance and
rehabilitation in this West European country. The results obtained clearly point the numerous advantages
of using the HDM-3 model, if regional specificities are realistically presented. Particular emphasis is
placed on the critical analysis of HDM-3 model results with respect to the great number of input
parameters used as a basis for modeling. The influence of various parameters on the performance of the
model is analyzed in detail. In a study of roads in Bavaria we found a prevalence of rehabilitation
strategies. Through a sensitivity analysis we showed that the roughness and traffic volume variables
dominate the outcome because user cost comprise 94 percent of the total cost and lessen the impact of
maintenance cost on the choice of optimum strategy.
[30]
C.A. Lenngren “ Rolling deflectometer meter data strategy dos and don’ts” Fifth international Conference
on the bearing capacity of roads and Airfields ; Trondheim ; 6-8 July 1998
ABSTRACT
Pavement management input data in Sweden are primarily based on surface characteristics. These data are
good for assessing user costs but they are of less value for estimating maintenance and reconstruction
costs. Structural assessment is very important and the falling weight deflectometer can be used for this
purpose. It has proven to be an excellent tool for overlay design. For this purpose its simplicity and
straightforwardness are well documented. However, the relatively low capacity per unit makes this device
less attractive for pavement management network level use. In addition the falling weight deflectometer
stationary operation mode requires lane closures and high user costs on high volume roads. The Swedish
National Road Administration decided that a high-speed rolling deflection tester should be tried using
lasers sensors mounted on a heavy truck. The Swedish Road and Transportation Institute had a working
system ready by the end of 1997. Forty sensors are mounted on a heavy truck to determine two transverse
profiles. One profile constitutes an unloaded case. The other profile just behind the rear wheels of the
vehicle constitutes the loaded case. The high sampling rate is adequate for filtering the macro texture of
the pavement. The present paper discusses some different strategies for optimizing the rolling deflection
tester usage.
Annex E-54
[31]
New Jersey experience with pavements, like many other states, goes back to the last century. Highways
constructed early in this century are still in service. During these extensive service lives, several
maintenance and rehabilitation activities were applied to keep the pavements in good condition. These
activities ranged from patching to full reconstruction. Also, most of New Jersey pavements have been
widened at least once. With all these factors, it is difficult to identify the limits of homogeneous sections
which should receive the same rehabilitation treatment.
[32]
Reduction in state expert personnel due to retirement, and lacking of funds are driving factors to make use
of the Falling Weight Deflectometer to optimize our pavement rehabilitation strategies. The case study,
presented here is an excellent example of a network level downsized to project level on 1-85 in the South
Hill Residency in the Richmond District, Virginia.
The initial approach, used by the newly appointed, maintenance manager is to core the flexible pavement
of the 40 mile stretch, at every mile post, in both directions (I.e. 80 cores). These were taken, whether we
have a distress or not. The cores and pavement condition survey together did not provide the maintenance
manager with enough information, so he can select his rehabilitation alternatives and optimize the use of a
limited fund. At this point the non-destructive testing team was called to provide FWD testing and
conduct field diagnostic investigation. In Virginia FWD testing and coring are not enough to completely
evaluate
the pavement structure, we furthermore trench the shoulder to expose the different layers and investigate
the effect of poor drainage. The presentation includes the FWD test results, the challenges in back
calculation, sieve analysis of subgrade soils, matching the as build plans with the existing pavement
structure, the damage caused due to lack of drainage, traffic data and projection, rehabilitation alterna-
tives, and plan of action in order that the proper treatment is applied to the specific pavement section,
while we still can afford the rehab work within the limited budget. In essence, we downsized a network
level of 80 miles to a successful project level of nine miles.
Annex E-55
[33]
ASTM Standards
Reprinted, with permission , from the Annual Book of ASTM Standards, copyright American Society for
Testing and Materials, 100 Bar Harbor Drive, West Conshohocken, PA 19428. Copies of ASTM standards
may be purchased from ASTM, phone: 610-832-9555, fax: 610-832-9585, e-mail: [email protected],
website: www.asm.org .
E1777-96 Standard Guide for Prioritization of Data Needs for Pavement Management
1.1 This guide identifies data needs for pavement management systems. It also addresses the relative
importance of various types of pavement data.
1.2 This guide was developed for use by federal, state, and local agencies, as well as consultants who
provide services to those agencies.
1.3 This guide describes a process and provides a set of recommendations that any agency may use to
develop a plan for acquiring pavement management data. Any individual agency may justifiably assign
higher or lower priority to specified data items depending on their needs and policy.
Annex E-56
1.4 This standard does not purport to address all of the safety concerns, if any, associated with its use. It is
the responsibility of the user of this standard to establish appropriate safety and health practices and
determine the applicability of regulatory limitations prior to use.
D4695-96 Standard Guide for General Pavement Deflection Measurements
1.1 This guide consists of recommendations for measuring vertical or normal pavement surface deflections
resulting from the application of a known transient load, a steady-state dynamic load, or an impulse load
applied by a non-destructive deflection testing (NDT) device. This deflections are measured with sensors
that monitor the vertical movement of the pavement surface directly under or at locations radially outward
(offset), or both, from the load center. Several offset deflection measurements at a specific test location
describe what is called a deflection "basin". Each NDT device is operated according to the standard
operating procedure applicable to the device. Also are given recommendations for collection of general
information regardless of the type of testing device used such as ambient temperature, pavement
temperature, equipment calibration, number of tests, and test location.
1.2 This guide is applicable for deflection measurements made on flexible (asphalt concrete (AC)), rigid
(Portland Cement Concrete (PCC) or continuously reinforced concrete (CRCP)), or composite (AC/PCC)
pavements.
1.3 This guide provides general information that is required for three suggested levels of testing effort, as
follows:
1.3.1 Level I -a general overview of pavement condition for network analysis.
1.3.2 Level II -a routine analysis of the pavement for purposes such as overlay or rehabilitation design
projects.
1.3.3 Level III -a detailed or specific analysis of the pavement, such as the evaluation of joint efficiency or
foundation support for PCC slabs.
1.4 The values stated in SI units are to be regarded as standard. Inch-pound units given in parentheses are
for information purposes only.
1.5 This standard may involve hazardous materials, operations, and equipment. This standard does not
purport to address all of the safety concerns, if any, associated with its use. It is the responsibility of the
user of this standard to establish appropriate safety and health practices and determine the applicability of
regulatory limitations prior to use.
D4694-96 Standard Test Method for Deflections with a Falling-Weight-Type Impulse Load Device
1.1 This test method covers the measurement of deflections of paved and unpaved surfaces with a falling-
weight-type impulse load device. These devices are commonly referred to as falling-weight deflectome-
ters.
1.2 This test method describes the measurement of vertical deflection response of the surface to an
impulse load applied to the pavement surface. Vertical deflections are measured on the load axis and at
points spaced radially outward from the load axis. An impulse load more nearly represents the moving
vehicle load-pulse applied to prototype pavements than does a static load.
1.3 The values stated in SI units are to be regarded as the standard.
1.4 This standard does not purport to address all of the safety concerns, if any, associated with its use. It is
the responsibility of the user of this standard to establish appropriate safety and health practices and
determine the applicability of regulatory limitations prior to use. A specific hazard statement is given in
Section 7.
Annex E-57
[34]
Annex E-58
[35]
SB: IRRD-OECD
TI: TESTING PAVEMENTS IN SPAIN WITH AN MT-15 CURVOMETER. (EVALUACION DE
PAVIMENTOS EN ESPANA CON CURVIAMETRO MT-15.)
AU: BENATOV-BB (EUROCONSULT N.T., S.A. SPAIN); PANTOJA-JR (EUROCONSULT N.T.,
S.A. SPAIN)
SO: PROCEEDINGS OF THE XIIITH WORLD MEETING OF THE INTERNATIONAL ROAD
FEDERATION, TORONTO, ONTARIO, JUNE 16 TO 20, 1997. 1997. pp-
PB: TRANSPORTATION ASSOCIATION OF CANADA (TAC), 2323 ST. LAURENT BLVD,
OTTAWA, ONTARIO, CANADA
PY: 1997
LA: ESPAGNOL
LS: ENGLISH
AB: Initially designed as equipment to receive pavements at the moment of their inauguration, the use of
the MT-15 CURVOMETER has changed over the last several years. It has become a piece of equipment
used to test pavement for Spain's Highway and Airport Network and a very appropriate tool for assessing
the process of laying down the layers which make up the road surface on highways and airport runways.
Consisting of a 15 meter-long chain development with three sensors inserted every 5 metres, the MT-15
CURVOMETER is able to measure deflexions at 18 km/h, with precisions up to the hundredth of a
millimetre (mm/100). This great capacity enables the MT-15 to carry out an initial assessment of a
highway the day and even the night before the opening of the road, to obtain the point zero of the roads
structural record. During the past several years, following a meticulous study carried out by the General
bureau of Spain's Highways that standardized deflexion tests, the CURVOMETER is now used for
systematic assessment of the Highway Network and for programming pavement upgrading works. The
MT-15 CURVOMETER has also been very well received by the Spanish airport administration, by
adding to its list of technical achievements, minimal interference with airport operations thereby enabling
a detailed assessment of a runway in a single night without having to officially shut down the runway. The
MT-15s ability to test even granular surfaces is enabling assessment of bearing capacities and layer
homogeneity of pavements during their construction. As a result of this ability, compacting defects can be
corrected before the top layer is even laid down, thereby improving pavement quality and reducing costs
of expensive subsequent repairs. For the covering abstract of this conference see IRRD number 872978.
(Author/publisher).
DE: CANADA-; 8018-; CONFERENCE-; 8525-; DEFLECTOGRAPH-; 6187-; PAVEMENT-; 2955-;
SURFACE-; 6438-; TEST-; 6255-; HIGHWAY-; 2755-; AIRPORT-; 2776-; SPAIN-; 8105-; CURVO-
METER-
SC: PAVEMENT-DESIGN (22); PROPERTIES-OF-ROAD-SURFACES (23)
AN: 897006
UD: 199811
CN: 9806RT361E
Annex E-59
[36]
SB: IRRD-OECD
TI: DEVELOPMENT AND IMPLEMENTATION OF A PAVEMENT REHABILITATION SYSTEM
FOR ISRAEL.
AU: COHEN-E (T & M TECHNOL & MANAGEMENT LTD., ISRAEL); UZAN-J (TECHNION,
HAIFA, ISRAEL); SAHAR-I (T & M TECHNOL & MANAGEMENT LTD., ISRAEL); COHEN-S (T &
M TECHNOL & MANAGEMENT LTD., ISRAEL)
SO: 3RD INTERNATIONAL CONFERENCE ON ROAD & AIRFIELD PAVEMENT TECHNOLOGY,
PROCEEDINGS VOLUME 2. 1998/04/28. pp1126-34 (9 Refs.)
PB: INFORMATION INSTITUTE OF SCIENCE AND TECHNOLOGY, MINISTRY OF COMMUNI-
CATIONS, P.R. CHINA, 240 HUIXINLI, CHAOYANG DISTRICT, BEIJING, 100029, CHINA (80$)
PY: 1998
RN: 962-8267-06-9
LA: ENGLISH
AB: A rehabilitation system was developed to meet the needs of the State of Israel, to upgrade the
deteriorated network to an acceptable level. The system is based on Pavement Condition Index (PCI) and
Present Serviceability Rating (PSR) surveys, Non Destructive Tests (NDT) of deflection basin measure-
ments, moduli backcalculation, a limited amount of drilling and a current design method for flexible
pavements. The system was used to rehabilitate 200 road sections. The methodology is described and
illustrated. Using the data as described above, the existing pavement structure was evaluated and
converted into an effective structural thickness. Using the above, the required Equivalent AC Overlay
(EACO) and the rehabilitation strategy were determined. A follow-up of some rehabilitated sections is
being conducted. For the covering abstract of the proceedings, see IRRD 490713.
DE: PAVEMENT-MANAGEMENT-SYSTEM; 3037-; REPAIR-; 3635-; ISRAEL-; 8050-; EVALUA-
TION-ASSESSMENT; 9020-; THICKNESS-; 5933-; STRENGTHENING-PAVEMENT; 3096-
SC: ECONOMICS-AND-ADMINISTRATION (10); PAVEMENT-DESIGN (22)
AN: 490856
UD: 199808
CN: 9806CD029E
Annex E-60
[37]
SB: IRRD-OECD
TI: A DATA TIME SERIES REGISTRATION SUBSYSTEM FOR HUNGARIAN ROAD DATA
BANK.
AU: GASPAR-L (INST FOR TRANSPORT SCIENCES LTD (KTI RT), HUNGARY); BORS-T (INST
FOR TRANSPORT SCIENCES LTD (KTI RT), HUNGARY)
SO: 3RD INTERNATIONAL CONFERENCE ON ROAD & AIRFIELD PAVEMENT TECHNOLOGY,
PROCEEDINGS VOLUME 2. 1998/04/28. pp1023-8 (3 Refs.)
PB: INFORMATION INSTITUTE OF SCIENCE AND TECHNOLOGY, MINISTRY OF COMMUNI-
CATIONS, P.R. CHINA, 240 HUIXINLI, CHAOYANG DISTRICT, BEIJING, 100029, CHINA (80$)
PY: 1998
RN: 962-8267-06-9
LA: ENGLISH
AB: This paper presents a data time series registration subsystem for Hungarian road data bank. The
subsystem covers 500Km long highways in the whole Hungarian national highway network, in which
knot-type (junction-type) highway reference system is applied. Every year the pavement condition data
measured (roughness, rutting depth, bearing capacity and surface defects) are collected into the subsystem
as well as the list of maintenance interventions (overlays, thin asphalt layers, surface dressings etc.). This
paper deals with the following respects: (1) the main goals of the subsystem; (2) the creation of the
subsystem; (3) the operation of the subsystem, including necessary actualization of condition data,
consideration of changes in highway network during previous year, and relationship between the road data
bank and the subsystem. For the covering abstract of the proceedings, see IRRD 490713.
DE: HUNGARY-; 8043-; DATA-BANK; 8614-; HIGHWAY-; 2755-; PAVEMENT-MANAGEMENT-
SYSTEM; 3037-; TIME-; 5414-; MAINTENANCE-; 3847-
SC: ECONOMICS-AND-ADMINISTRATION (10)
AN: 490846
UD: 199808
CN: 9806CD019E
Annex E-61
[38]
SB: IRRD-OECD
TI: DEVELOPMENT OF PERFORMANCE PREDICTION MODEL FOR PROJECT LEVEL PMS
APPLICATION IN TURKEY.
AU: OGRAS-T (HIGHWAYS, TURKEY); HAAS-R (UNIVERSITY OF WATERLOO)
SO: PROCEEDINGS OF THE XIIITH WORLD MEETING OF THE INTERNATIONAL ROAD
FEDERATION, TORONTO, ONTARIO, JUNE 16 TO 20, 1997. 1997. pp- (3 Refs.)
PB: TRANSPORTATION ASSOCIATION OF CANADA (TAC), 2323 ST. LAURENT BLVD,
OTTAWA, ONTARIO, CANADA
PY: 1997
LA: ENGLISH
AB: Pavements represent the largest single area of investment in highway transport infrastructure. Like
any large investment, they deserve good management. This is particularly so, because pavements are also
subjected to fairly rapid deterioration. A total pavement management system (PMS) consists of a
coordinated' set of activities, all directed toward achieving the best value for the available public funds in
providing and operating smooth, safe and economical pavements. The measurement and prediction
(modeling) of pavement performance is a cornerstone of any PMS for both network and project level,
because it is necessary for both the financial planning/budgeting and pavement life cycle economic
evaluation. A pilot PMS for Turkey's roads has been carried out for 72 hot-mix asphalt concrete sections
within the Ankara region, The necessary data (including FWD deflections, surface distress, roughness, etc.
were collected and evaluated. A follow-up to the network project was a study to develop a project level
performance prediction model for Turkey's roads by using the network data. The results of the following
study are reported in this paper. In one part of the study, a relationship between PQI (Pavement Quality
Index) vs. cumulative traffic loading and age is developed by using multivariate linear regression analysis
after classification of the sections according to thickness (EGT), subgrade' strength (MK) and cumulative
traffic loading (ESAL). In another part, a set of performance curves in terms of PSI vs. ESAL are drawn
for different SN (structural number) values. These SN values are calculated by running the software
DARWin of AASHTO'93 for Turkey's condition representing different compositions for different traffic
and different subgrade strengths. In the third part of the study, the OPAC (Ontario Pavement Analysis or!
Costs) performance model which is based on a relationship between subgrade deflection and pavement
deterioration, is modified for Turkey's roads which are extremely high trafficked when compared to
Ontario's. In order to achieve more reliable performance curves, it is recommended that all these
approaches should be further verified by means of periodic data collection and evaluation. For the
covering abstract of this conference see IRRD number 872978. (Author/publisher).
DE: CANADA-; 8018-; CONFERENCE-; 8525-; PAVEMENT-; 2955-; PAVEMENT-
MANAGEMENT-SYSTEM; 3037-; HIGHWAY-; 2755-; BEND-ROAD; 2872-; PAVEMENT-DESIGN;
3055-; TURKEY-
SC: ECONOMICS-AND-ADMINISTRATION (10); PAVEMENT-DESIGN (22)
AN: 896978
UD: 199808
CN: 9804RT433E
Annex E-62
[39]
Aleš Hoçevar, Use of Falling Weight Deflectometer at Network Level, Zbornik referatov, 4. slovenski
kongres o cestah in prometu, Portorož 1998 ( ISBN 961-90496-5-9 )
Summary:
Pavement management systems require a large number of input data to be obtained by monitoring the
condition of the road network. The most important ones include longitudinal and transverse unevenness,
surface distress, cracking and skid resistance. The bearing capacity of pavements (or structural adequacy),
however, is also gaining in importance. Structural adequacy is usually determined by the size of measured
deflections, by knowing the shape of the deflection bowl and deflection radius, as well as by using various
statistical parameters.
In the past, deflection measurements were mainly executed at project level for the purpose of determining
necessary treatments (overlays, strengthening, etc.) on a certain road or on a section of a road. This
situation is changing. Deflections are increasingly being measured at network level. In the paper that
follows, the main reasons for performing measurements at network level are presented together with
results of recent studies that confirm this. Finally, some general criteria are provided that help determine
when a network level approach deserves considerations.
[40]
Aleš Hoçevar, Slovenian Experience in Using the Falling Weight Deflectometer for Determining the
Structural Adequacy of Roads, Zbornik referatov, 4. slovenski kongres o cestah in prometu, Portorož 1998
( ISBN 961-90496-5-9 )
Slovenian Experience in Using the Falling Weight Deflectometer for Determining the Structural
Adequacy of Roads
Summary:
This paper has been written at a time which coincides with the end of the first round of deflection
measurements. These were executed on the main and regional roads in Slovenia using a Falling Weight
Deflectometer. The paper describes the methodology used and provides reasons for the network level
approach. The primary aim was to obtain a complete overview of the structural adequacy of the road
network. This would enable the identification of all critical sections of roads with respect to the bearing
capacity. Other reasons are also mentioned.
The paper also reports some general results of the concluded monitoring. A short description of the
adopted procedure with other relevant information is also provided. In the conclusion, the author makes an
assessment of the reasonableness for the decision to use an FWD at network level.
Annex E-63
[41]
SB: IRRD-OECD
TI: PAVEMENT CONDITION ASSESSMENT IN NORTHERN IRELAND.
AU: ORR-D (ROADS SERVICE HQ); KENNEDY-C (WDM LTD)
SO: HIGHWAYS AND TRANSPORTATION. 1998/09. pp24-6
PB: INSTITUTION OF HIGHWAYS AND TRANSPORTATION, 6 ENDSLEIGH STREET, LON-
DON, WC1H 0DZ, UNITED KINGDOM
PY: 1998
IS: 0265-6868
LA: ENGLISH
AB: The Roads Service agency of Northern Ireland's Department of the Environment is responsible for
24,490km of public roads. It aims to ensure the provision of a safe and effective road network throughout
Northern Ireland, while protecting environmental quality. This article shows how pavement management
contributes to that aim by maintaining the roads in a reasonable and serviceable condition. At network
level, information about road condition must measure relative performance from year to year and between
different parts of the network, and establish what investment is needed to bring the network up to an
agreed standard. At project level, information must identify where repairs are needed to maintain safety
and structural capacity, and assign a priority order for them. Measurement equipment and methods used
include: (1) an automated lorry-mounted deflectograph to assess strength of the pavement below; (2) a
SCRIM (Side-force Coefficient Routine Investigation Machine) continuously measuring wet skidding
resistance; (3) an HRM (High speed Road Monitor) providing a quick and effective overview of road
condition; and (4) the computer-based MARCH (Maintenance Assessment, Rating and Costing for
Highways) system. The article also discusses safety inspections, key performance indicators, and future
developments.
DE: PAVEMENT-; 2955-; QUALITY-; 9063-; EVALUATION-ASSESSMENT; 9020-; UNITED-
KINGDOM; 8119-; ROAD-NETWORK; 2743-; MAINTENANCE-; 3847-; SPECIFICATION-
STANDARD; 0139-; PAVEMENT-MANAGEMENT-SYSTEM; 3037-; INVESTMENT-; 0193-;
MEASUREMENT-; 6136-; EQUIPMENT-; 3674-; METHOD-; 9102-; DEFLECTOGRAPH-; 6187-;
SKIDDING-RESISTANCE; 3031-; PROGRAM-COMPUTER; 8646-
SC: ECONOMICS-AND-ADMINISTRATION (10); EQUIPMENT-AND-MAINTENANCE-
METHODS (61)
AN: 492620
UD: 199811
CN: 9811TR435E
Annex E-64
ANNEX F
FWD CALIBRATION PROCEDURES
September 2002
Annex F-00-1
Preface
This draft standard will probably be under the responsibility of FEHRL. No final agreement has been
achieved on this issue yet (status August 2004), although FEHRL has volunteered to promote acceptance
of the pre-draft standard with CEN. Since CEN has not entered the pre-draft version of the standard in
their system, no number has been assigned to this version of the standard yet..
The standard has been prepared by Task Group 3 'Falling weight deflectometer calibration' of COST
Action 336 and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling weight
deflectometer calibration' of the European Commission. The work was continued in the project 'SpecifiQ'
(Specifications for a harmonised European calibration station for improved FWD measurement of road
quality) funded by the European Commission under the 4FP scheme of 'Standards, Measurements and
Testing' under contract number SMT4-CT98-5518.
This pre-draft European Standard consists of twelve parts. Part 1 covers the field of application of the
FWD and specifies the rules of determination of the periodicity of the calibration of the FWD and its
components and the interaction between the other parts of this standard. Part 1 also describes the terms
and definitions that apply to this standard. Parts 2 to Part 12 included specify the requirements of the
FWDs and their its components.
The following pages of this annex contain the pre-draft standard prepared in CEN layout.
Annex F-00-2
DRAFT
EUROPEAN STANDARD prEN xxxxx-1
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-1:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
Introduction
- collect data to estimate the structural capacity of the pavement structure and the material prop-
erties of the individual pavement layers;
- collect data to estimate the load transfer ability of joints in jointed rigid and composite pave-
ments;
1 Scope
This European Standard applies to the calibration of the falling weight deflectometer, a non-
destructive device used for the structural evaluation of layers of roads and other pavement structures.
This standard covers the interaction of all falling weight deflectometer calibration and calibration
verification procedures as described in the Parts 2-12 of this standard. This standard also describes
when and at what intervals in time the various calibration (verification) activities should be performed.
The standard specifies by whom the activities should be performed.
Distinction is made between calibration activities that should be performed by the FWD user, and
activities that should be performed by the FWD manufacturer and the independent calibration station.
Annex F-01-2
Page 3
prEN xxxxx-1:2002
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
prEN xxxxx-6 Laboratory reference calibration of dismounted falling weight deflectometer deflection
sensors
prEN xxxxx-7 In-situ reference calibration of dismounted falling weight deflectometer deflection
sensors
prEN xxxxx-8 Dynamic reference calibration of falling weight deflectometer load cell
prEN xxxxx-9 Static reference calibration of falling weight deflectometer load cell
For the purpose of this European Standard, the following terms and definitions apply:
3.1
calibration
corrective activity in which the gain factors or calibration factors of the instrument or device under test
are adjusted to match the readings of the instrument or device under test within the specified limits to
those of reference instrumentation or to the mean of a set of readings of similar type of instrumentation
3.2
calibration station
organisation that performs FWD calibration and verification procedures that should not and cannot be
performed by the FWD user for reasons of available instrumentation, expertise and independency
3.3
deflection bowl
envelope curve connecting the peak values of the deflection time histories as a function of the offset of
the deflection sensor
Annex F-01-3
Page 4
prEN xxxxx-1:2002
3.4
deflection sensor
instrument used to measure the surface deflection at a given offset resulting from the load impact of a
falling weight drop
3.5
deflection time history
output signal of the deflection sensor in terms of deflection versus time
3.6
falling weight deflectometer
trailer or van mounted equipment that applies an impact load on a pavement structure by means of a
mass dropping on a set of buffers mounted to a loading plate resting on the pavement surface, and that
measures the resulting vertical deflections of the pavement surface
3.7
falling weight deflectometer user
the operator using the FWD on an operational basis
3.8
load
peak value of the load time history
3.9
load cell
transducer unit used to measure the load imparted onto it
3.10
load time history
output signal of the load cell in terms of load versus time
3.11
loading plate
base plate at the bottom of the falling weight deflectometer that rests on the pavement surface and
transmits the load onto the pavement structure
3.12
long-term repeatability
capacity of a falling weight deflectometer to produce consistent results on a specific test site for each
day of testing under identical climatic conditions over a time interval of months or years
3.13
normalisation
linear interpolation technique used to convert readings to values that would have been obtained under
target conditions
3.14
pulse duration
part of the time history elapsed between the onset of the ascent of the time history and the moment that
the descent returns to the zero level again
Annex F-01-4
Page 5
prEN xxxxx-1:2002
3.15
pulse rise time
part of the time history elapsed between the onset of the ascent of the time history and the moment that
the time history reaches the peak value
3.16
reference calibration
corrective activity in which the response of the instrument or device under test is matched to the out-
put of a reference instrument or device not forming part of the falling weight deflectometer or instru-
ment or device under test
3.17
relative calibration
technique used to fine-adjust within allowed tolerances the response of each similar type of instrument
under test, so that equivalent measurements are obtained. A direct result of this technique is usually
the determination of a set of multipliers necessary to keep the measurements derived from the instru-
ments equivalent.
3.18
short-term repeatability
capacity of a falling weight deflectometer to produce consistent results on a specific test site for a
series of multiple drops without lifting the loading plate
3.19
verification
activity in which either the readings of the instrument or device under test are compared to those of
reference instrumentation or to the mean of a set of readings of similar type of instrumentation, or in
which fine-adjustment within the specified error band is applied
4.1.1 Verify the position and offset of FWD deflection sensors along the raise/lower bar and adjust
mounting to the targeted offsets. Part 2 of this standard presents the procedure to be followed. The
flow chart in Figure 1 presents the procedure graphically.
each month
Part 2
no
OK ? Adjust
yes
End
Annex F-01-5
Page 6
prEN xxxxx-1:2002
4.1.2 This procedure should be conducted at least once per month and each time when the deflection
sensors have to moved or have been moved along the raise/lower bar or more often as considered nec-
essary by the FWD user.
NOTE: FWD testing on airfields on one day and testing on low-volume roads on the other day
with different deflection sensor positions, are typical moments for use of the verification pro-
cedure.
4.1.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
4.2.1 Verify whether the FWD under test is capable of producing consistent results on a specific test
site. Apply a series of multiple drops on a specific test point and determine whether the variation in
load and deflection agrees with the specified limits. Part 3 of this standard presents the procedure to be
followed. The flow chart in Figure 2 presents the procedure graphically. The numbers in the circles
refer to flow chart numbering and other calibration (verification) procedure(s).
each month
Part 3
no
OK ? 5
yes
End
4.2.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
4.3.1 Verify whether the FWD under test is capable of producing consistent results on a specific test
site over the year. Start with verifying whether the deflection sensors are mounted at the targeted off-
sets. Apply a series of multiple drops on a specific test point on a certain day in the year. Compare the
data with deflections predicted by the trend line based on previous testings and temperature recordings
at the same test station. Part 4 of this standard presents the procedure to be followed. The flow chart in
Annex F-01-6
Page 7
prEN xxxxx-1:2002
Figure 3 presents the procedure graphically. The numbers in the circles refer to flow chart numbering
and other calibration (verification) procedure(s) needed for this procedure.
NOTE: The procedure reveals whether unexpected absolute changes of deflection have oc-
curred.
each month
Part 2
no
OK ? Adjust
Part 4
OK ? 8
End
4.3.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
4.4.1 Verify similarity of the response of each of the FWD deflection sensors. If one or more sen-
sors generate deviating results, then the deflection sensor calibration factors should be fine adjusted. If
large adjustments are required, then the sensors should be subject to closer investigation. Part 5 of this
standard presents the procedure to be followed. The flow chart in Figure 4 presents the procedure
graphically.
4.4.2 This procedure should be conducted at least once per period of six months or more often as
considered necessary by the FWD user. When the results do not meet the requirements, then the pro-
cedure should be repeated. Cases of persistent non-compliance invalidates the data collected and ne-
cessitates closer investigation of the instrument under test for removing this problem.
Annex F-01-7
Page 8
prEN xxxxx-1:2002
4.4.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
each 6 months
Part 5
no Check
OK ?
Adjust
yes
End
4.5.1 Reference calibrate the FWD temperature probe to ensure that the probe measures the air and
pavement temperature accurately. Part 11 of this standard presents the procedure to be followed. The
flow chart in Figure 5 presents the procedure graphically.
NOTE: Changes of temperature affect stiffness of the asphalt layers in a pavement structure.
For adequate processing of this effect, reliable stiffness-temperature relationships should be
used, but also accuracy of the pavement temperature recording and the temperature probe it-
self should be beyond any doubt.
each year
Part 11
no Check
OK ? Adjust or
Repair
yes
End
Annex F-01-8
Page 9
prEN xxxxx-1:2002
when another temperature probe calibration procedure is used or when the temperature probe carries a
valid certificate.
4.5.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
4.6.1 Reference calibrate the FWD oedometer or distance measuring instrument to ensure that sub-
sequent comparison with other field data, e.g. results of distress mapping, can be performed with suffi-
cient accuracy. Part 12 of this standard presents the procedure to be followed. The flow chart in Figure
6 presents the procedure graphically.
NOTE: The calibration is also beneficial to additional field experiments. These activities, e.g.
coring can be performed at the same test position visited by the FWD.
each 6 months
Part 12
no
OK ? Adjust
yes
End
4.6.3 This procedure may be used by the FWD user, the FWD manufacturer and the calibration
station.
5.1 Laboratory reference calibration of dismounted FWD deflection sensors for pulse dura-
tion effect
5.1.1 Verify whether the signal processing electronics of the FWD are capable of producing correct
peak values of deflection in cases of varying amplitudes and durations of deflection pulses. The sec-
ond objective is to determine the reference calibration factors. The sensor is subjected to various series
of deflection pulses consisting of multiple combinations of displacement amplitude and deflection
pulse rise time. The output of the sensor is compared to the output of a reference instrument.
Annex F-01-9
Page 10
prEN xxxxx-1:2002
5.1.2 Determine the calibration factors for adjusting deflection sensor output to reference instrumen-
tation. Use the readings around the routinely used pulse duration for this computation.
once
Part 6
normal
no Check
OK ?
Repair
yes
Part 7
no
OK ?
yes
Part 5
no
OK ?
yes
End
Figure 7 - Flow chart #9 for laboratory reference calibration of dismounted deflection sensors
5.1.3 This procedure should be conducted at least once for any new FWD model or release of an
update of a combination of deflection sensor and signal processing unit in case the effect of varying
deflection pulse duration on the accuracy of peak value of deflection is determined. This procedure
should be conducted at least once per two years in case the deflection sensors need to be calibrated.
Cases of persistent non-compliance invalidates the data collected and necessitates closer investigation
of the instrument under test for removing this problem.
5.1.4 This procedure may be used by the FWD manufacturer and the calibration station.
Annex F-01-10
Page 11
prEN xxxxx-1:2002
5.2 Laboratory reference calibration of dismounted FWD deflection sensors for calibration
factor determination
5.2.1 Determine the reference calibration factors of deflection sensors. The sensor is subjected to
various series of deflection pulses consisting of multiple combinations of displacement amplitude and
deflection pulse rise time. The output of the sensor is compared to the output of a reference instrument.
Part 6 (short version) of this standard presents the procedure to be followed. The procedure for relative
calibration of deflection sensors (see Part 5) should be applied as well. The flow chart in Figure 7
presents the procedure graphically. The numbers in the circles refer to flow chart numbering and the
other calibration (verification) procedure(s) needed for this procedure.
each 2 years
either
or
Part 6
Part 7
short
no Check
OK ?
Repair
yes
Part 5
no
OK ?
yes
End
5.2.3 This procedure should be conducted at least once per two years. Cases of persistent non-
compliance invalidates the data collected and necessitates closer investigation of the instrument under
test for removing this problem.
5.2.4 This procedure may be used by the FWD manufacturer and the calibration station.
Annex F-01-11
Page 12
prEN xxxxx-1:2002
5.3.1 Verify whether the FWD deflection sensor produces correct peak values of deflection with the
FWD as load generator. In this procedure the FWD deflection sensor is dismounted and mounted to a
test holder at some distance from the FWD. A test rig with reference instrumentation is positioned
over the holder without having contact with it for (a) recording the reference peak value of deflection
due to the FWD load impact and (b) checking whether the test rig has moved prior to recording the
reference reading. Part 7 of this standard presents the procedure to be followed. The procedure for
relative calibration of deflection sensors (see Part 5) should be applied as well. The flow chart in
Figure 7 presents the procedure graphically. The numbers in the circles refer to flow chart numbering
and the other calibration (verification) procedure(s) needed for this procedure.
NOTE 2: Since the test rig is positioned not too far from the FWD loading plate, vibra-
tions from the FWD load impact will inevitably reach the test rig and have this rig vertically
deflected or rotated. However, with proper design of the test pavement structure, the test rig
and well chosen stiffness and damping characteristics of the rig, vertical movement of the ref-
erence sensor and the deflection sensor under test will be detected later than measuring the
peak value of deflection.
5.3.2 The laboratory reference calibration procedure for determining calibration factors for deflec-
tion sensors (see Part 6 short version) may serve as alternative calibration procedure.
5.3.3 This procedure should be conducted at least once per two years. Cases of persistent non-
compliance invalidates the data collected and necessitates closer investigation of the instrument under
test for removing this problem.
5.3.4 This procedure may be used by the FWD manufacturer and the calibration station.
5.4.1 Apply dynamic reference calibration of the FWD load cell, while mounted to the FWD unit, to
ensure that the load cell measures the peak value of the load pulse accurately. A platform with refer-
ence load cell should be positioned beneath the FWD loading plate for recording the reference load.
Part 8 normal version, of this standard presents the procedure to be followed. The procedure for short-
term repeatability verification (see Part 3) should be applied as well. The flow chart in Figure 9 pre-
sents the procedure graphically. The numbers in the circles refer to flow chart numbering and the other
calibration (verification) procedure(s) needed for this procedure.
5.4.2 This procedure should be conducted at least once per two years or more often as considered
necessary by the FWD user. When the results do not meet the requirements, then the procedure should
be repeated. Cases of persistent non-compliance invalidates the data collected and necessitates closer
investigation of the instrument under test for removing this problem.
5.4.3 This procedure may be used by the FWD manufacturer and the calibration station.
Annex F-01-12
Page 13
prEN xxxxx-1:2002
each 2 years
either
or
Part 8
Part 9
normal
no Check
OK ?
Repair
yes
Part 3
no
OK ?
yes
End
5.5.1 Apply static reference calibration of the FWD load cell, while dismounted from the FWD unit,
to ensure that the load cell measures the peak value of the load pulse accurately. Part 9 of this standard
presents the procedure to be followed. The procedure for short-term repeatability verification (see
Part 3) should be applied as well. The flow chart in Figure 9 presents the procedure graphically. The
numbers in the circles refer to flow chart numbering and the other calibration (verification) proce-
dure(s) needed for this procedure.
5.5.2 This procedure should be conducted at least once per two years when the dynamic reference
calibration of the load cell cannot be performed. When the results do not meet the requirements, then
the procedure should be repeated. Cases of persistent non-compliance invalidates the data collected
and necessitates closer investigation of the instrument under test for removing this problem.
5.5.3 This procedure may also be applied when a new FWD load cell is mounted in the FWD unit.
NOTE: For recalibration of existing FWDs, preference is given to perform the dynamic refer-
ence calibration procedure. In some cases however, no reference platform is at hand, whereas
test rigs might be available for calibration of a dismounted load cell.
Annex F-01-13
Page 14
prEN xxxxx-1:2002
5.5.4 This procedure may be used by the FWD manufacturer and the calibration station.
5.6.1 Determine the FWD correlation factor for an FWD participating in a correlation trial. In this
procedure a group of FWDs is relatively compared against a reference group of FWDs that form part
of the whole group under test. The objective of determination of the correlation factor is to enhance
reproducibility among FWDs. Part 10 of this standard presents the procedure to be followed. The flow
chart in Figure 10 presents the procedure graphically. The numbers in the circles refer to flow chart
numbering and the other calibration (verification) procedure(s) needed for this procedure.
NOTE 1: The procedure was written with the following in mind. Even when all FWD
load cells and deflection sensors have been calibrated satisfactorily, reproducibility among
FWDs may not be achieved. This may be due to differences in thicknesses and properties of
the rubber pads under the loading plate, properties of the rubber buffers, etc, all leading to dif-
ferences in load pulse shape and load pulse rise time. For this reason, use of FWD correlation
factors will enhance reproducibility among FWDs.
NOTE 2: This procedure is primarily intended for use in FWD groups with membership
of FWDs with limited differences in load pulse durations. Use of widely different load pulse
durations will result in deflections for which differences in peak values may be influenced by
differences in pavement structure and structural support of subgrade. More complex conver-
sion techniques are needed under those circumstances to determine appropriate FWD correla-
tion factors for all participating FWD equipment.
each 2 years
10
either
or
Part 8
short
no
OK ? 7
yes
Part 10
no
OK ? 7 9
yes
End
Annex F-01-14
Page 15
prEN xxxxx-1:2002
NOTE: The procedure used is not completely identical to the procedure described in 4.2, due
to the lack of opportunity to analyse repeatability data during the day of testing.
5.6.3 Determination of the load cell calibration factor may form part of the FWD correlation trial.
Part 8 'Short test procedure' of this standard presents the procedure to be followed.
5.6.4 This procedure should be conducted at least once per two years. Cases of non-compliance
invalidate the data collected and necessitates closer investigation of the FWD under test for removing
this problem.
5.6.5 This procedure may be used by the calibration station or any other organisation with sufficient
expertise in this field.
Annex F-01-15
DRAFT
EUROPEAN STANDARD prEN xxxxx-2
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-2:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard covers the positioning and re-positioning of FWD deflection sensors along the
raise/lower bar carrying the sensor holders.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
3 Apparatus
- Cardboard or soft board or any other equivalent system on which sensor positions can leave
their print or mark
4 Procedure
4.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard and each time when the deflection sensors should be repo-
sitioned or have been repositioned along the raise/lower bar.
4.2 Preparation
Annex F-02-2
Page 3
prEN xxxxx-2:2002
4.2.1 Prepare a cardboard plate or soft board plate or equivalent medium of more than 2 m length
and mark this with the requested FWD deflection sensor positions. In the remainder of this standard
only the term cardboard plate will be used. Label the settings. When marking the settings, ensure that
in the actual verification, the cardboard plate will be laid attached to the FWD loading plate.
4.2.2 Park the FWD trailer on a smooth, even floor or pavement. Lower the FWD loading plate. Lay
the cardboard plate described in the previous step at the floor or pavement attached to the FWD load-
ing plate under the raise/lower bar so that the pins under the deflection sensors can punch holes in the
plate.
NOTE: The test procedure can also be performed by using sheets of clean white paper. The
pins of the deflection sensors should be inked prior to testing. When lowering the sensor bar,
the pins will leave a mark on the paper. Other procedures based on the same principle of test-
ing may be used as well.
4.3 Test
Check whether the FWD deflection sensor positions on the FWD raise/lower bar match with the mark-
ings on the cardboard plate. Loosen the FWD deflection sensor holders from their current positions
when needed and reposition them to the requested positions. Tighten the holders again. Positioning
should be performed with an absolute tolerance of ±(4 mm + 0,5 % of the radial distance).
5 Report
b) FWD manufacturer;
g) location of verification;
Annex F-02-3
DRAFT
EUROPEAN STANDARD prEN xxxxx-3
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-3:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies a method to verify whether the falling weight deflectometer under test is capa-
ble of producing consistent peak values of load and deflection on a specific test site by using a series
of successive drops without lifting the loading plate.
The method of verification of short-term repeatability described in this standard may be applied prior
to deflection testing under operational conditions.
The procedure described in this standard may be combined with the long-term repeatability verifica-
tion procedure described in Part 4 of this standard.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-03-2
Page 3
prEN xxxxx-3:2002
sdj standard deviation of normalised deflections measured by Deflection Sensor j over NK drops
(µm)
sF standard deviation of load over all drops (kN)
ui,j unnormalised deflection measured by Deflection Sensor j at Drop i (µm)
4 Apparatus
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
NOTE: Use of this standard is recommended at the first test site of each day of operational
testing to verify whether the equipment produces consistent results.
5.2 Preparation
5.2.1 Position the FWD on a smooth, level, sound asphalt pavement structure with no visible cracks,
where a peak deflection in the order of 250 µm to 600 µm can be produced in the load centre when
using the selected target load level.
5.2.2 Warm up the FWD rubber buffers and condition the test point by repeating a sequence of ten
drops until the loads and deflections that are registered are nearly uniform. The deflections in this
sequence of ten drops should not be showing a steadily increasing or decreasing trend. If liquefaction
or compaction is indicated by the warm-up data, or when the required deflection level cannot be
achieved, relocate the FWD to another pavement.
5.3 Test
Set the drop height and drop mass to generate the selected target load level. Apply two seating drops,
for which no data is recorded, followed by ten replicate drops, for which peak values of load and de-
flection are recorded. Use only these last ten drops in the analysis. Do not raise the FWD loading plate
during the test.
6 Analysis
6.1.1 Normalise all deflections with the use of linear interpolation techniques to a reference load
level that does not depart more than ten percent from the actually applied load (Eq. 1). Determine the
mean deflection of each deflection sensor for the set of ten drops (Eq. 2).
d i, j = u i, j x FR Fi (1)
Annex F-03-3
Page 4
prEN xxxxx-3:2002
NK
d j = ∑ d i, j NK (2)
i =1
The standard deviation of the load recorded in the series of ten drops shall be less than, or equal to two
percent of the mean of the recorded values. If the actual standard deviation exceeds the requirement,
then the repeatability verification should be repeated at another pavement. Failure to satisfy the re-
peatability criteria again, necessitates closer investigation of the deflection sensors and their holders on
the raise/lower bar. The non-compliance invalidates data collected by the FWD under test.
6.1.3 Determine the standard deviation of normalised deflections of each deflection sensor (Eq. 4).
⎡ NK ⎛ NK ⎞ ⎤
2
sd j = ⎢ NK x ∑ d i2, j − ⎜ ∑ d i , j ⎟ ⎥
⎜ ⎟ [ NK x ( NK − 1 ) ] (4)
⎢ m =1 ⎝ m =1 ⎠ ⎥⎦
⎣
The standard deviation of the normalised deflections, recorded in the series of ten drops shall be less
than, or equal to 2 µm in case the mean of normalised deflections is less than, or equal to 40 µm. The
standard deviation of the normalised deflections, recorded in the series of ten drops shall be less than,
or equal to the sum of 1,5 µm and 1,25 % of the mean of the recorded normalised values, in case this
mean is greater than 40 µm. If the actual standard deviation of one or more deflection sensors exceeds
the specified values, then the repeatability verification should be repeated at another pavement. Failure
to satisfy the repeatability criteria again, necessitates closer investigation of the deflection sensors and
their holders on the raise/lower bar. The non-compliance invalidates data collected by the FWD under
test.
7 Report
b) FWD manufacturer;
Annex F-03-4
Page 5
prEN xxxxx-3:2002
m) Analysis results;
n) Declaration whether FWD or FWD component under test complies with the specifications.
Annex F-03-5
DRAFT
EUROPEAN STANDARD prEN xxxxx-4
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-4:2002
Foreword
This draft standard is under the responsibility of cccccccccc and has been prepared by Task Group 3
"Falling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies the determination of the long-term repeatability of the Falling Weight Deflec-
tometer (FWD) on a carefully selected test site producing deflection data with limited annual and/or
seasonal variation. This standard provides a procedure for easy verification of accuracy of FWD de-
flection output at the home base of the FWD user and may be used to investigate whether deflections
depart form the expected outcome.
This standard may also be used to develop a trend pattern of deflections at a selected test site with date
of testing and pavement temperature as predictive input.
The procedure described in this standard may be combined with the short-term repeatability verifica-
tion procedure described in Part 3 of this standard.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-04-2
Page 3
prEN xxxxx-4:2002
4 Apparatus
- Thermometer
- Clock
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
5.2.1 In the first application of this standard, the FWD under test should be equipped, and drop
height and deflection sensor offset should be set as used in normal operation configuration (see Part 2
of this standard). These settings will be termed as default settings. In any future use of this standard,
settings should be identical to the default settings. Keep a record of the default settings.
5.2.2 Select a smooth, level, and sound pavement structure with no visible cracks on which deflec-
tions may be measured which will hardly change with time of the year. This test site should preferably
be shielded from direct solar radiation and other climatic influences. Deflections measured in se-
quences of multiple drops should not be showing a steadily increasing or decreasing trend. If so, relo-
cate the FWD to another pavement.
5.2.3 Mark the position where the loading plate of the FWD rests so that it can be relocated pre-
cisely on the same spot at another day of testing. This may be done by paint, or by marking a small
divot in the pavement with a chisel. Also mark the direction in which the deflection sensor beam
points. Include a description of the test position in the record of the default settings.
Annex F-04-3
Page 4
prEN xxxxx-4:2002
Verify if deflection sensors are mounted at the default offsets. In case of non-compliance, reposition
deflection sensors to the correct offset (use procedure in Part 2 of this standard). Check whether the
FWD is equipped according to the default settings.
5.4 Test
5.4.1 Move the FWD to the test mark and warm up the FWD rubber buffers and condition the test
point by repeating a sequence of ten drops until the loads and deflections that are registered are nearly
uniform.
5.4.2 Use the selected target load level. Apply two seating drops, for which no data is recorded,
followed by ten replicate drops, for which peak values of load and deflection is recorded. Only these
last ten drops will be used in the analysis. Do not raise the FWD loading plate during the test.
5.4.3 Measure the pavement temperature at the mid-depth of the asphalt concrete layer or cement
concrete slab. Register the temperature in °C with one digit placed beyond the decimal point. The
accuracy of the temperature-measuring device should be ±0.5 °C. Record day of the year and clock
time in hours (24 hour system) and minutes (e.g. 14:35).
6 Analysis
Normalise all deflections with the use of linear interpolation techniques to a reference load level that
does not depart more than ten percent from the actually applied load (Eq. 1). Determine the mean de-
flection of each deflection sensor for the set of ten drops (Eq. 9.2).
d i, j = u i, j x FR Fi (1)
NK
d j = ∑ d i, j NK (2)
i =1
6.2.1 If the procedure described in this standard was performed less than or equal to five times for
the selected default settings, then store the normalised centre deflection, pavement temperature or floor
temperature, and the date of testing data for later development of a trend pattern.
6.2.2
6.2.3 If the procedure described in this standard was performed for the sixth time or more, then store
the data specified in 6.2.1 and use the data of all previous tests for the development of a trend line. Use
the data of the test just performed for checking against the trend line (see 6.3).
6.2.4 For development of the trend line, perform a least squares regression, where the dependent
variable is the normalised centre deflection, and the independent variables are the pavement tempera-
ture or floor temperature and the date of testing (Eq. 3). Express the date of testing as the proportion of
the year expired since January 1st. The predicted centre deflection is termed reference deflection.
dR = β 0 + β1T + β 2 sin 2π t + β3 cos 2π t (3)
Annex F-04-4
Page 5
prEN xxxxx-4:2002
6.3 Expectation
For checking the result of the test just performed, enter the date of testing and the pavement tempera-
ture or floor temperature in Eq. 3 to calculate the reference deflection. Compute the deviation of the
measured centre deflection from the reference deflection according (Eq. 4).
d − dR
D= 0 (4)
dR
The recording of a deviation of more than 0,040 invalidates the test results. Identify the source of the
problem, correct it, and repeat the repeatability verification procedure. Persistent presence of variation
in the test data in excess of the trend data, necessitates closer investigation for removing this problem.
NOTE 1: Mounting of a new FWD loading plate and/or FWD load cell may result into
deflections that may depart form the trend pattern found for the 'old' settings, although the
equipment was well calibrated both before and after mounting of the new components.
NOTE 2: Deviations from the expected pattern need not always to be found in the fal-
ling weight deflectometer, but may also be due to inaccurate temperature recording.
7 Report
b) FWD manufacturer;
l) analysis results;
m) declaration whether FWD or FWD component under test depart from the expected pattern
Annex F-04-5
DRAFT
EUROPEAN STANDARD prEN xxxxx-5
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-5:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies a method for the determination of the differences and allowable variation in
deflections recorded by all deflection sensors of the falling weight deflectometer in a series of multiple
drops when all deflection sensors are mounted in a stand, in a way that they all are exposed to the
same deflection. The objective of the test is to verify similarity of the response of each of the deflec-
tion sensors.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-05-2
Page 3
prEN xxxxx-5:2002
4 Apparatus
- Falling weight deflectometer (FWD) including control and signal processing electronics
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
5.2 Preparation
5.2.1 Remove all deflection sensors from their holders on the raise/lower sensor bar of the FWD.
Make sure that the deflection sensors are labelled (e.g., from 1 to 7 in case of seven deflection sensors)
with respect to their normal position on the FWD. The centre position is position number '1'.
5.2.2 Position the deflection sensors in the deflection sensor stand and label the deflection sensor
levels from 'A' to 'G' in case of seven deflection sensors. The top level should be labelled 'A'.
5.2.3 Support the deflection sensor stand in a vertical position. Position it as close to the FWD load-
ing plate as possible where a peak deflection in the order of 250 µm to 600 µm may be produced when
using the selected target load level. Mark the location where the stand rests so that it can be relocated
precisely on the same spot.
5.2.4 Warm up the FWD rubber buffers and condition the test point by repeating a sequence of ten
drops until the loads and deflections that are registered are nearly uniform. The deflections in this
sequence of ten drops should not be showing a steadily increasing or decreasing trend. If liquefaction
or compaction is indicated by the warm-up data, or when the required deflection level cannot be
achieved, relocate the FWD to another pavement.
5.3 Test
5.3.1 Lower the FWD loading plate. Do not raise the loading plate and do not move the FWD dur-
ing testing. This will assure a constant distance between the centre of the load and the base of the de-
flection sensor stand. Press the stand firmly to the pavement or floor.
5.3.2 Set the drop height and drop mass to generate the target load level. Apply two seating drops,
for which no data is recorded, followed by ten replicate drops, for which peak values of load and de-
flection is recorded. Use only the last ten drops in the analysis.
Annex F-05-3
Page 4
prEN xxxxx-5:2002
6 Analysis
6.1.1 Normalise all deflections with the use of linear interpolation techniques to the reference target
load level (Eq. 1).
d i, j = u i, j x FR Fi (1)
6.1.2 Determine the overall mean (Eq. 2) of all normalised deflections for all drops and all deflec-
tion sensors. Determine the maximum-minimum range (Eq. 3) of all normalised deflections per deflec-
tion sensor for all drops. This deflection range should be equal to or less than 4 µm if the system is
working properly and the test was conducted carefully. If the deflection range exceeds 4 µm, the rela-
tive calibration should be repeated.
NK NG
dM = ∑ ∑ d i, j ( NK x NG ) (2)
i =1 j=1
( ) (
rd j = max imum d1, j ; d 2, j ;K; d NK , j − min imum d1, j ; d 2, j ;K; d NK , j ) (3)
6.1.3 Determine the mean deflection of each deflection sensor for the set of ten drops (Eq. 4). Com-
pute the ratio of the overall mean to the deflection sensor mean for each deflection sensor (Eq. 5).
Define this ratio as means ratio.
NK
d j = ∑ d i, j NK (4)
i =1
R j = dM d j (5)
6.1.4 Computed means ratios between 0,995 and 1,005 inclusive are considered to be equivalent to
a ratio of 1,000. In other words, no adjustment is required. When the means ratios for one or more
deflection sensors fall outside the range 0,995 to 1,005, position the deflection sensors at other levels
in the deflection sensor stand, and repeat the relative calibration process. If the position of the deflec-
tion sensor in the stand is the source of means ratios outside the range mentioned above, then the con-
nections in the stand should be checked, and the test should be repeated again. Otherwise, current
calibration factors of the deflection sensor should be adjusted with the computed means ratios as mul-
tiplication factors. Means ratios less than 0,98 or greater than 1,02 are indicative of a damaged deflec-
tion sensor or damaged signal processing unit, which should be replaced or repaired.
7 Report
b) FWD manufacturer;
Annex F-05-4
Page 5
prEN xxxxx-5:2002
g) location of calibration;
l) analysis results;
m) declaration whether FWD or FWD component under test complies with the specifications.
Annex F-05-5
DRAFT
EUROPEAN STANDARD prEN xxxxx-6
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-6:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This protocol covers the determination of the capability of the combination of deflection sensor and
signal processing electronics of producing correct peak values of deflection when deflection sensors
are subject to single shock pulses of varying deflection amplitude and deflection pulse duration.
This protocol covers the determination of the deflection sensor calibration factor for load pulse dura-
tions at operational testing.
2 Referenced documents
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-06-2
Page 3
prEN xxxxx-6:2002
4 Apparatus
- Falling Weight Deflectometer including control and signal processing electronics and deflec-
tion sensors.
- Concrete inertial block or stable solid floor both with background vibrations of less than 1 µm.
- Vibration testing system (shaker + power amplifier). The system must be capable to generate
single deflection impulses of different amplitude up to 2 mm with a shape and duration range
matching FWD generated deflection impulses.
- Spirit level.
- Clamps to mount the deflection sensors to the coil of the shaker. Light metals should be used
to reduce inertial effects as much as possible.
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
5.2 Preparation
5.2.1 Assure proper operation of the reference displacement transducer and all other auxiliary
equipment. If the transducer cannot be made to operate smoothly or accurately, do not continue with
the calibration verification. In the remainder of this protocol the text applies to the use of a displace-
ment transducer for measuring the reference displacement. A servo-accelerometer may be used for
measuring reference displacement as well. The text in the following steps needs to be interpreted ac-
Annex F-06-3
Page 4
prEN xxxxx-6:2002
5.2.2 Assure that the reference calibration system is placed on a concrete block well supported, or
on a solid stable test rig both isolated from vibrations from the surrounding building as much as possi-
ble. Secure the reference displacement transducer in its holder on the reference system, so that it is
near the null point (e.g. zero voltage output). Verify with a spirit level that the reference system is
horizontal and that the transducer is vertical in its holder. If these conditions are not met, adjust the
positions of the system and the holder to attain horizontality and verticality.
5.2.3 Remove the deflection sensors from their holders on the raise/lower bar. Verify that the de-
flection sensors are free of dirt and grime which would adversely affect their seating in the reference
system sensor holder. Clean the base to assure that it is clean.
5.2.4 Place the deflection sensors in the deflection sensor holder of the reference system and clamp
them to the system so that the casing of the deflection sensor cannot move in the system when shock
pulses are to be applied.
5.2.5 Connect the deflection sensors to the FWD control and signal processing electronics and zero
the electronics. Test whether the triggering operates satisfactorily.
5.3.1 Before beginning any calibration work it is necessary that there are no filters and/or smoothing
features in operation in the FWD. Apply single shock pulses of various deflection level and pulse rise
time as tabulated in Table 1. Use three replicates per combination of deflection level and pulse rise
time. Register the deflection time histories of the reference displacement transducer and all deflection
sensors mounted. Check whether the time histories are free of disturbances. Try to remove the source
of these disturbances. If this source cannot be found, do not continue with the calibration.
Table 1 - Deflection level-pulse rise time combinations
NOTE The combination of short pulse time and high deflection level is not used, because
most shaker systems might have difficulty of generating this type of single shock pulse.
5.3.2 Complete the sequence of single shock pulses as shown in Table 1. Pause after each shock
pulse. Register at each shock pulse:
Annex F-06-4
Page 5
prEN xxxxx-6:2002
NOTE 1: Problems may arise in exact determination of the rise times of the shock pulses,
since the onset of the pulse is ill defined. It is recommended to use that point in time as onset,
where the deflection recorded by the reference displacement transducer reaches a value of 2 %
of its peak value on the ascent of the pulse.
NOTE 2: Check whether all test data have been stored successfully. Unsuccessful test data
due to disturbing background vibration should be discarded.
5.4.1 Before beginning any calibration work it is necessary that there are no filters in operation in
the FWD. Set the vibration system to generate single shock pulses with a duration within a band of
25 % around the operationally used load pulse duration of the FWD under test.
5.4.2 Apply single shock pulses of at least ten deflection levels ranging from 50 µm to 600 µm. Use
three replicates per deflection level. Pause after each shock pulse. Register at each shock pulse:
5.4.3 Register the deflection time histories of the reference displacement transducer and all deflec-
tion sensors mounted. Check whether the time histories are free of disturbances. Unsuccessful test data
due to disturbing background vibration should be discarded.
6 Analysis
6.1.1 Compute per valid shock pulse the difference between the peak value of deflection of the
FWD deflection sensor and the reference displacement transducer. Compute the ratio of this difference
to the peak value of deflection generated by the reference transducer (Eq. 1). This ratio will be termed
as 'deviation ratio'.
D i = ( d i − dR i ) dR i (1)
6.1.2 Perform a least squares regression for all the deviation ratios for each deflection sensor, where
the dependent variable is the deviation ratio of the FWD deflection sensor under test, and the inde-
pendent variable is the pulse rise time of the reference displacement transducer expressed in ms (Eq. 2
and Eq. 3). The slope of the regression line, ß1 represents the sensitivity of the system of changes of
pulse rise time on accuracy of the peak value of deflection.
⎡ NS NS NS NS ⎤ ⎡ NS ⎛ NS ⎞ ⎤
2
β 0 = ⎢∑ D i ∑ x i2 − ∑ x i ∑ x i D i ⎥ ⎢ NS x ∑ x i2 − ⎜ ∑ x i ⎟ ⎥
⎜ ⎟ (2)
⎣ i =1 i=1 i =1 i =1 ⎦ ⎢ i =1 ⎝ i=1 ⎠ ⎥⎦
⎣
Annex F-06-5
Page 6
prEN xxxxx-6:2002
⎡ NS NS NS ⎤ ⎡ NS ⎛ NS ⎞
2 ⎤
β1 = ⎢ NS x ∑ x i D i − ∑ x i ∑ D i ⎥ ⎢ NS x ∑ x i2 − ⎜ ∑ x i
⎜
⎟
⎟
⎥ (3)
⎣ i =1 i =1 i =1 ⎦ ⎢ i =1 ⎝ i=1 ⎠ ⎥
⎣ ⎦
6.1.3 The absolute value of the slope of the regression line, ß1 should not be more than
0,0010 m/m/ms. Higher values indicate non-negligible influences of deflection pulse duration on the
peak value of deflection. In case of values in excess of the specified value, FWD deflection sensors
and signal processing electronics should be checked by the FWD manufacturer for removing this prob-
lem. When the FWD manufacturer has corrected the problem, the procedure specified in this standard
should be performed again.
6.1.4 The presence of a standard error of the slope of the regression line (Eq. 4) in excess of 0,0025
invalidates the deflection sensor test results. Identify the source of the problem, correct it, and repeat
the calibration test. Persistent lack of accuracy of the FWD deflection sensor data in excess of speci-
fied values necessitates searching the equipment by the FWD manufacturer for removing this problem.
When the FWD manufacturer has corrected the problem, this calibration verification procedure should
be performed again.
NS NS NS
∑ Di2 - β0 ⋅ ∑ Di - β1 ⋅ ∑ xi Di
NS
sβ1 =
i =1 i =1 i =1
2
⋅ (4)
NS ⎛ NS ⎞ NS - 2
NS ⋅ ∑ xi2 - ⎜⎜ ∑ xi ⎟⎟
i =1 ⎝ i =1 ⎠
6.2.1 Compute per valid shock pulse with a pulse duration within 25 percent around the routinely
used load pulse duration, the ratio of the peak value of the reference displacement transducer to the
peak value of deflection of the FWD deflection sensor (Eq. 5). Compute the mean of the computed
ratio's (Eq. 6). This mean is defined as the 'deflection sensor calibration factor'.
R j = dR j d j (5)
NT
RM = ∑ R j NT (6)
j=1
6.2.2 Standard error of the in-situ deflection sensor calibration factor (Eq. 7) in excess of 0,0020
invalidates the deflection sensor calibration. Identify the source of the problem, correct it, and repeat
the calibration procedure.
⎡ NT ⎛ NT ⎞ ⎤
2
sR mean = ⎢ NT x R 2 − ⎜ R ⎟ ⎥
⎢ ∑ j ⎜∑ j ⎟ ⎥ [ NT x ( NT − 1 ) ] (7)
⎣⎢
j=1 ⎝ j=1 ⎠ ⎦⎥
7 Report
Annex F-06-6
Page 7
prEN xxxxx-6:2002
o) analysis results;
p) declaration whether FWD component under test complies with the specifications.
Annex F-06-7
DRAFT
EUROPEAN STANDARD prEN xxxxx-7
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-7:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This protocol covers the determination of the in-situ deflection sensor calibration factor in a series of
multiple drop heights and drops generated by the FWD while having the sensor dismounted from the
raise/lower bar and mounted to a test holder on a floor or pavement and having the sensor connected to
the FWD.
This test shows whether the release of the mass of the FWD affects accuracy of the deflection readings
and whether the vibrations due to the release have attenuated quickly enough. If the release of the
falling mass and the in this way generated vibrations appear to have a disturbing influence on the re-
corded deflections, then the sensors and the signal processing electronics should be checked by the
FWD manufacturer for removing this problem. Also lack of accuracy of the FWD deflection sensor
data in excess of specified values necessitates a check-up of the equipment by the FWD manufacturer.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-07-2
Page 3
prEN xxxxx-7:2002
4 Apparatus
- Falling Weight Deflectometer including control and signal processing electronics and deflec-
tion sensors.
- Separate test holder for mounting deflection sensor to the test pad or test floor (see Arrow 3 in
Figure 1 and Figure 2).
- Test rig with solid supports and horizontal beam with separate holder at the end or in the cen-
tre of the beam (depending on system used) allowing mounting of reference instrumentation
(see Figure 1 and Figure 2) The design of the test rig must allow for mounting of a deflection
sensor for measuring vertical vibrations of the centre of the test rig. The test rig should be de-
signed to create sufficient phase lag through the supports of the test rig (see Arrow 5 in Figure
1 and Figure 2) to have the centre of the test rig moving later under FWD load impact than the
time for the peak value of deflection recorded at the sensor (see Arrow 4 in Figure 1 and
Figure 2) in the separate holder.
- Reference high accuracy deflection sensor (see Arrow 1 in Figure 1 and Figure 2).
- Clamps to mount the deflection sensors to the test rig and test holder.
Annex F-07-3
Page 4
prEN xxxxx-7:2002
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
Select a smooth, level, sound test pad or test floor with no visible cracks, where a peak deflection
between 250 µm and 600 µm can be produced when using drop height 3.
5.3.1 Position the separate deflection sensor holder within a radius of 1 m from the FWD loading
plate. Remove the deflection sensor from the holder on the raise/lower bar and mount the sensor (see
Arrow 4 in Figure 1 and Figure 2) in the test holder (see Arrow 3 in Figure 1 and Figure 2). The holder
should be properly seated to the test floor or pavement. Connect the deflection sensor to the FWD
control and signal processing electronics. Test whether the triggering operates satisfactorily.
5.3.2 Position the test rig over the holder. Mount the reference displacement transducer (see Arrow
2 in Figure 1 and Figure 2) or equivalent reliable and accurate instrumentation to the test rig. Assure
that the reference instrumentation is properly aligned and in contact with the deflection sensor under
test so that they both will be subject to the same deflection.
5.3.3 Mount a reference deflection sensor to the top of the beam of the test rig (see Arrow 1 in
Figure 1 and Figure 2) to measure the response of the test rig at the position of the instrumentation.
This sensor has to measure the movement and vibration of the test rig generated by the FWD load
impact. The sensor should be properly aligned with the two instruments described in the previous
steps.
5.4 Warming-up
Annex F-07-4
Page 5
prEN xxxxx-7:2002
Before beginning any calibration work it is necessary that there are no data filters or peak smoothing
features in operation in the FWD. Verify that the smoothing option has been turned off. Apply a series
of ten warm-up drops immediately prior to beginning calibration, to assure that the rubber buffers have
been thoroughly warmed up.
5.5.1 Choose three drop heights for the FWD mass. Set the FWD mass to produce peak values of
loads at drop height 2 that are typical for peak loads at operational testing. Set drop height 3 to the full
drop height. The lower drop height 1 should not be less than 20% of the full drop height.
5.5.2 Lower the FWD loading plate and complete the following sequence of drops for recording
peak values of deflection of the sensor under test and the reference instrumentation:
- 5 drops at drop height 3, with a pause after each drop except the last.
Stop after the last drop (FWD loading plate remains down).
Record per drop for each of the three instruments not only the peak value of deflection but also the
deflection time history.
NOTE: It is useful to program six drops at each height, rather than five, so that one can con-
sider a 'spare' in case a drop is missed by the reference system instrumentation. If the first five
drops are successfully recorded, then the data for the sixth drop can be discarded.
6 Analysis
The time history graphs of the test rig deflection sensor and the other two instruments should be
viewed to determine if the test rig has moved, rotated or vibrated prior to capturing the peak value of
deflection of the deflection sensor under test and the reference displacement transducer. The deflection
of the test rig sensor should not exceed 2 µm at the moment in time that the peak values of deflection
for the other two instruments are measured. Deflection levels in excess of this value invalidate the test.
The test should then be repeated. If the position of the test rig is the source of the problem, then the
test rig should be repositioned prior to repeating the test.
6.2.1 Compute per valid drop the ratio of the peak value of deflection of the reference displacement
transducer to the deflection sensor under test (Eq. 1).
R i,k,m = FR i,k ,m Fi,k ,m (1)
Annex F-07-5
Page 6
prEN xxxxx-7:2002
6.2.2 Compute the mean of the calculated ratios and term this mean 'in-situ deflection sensor cali-
bration factor' (Eq. 2).
NH NK
RM = ∑ ∑ R i,k ( NH x NK ) (2)
k =1 i =1
6.2.3 Standard error of the in-situ deflection sensor calibration factor (Eq. 3) in excess of 0,0020
invalidates the in-situ reference deflection sensor calibration. Identify the source of the problem, cor-
rect it, and repeat the calibration procedure.
⎡ NK x NH ⎛ NK x NH ⎞ ⎤
2
sRM = ⎢ NK x NH x
⎢ ∑ R i,k − ⎜ ∑ R i,k ⎟ ⎥⎥
2 ⎜ ⎟ [ NK x NH x ( NK x NH − 1 ) ] (3)
⎣⎢
i =1, k =1 ⎝ i=1, k =1 ⎠ ⎦⎥
6.2.4 A computed in-situ deflection sensor calibration factor between 0,995 and 1,005 inclusive is
considered to be equivalent to a ratio of 1,000. In other words no adjustment is required. An in-situ
deflection sensor calibration factor less than 0,98 or greater than 1,02 is indicative of a damaged de-
flection sensor, which should be replaced or repaired.
7 Report
o) analysis results;
Annex F-07-6
Page 7
prEN xxxxx-7:2002
p) declaration whether the FWD component under test complies with the specifications.
Annex F-07-7
DRAFT
EUROPEAN STANDARD prEN xxxxx-8
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-8:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies a method for the dynamic determination of the load cell calibration factor of
the falling weight deflectometer (FWD) with the load cell mounted to the FWD loading plate. Use of
the procedure ensures that the unit under test measures the peak value of the load pulse accurately.
Procedures are applied to detect any non-linearity in the output of the FWD load cell.
This standard specifies a test and analysis procedure for the determination of the load cell adjustment
factor of the falling weight deflectometer (FWD) during a FWD correlation trial. For this case the
short procedure should be applied whereas in all other cases the normal procedure should be used.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-08-2
Page 3
prEN xxxxx-8:2002
NT number of trials
Ri,k,m ratio of reference load to FWD load of Drop i, Drop height k and Trial m
Ri,m ratio of reference load to FWD load of Drop i irrespective of drop height and Trial m
RM FWD load cell adjustment factor based on data of NT trials
Rm FWD load cell adjustment factor based on data of Trial m
sRm standard error of FWD load adjustment factor of Trial m
sT standard deviation of FWD load adjustment factor over NT trials
4 Apparatus
- Reference load cell platform with a minimum diameter of 300 mm and a maximum diameter
of 450 mm. The platform should consist of either three load cells or a single custom-made
wide based load cell sandwiched between two plates to constitute a stable platform. The upper
plate should be made of light metal. The lower plate may be made of stainless steel
- In case of three reference load cells, the load cells should have been matched
- The reference load cell platform should comply with the following specifications for static
conditions:
- error band encompassing non-linearity, repeatability and hysteresis: ±0,2 kN on the
full scale range if the maximum load of the FWD under test is less than 150 kN, oth-
erwise ±0,4 kN on the full scale range.
- user temperature range of 0 °C to +40 °C
- Spirit level
- Thermometer
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. When the results of application of this standard give reason to further investigation and/or repair
or servicing of components, then this procedure should be repeated after the equipment has been re-
turned from repair or servicing.
Use the test procedures described in 5.3 and 5.5 in case the dynamic load cell calibration forms part
the FWD correlation trial (see Part 10 of this standard). This version is termed the short test procedure
Annex F-08-3
Page 4
prEN xxxxx-8:2002
in this standard. In all other situations the procedures described in 5.3 and 5.4 should be applied. This
version is termed the normal test procedure in this standard.
NOTE: The normal test procedure is appropriate for an accurate dynamic calibration of the
FWD load cell. However, the procedure is too time consuming to be applied on a single day of
testing of several FWD load cells. For this reason, the short version of the calibration proce-
dure was developed for the purpose of the FWD correlation trial.
5.3 Positioning
5.3.1 Position the reference load cell platform so that the lower plate is on a stiff, smooth surface
with a centre deflection between 250 µm and 600 µm using a load level of 50 kN. Verify that there is
no sand or other loose debris under the platform. Use plaster if needed to guarantee proper support.
Verify with a spirit level that the platform is horizontal.
5.3.2 Connect the reference load cell platform to the control and signal processing electronics and
data acquisition system. Position the FWD loading plate above the reference load cell platform, mak-
ing sure that the reference load cell platform and the FWD loading plate are properly aligned. Verify
that the FWD loading plate is properly resting at the platform. All connectors should be inspected and,
if necessary, cleaned and firmly seated.
5.4.1 Before beginning any calibration work it is necessary that there are no data filters in operation
in the FWD. Verify that the smoothing option has been turned off. Apply a series of ten warm-up
drops immediately prior to beginning calibration, to assure that the rubber buffers have been thor-
oughly warmed up.
5.4.2 When the test is performed indoors, sufficient time should be allowed for the FWD to equili-
brate to indoor temperature if it had been outdoors at an air temperature below +5 °C or above +25 °C.
If the test is performed outdoors, ambient temperature may only range from +5 °C to +25 °C.
5.4.3 Choose three drop heights for the FWD mass. Set the FWD mass to produce peak values of
loads at drop height 2 that are typical for loads at operational testing. Set drop height 3 to the full drop
height. The lower drop height 1 should not be less than 20% of the full drop height.
5.4.4 Lower the FWD loading plate and complete the following sequence of drops for recording the
load time histories and peak values of load of the FWD load cell and the reference load cell:
- 5 drops at drop height 3, with a pause after each drop except the last
Stop after the last drop (FWD loading plate remains down).
NOTE: It is useful to program six drops at each height, rather than five, so that one can con-
sider a 'spare' in case a drop is missed by the reference system instrumentation. If the first five
drops are successfully recorded, then the data for the sixth drop can be discarded.
Annex F-08-4
Page 5
prEN xxxxx-8:2002
5.4.5 Label the series of drops described in 5.4.4 as 'Trial 1'. Repeat the procedure described and
label this test as 'Trial 2'.
5.5.1 Before beginning any verification work it is necessary to ensure that data filtering and peak
smoothing of the reference load cell resembles that of the FWD load cell under test as much as possi-
ble (see Part 10 of this standard). Apply a series of ten warm-up drops immediately prior to beginning
the verification action, to assure that the rubber buffers have been thoroughly warmed up.
5.5.2 Copy the settings of the configuration of FWD mass and drop height used in the FWD correla-
tion trial (see Part 10 of this standard).
5.5.3 Lower the FWD loading plate and complete the following sequence of drops for recording the
load time histories and peak values of load of the FWD load cell and the reference load cell:
- 3 seating drops at the target load level (no data recorded), followed by a pause
- 5 drops at the target load level, with a pause after each drop
Stop after the last drop (FWD loading plate remains down).
NOTE: It is useful to program six drops, rather than five, so that one can consider a 'spare' in
case a drop is missed by the reference system instrumentation. If the first five drops are suc-
cessfully recorded, then the data for the sixth drop can be discarded.
6 Analysis
6.1.1 The load time histories should be viewed to determine if noise is of concern before rejecting
the test data. If noise is of concern, identify the source of the problem and report this result.
6.1.2 Compute per drop the ratio of the reference load to the FWD load
R i,k ,m = FR i,k ,m Fi,k ,m (1)
6.1.3 Determine the FWD load cell adjustment factor (Eq. 2) based on the data of Trial 1.
⎛ NH NK ⎞
Rm = ⎜⎜ ∑ ∑ R i,k ,m ⎟⎟ ( NH x NK ) (2)
⎝ k =1 i =1 ⎠
Repeat the procedure for the data of Trial 2. If the difference in results of Eq. 2 of the two trials is less
than, or equal to 0,003, then a third test need not to be performed. Average the results of the first two
trials and term this mean as FWD load cell calibration factor. If the results of the two trials do not
agree within 0,003, then a third test should be performed.
Annex F-08-5
Page 6
prEN xxxxx-8:2002
6.1.4 Determine the standard error of the FWD load cell adjustment factor (Eq. 3). Standard errors
in excess of 0,0020 invalidate the FWD load cell calibration. Identify the source of the problem, cor-
rect it, and repeat the FWD load cell calibration procedure.
⎡ NK x NH ⎛ NK x NH ⎞ ⎤
2
sR m = ⎢ NK x NH x ∑ R i2,m − ⎜ ∑ R i,m ⎟ ⎥
⎜ ⎟ ⎥
[ NK x NH x ( NK x NH − 1 ) ] (3)
⎢ i =1 ⎝ i =1 ⎠ ⎦
⎣
6.1.5 If three trials are performed, compute the mean and the standard deviation of the three results
of the FWD load cell adjustment factor (Eq. 4 and Eq. 5). If the standard deviation is equal to or less
than 0,003, then term the computed mean of the three trials as FWD load cell calibration factor. If the
standard deviation exceeds 0,003, then repeat the entire test. Also verify the standard error from 6.1.4.
NT
RM = ∑ Rm NT (4)
m =1
⎡ NT ⎛ NT ⎞ ⎤
2
sT = ⎢ NT x ∑ R m − ⎜ ∑ R m ⎟ ⎥
2
⎜ ⎟ ⎥ [ NT x ( NT − 1 ) ] (5)
⎢ m =1 ⎝ m =1 ⎠ ⎦
⎣
6.1.6 Compute the new gain factor by multiplying the FWD load cell calibration factor by the cur-
rent gain factor.
6.2.1 The load time histories should be viewed to determine if noise is of concern before rejecting
the test data. If noise is of concern, identify the source of the problem and report this result.
6.2.2 Compute per drop the ratio of the FWD load to the reference load (Eq. 1). Determine the FWD
load cell adjustment factor (Eq. 2) based on the collected data.
7 Report
b) FWD manufacturer;
Annex F-08-6
Page 7
prEN xxxxx-8:2002
j) location of calibration;
n) analysis results;
o) declaration whether the FWD or FWD component under test complies with the specifications.
Annex F-08-7
DRAFT
EUROPEAN STANDARD prEN xxxxx-9
2nd draft
NORME EUROPÉENNE
English version
Français Deutsch
Déflectomètre à masse tombante (à boulet) Fallgewichtsdeflektometer
Calibration Kalibrierung
Partie 9 : Calibration statique de référence Teil 9: Statische Referenzkalibrierung
de la cellule de charge der FWD-Lastzelle
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-9:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies a method for the static determination of the load cell calibration factor of the
falling weight deflectometer (FWD) with the FWD load cell mounted co-axially in series with a refer-
ence load cell. Use of the procedure ensures that the unit under test measures the peak value of the
load pulse accurately. Procedures are applied to detect any non-linearity in the output of the FWD load
cell.
This standard shall be used in case dynamic reference calibration of the FWD load cell (see Part C1 of
this standard) cannot be performed due to lack of reference instrumentation at the location of the FWD
system carrying the load cell.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
prEN xxxxx-8 Dynamic reference calibration of falling weight deflectometer load cell
Annex F-09-2
Page 3
prEN xxxxx-9:2002
4 Apparatus
- Dismounted FWD load cell including control and signal processing electronics
- Rigid test rig for mounting FWD load cell and reference load cell co-axially
- Hydraulic press or similar loading system capable to produce load beyond maximum load
level of FWD load cell
- Reference load cell. This load cell must comply with the following specifications for static
conditions:
- error band encompassing non-linearity, repeatability and hysteresis: ±0,2 kN on the
full scale range if the maximum load of the FWD under test is less than 150 kN, oth-
erwise ±0,4 kN on the full scale range.
- user temperature range of 0 °C to +40 °C
5 Procedure
5.1 Periodicity
This protocol must be conducted as often as specified in the calibration scheme described in Part 1 of
this standard or more frequently as considered necessary by the FWD user. When the results of appli-
cation of this standard give reason to further investigation and/or repair or servicing of components,
then this procedure must be repeated after the equipment has been returned from repair or servicing.
5.2 Preparation
5.2.1 Mount the reference load cell and the FWD load cell under test in a test rig, making sure that
both load cells are properly aligned. Connect both load cells with the signal processing equipment and
data acquisition electronics.
5.2.2 Lower the actuator of the hydraulic press until it just touches the upper load cell and no exter-
nal load is applied to the set of load cells. In case the hydraulic press is mounted under the two load
cells, perform a similar procedure. Zero both load cells.
5.3 Test
5.3.1 Slowly increase the pressure in the hydraulic press until the nominal range of the load cell
under test has been reached. Record the simultaneous readings of the reference load cell and the FWD
load cell for at least ten load levels spread over the whole range.
Annex F-09-3
Page 4
prEN xxxxx-9:2002
5.3.2 Label the series of samples described in 5.3.1 as 'Trial 1'. Repeat the procedure described and
label this test as 'Trial 2'.
6 Analysis
6.1.1 Compute per sample the ratio of the reference load and the FWD load (Eq. 1).
R i,m = FR i,m Fi,m (1)
6.1.2 Determine the standard error of the ratio computed in the previous step (Eq. 2). Standard er-
rors in excess of 0,0020 invalidates the FWD load cell calibration. Identify the source of the problem,
correct it, and repeat the FWD load cell calibration procedure.
⎡ NS ⎛ NS ⎞ ⎤
2
sR m = ⎢ NS x ∑ R i2,m − ⎜ ∑ R i,m ⎟ ⎥
⎜ ⎟ ⎥ [ NS x ( NS − 1 ) ] (2)
⎢ i =1 ⎝ i =1 ⎠ ⎦
⎣
6.1.3 Determine the FWD load cell adjustment factor (Eq. 3) based on the data of Trial 1. Repeat
the procedure for the data of Trial 2. If the difference in results of the two trials is less than, or equal to
0,003, then a third test need not to be performed. Average the results of the first two trials and term
this mean as FWD load cell calibration factor. If the results of the two trials do not agree within 0,003,
then third test should be performed.
NS
R m = ∑ R i,m NS (3)
i =1
6.2.1 If two trials are performed, average the FWD load cell adjustment factors of the first two trials
and term this mean FWD load cell calibration factor.
6.2.2 If three trials are performed, compute the mean and the standard deviation of the three results
of the FWD load cell adjustment factor (Eq. 4 and Eq. 5). If the standard deviation is equal to or less
than 0,003, then term the computed mean of the three trials FWD load cell calibration factor. If the
standard deviation exceeds 0,003, then repeat the entire test. Also verify the standard error from the
previous step.
NT
RM = ∑Rm NT (4)
m =1
⎡ NT ⎛ NT ⎞ ⎤
2
sT = ⎢ NT x ∑ R 2m − ⎜ ∑ R m ⎟ ⎥
⎜ ⎟ ⎥ [ NT x ( NT − 1 ) ] (5)
⎢ m =1 ⎝ m =1 ⎠ ⎦
⎣
6.2.3 Compute the new gain factor by multiplying the FWD load cell calibration factor by the cur-
rent gain factor.
Annex F-09-4
Page 5
prEN xxxxx-9:2002
7 Report
b) FWD manufacturer;
i) Location of calibration;
m) Analysis results;
n) Declaration whether FWD or FWD component under test complies with the specifications.
Annex F-09-5
DRAFT
prEN xxxxx-10
EUROPEAN STANDARD 2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-10:2002
Contents
Contents...................................................................................................................................................2
Foreword .................................................................................................................................................3
1 Scope................................................................................................................................................3
4 Apparatus........................................................................................................................................4
5 Procedure ........................................................................................................................................5
5.1 Periodicity ...................................................................................................................................5
5.2 Organisation ................................................................................................................................5
5.3 Load.............................................................................................................................................5
6 Preparation .....................................................................................................................................5
6.1 Selection of test stations ..............................................................................................................5
6.2 Temperature test point .................................................................................................................6
6.3 FWD test configuration ...............................................................................................................6
6.4 Eligibility.....................................................................................................................................6
7 Procedure ........................................................................................................................................7
7.1 Load calibration verification .......................................................................................................7
7.2 FWD test set-up...........................................................................................................................7
7.3 Pavement temperature .................................................................................................................7
7.4 Deflection testing.........................................................................................................................7
8 Analysis ...........................................................................................................................................8
8.1 Pavement temperature .................................................................................................................8
8.2 Outliers ........................................................................................................................................8
8.3 Correlation factor.........................................................................................................................8
8.4 Repeatability Test......................................................................................................................10
8.5 Verification................................................................................................................................11
9 Precision ........................................................................................................................................11
10 Report............................................................................................................................................11
Annex A.................................................................................................................................................13
Annex F-10-2
Page 3
prEN xxxxx-10:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
This draft European Standard has one normative annex. Annex A of this standard (not Annex A of
COST 336 report) lists the requirements for organisation of a falling weight deflectometer (FWD)
correlation trial.
1 Scope
This standard specifies the determination of the correlation factor for each falling weight deflectometer
(FWD) participating at an FWD correlation trial with participation of at least five participating FWDs
with almost similar load pulse durations.
This standard specifies the determination of the reference deflection bowl on the basis of a group eli-
gible FWDs. Eligibility of FWDs is established by the organising panel of the FWD correlation trial
(see Annex A of this standard) prior to testing.
This standard specifies the determination of the short-term repeatability of FWD load and deflections.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
ISO 5725 Precision of test methods - Determination of repeatability and reproducibility for a
standard test method by inter-laboratory tests
Annex F-10-3
Page 4
prEN xxxxx-10:2002
dHj,k maximum deflection of Deflection Sensor j of eligible set of FWDs at Test Station k (µm)
di,j,k,m normalised deflection measured by Drop i of Deflection Sensor j of FWD m at Test Station k
(µm)
Dj,k,m deviation of normalised deflection of weighted mean deflection
dj,k,m normalised deflection measured by Deflection Sensor j of FWD m averaged over NK drops at
Test Station k (µm)
dj,k,n normalised deflection measured by Deflection Sensor j of FWD n averaged over NK drops at
Test Station k (µm)
dLj,k minimum deflection of Deflection Sensor j of eligible set of FWDs at Test Station k (µm)
dRj,k reference deflection by Deflection Sensor j at Test Station k (µm)
dWj,k weighted mean deflection by Deflection Sensor j at Test Station k (µm)
Fi,k,m peak value of Load at Drop i of FWD m at Test Station k (kN)
FR preselected target load (kN)
i drop label
j deflection sensor label
k test station label
m FWD label
n reference FWD label
NG number of deflection sensors
NK number of drops
NM number of test stations
NN number of eligible FWDs
NR number of reference FWDs
Rj,k,m ratio of reference deflection to deflection measured by FWD m for Deflection Sensor j and
Test Station k
Rj,m deflection sensor correlation factor of FWD m
Rm FWD correlation factor of FWD m
sdj,k,m standard deviation of normalised deflections measured by Deflection Sensor j of FWD m over
NK drops at Test Station k (µm)
sDj,m standard deviation of deviations from weighted mean deflection per deflection sensor and
FWD
sFk,m standard deviation of load of FWD m over NK drops at Test Station k (kN)
sRj,m standard deviation of deviations of ratio's of reference deflection to deflection measured by
FWD m for Deflection Sensor j over all test stations in reproducibility test
ui,j,k,m unnormalised deflection measured by Drop i of Deflection Sensor j of FWD m at Test Sta-
tion k (µm)
Number eligible FWDs from 1 to NN included; number other FWDs from NN+1 to NF included.
4 Apparatus
- Straight edge of 1,2 m for measuring cross-fall, gradient and rut depth;
- Spirit level;
Annex F-10-4
Page 5
prEN xxxxx-10:2002
- Clock
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the organis-
ing panel. The panel should keep records of the test data and analysis results.
5.2 Organisation
5.2.1 An organising panel should initiate the FWD correlation trial and recruit the FWDs. The panel
should also appoint an organisation and executive officer familiar with FWDs for the actual organisa-
tion and preparation of the test, the analysis of the data and reporting of the results. The organising
panel should inspect whether the procedures followed are in conformity with this standard. Annex A
of this standard specifies the personnel requirements for organising an FWD correlation trial.
5.2.2 The organising panel should take a decision which of the participating FWDs should be eligi-
ble for the determination of the reference deflection bowls.
5.3 Load
The organising panel should decide on inclusion of the dynamic load cell calibration verification pro-
cedure in the trial to investigate the accuracy of the FWD load readings. Part 8 (short version) of this
standard contains all the details of the procedure and is dedicated for use in combination with the pro-
cedure described in this part.
6 Preparation
6.1.1 Select at least nine test stations, all with asphalt wearing course or seal coat, prior to the day of
testing in order to test FWD response on various combinations of pavement structure and subgrade
support. Subgrade stiffness should range from weak to stiff, whereas thickness of asphalt layer and
base course should range from thin to thick. The test stations will be used for the determination of the
FWD correlation factor. The test stations should be visited twice during testing. The set of test stations
will be termed 'Set A' in this standard.
Annex F-10-5
Page 6
prEN xxxxx-10:2002
6.1.2 Select three test stations, all with asphalt wearing course or seal coat, prior to the day of test-
ing in order to test FWD response on various combinations of pavement structure and subgrade sup-
port. Subgrade stiffness should range from weak to stiff, whereas thickness of asphalt layer and base
course should range from thin to thick. The test stations will be used for determination of the short-
term repeatability. The stations should be visited only once during testing. The set of test stations will
be termed 'Set B' in this standard.
6.1.3 The choice of the test stations in Sets A and B should be influenced by factors such as conven-
ience of access and safety of operation. The pavement sections should preferably not be too heavily
trafficked. Each test station should have a uniform road surface and preferably not be subject to
sun/shadow effect making it difficult to determine pavement temperatures accurately. No cracking
should be identified at the pavement surface 5 m at either side of each test station; otherwise the de-
flection bowl is liable to be excessively variable among devices. The cross-fall and the gradient of
each test station should not exceed four percent.
6.1.4 Test stations may be chosen either mid-lane or in the nearside wheelpath. In the latter case, the
rutting in the wheelpath should be less than 2 mm deep. Test stations of Set A should be grouped in
test sites with limited spatial distance between the first and last station (preferably not more than
1 km).
6.1.5 Mark each test station with a circle with a radius of 0,4 m.
Drill a narrow hole at the first station of each test site of Set A to facilitate recording of pavement
temperature at the day of testing. The holes must be drilled to the mid-depth of the asphalt layer but
never deeper than 100 mm. Holes may be drilled and temperature may be recorded at each test station.
6.3.1 Each participating FWD should use a loading plate with a diameter of 300 mm and should be
equipped as in routine operation for that specific device. Use at least five deflection sensors. In case of
more than seven deflection sensors, analyse only the data produced by the first series of seven deflec-
tion sensors. Space the deflection sensors at intervals of 300 mm from the load centre, starting at the
load centre, irrespective of the settings commonly used in operational testing. In case of six deflection
sensors, omit the offset 1500 mm. In case of five deflection sensors omit the offsets 1200 mm and
1500 mm.
6.3.2 Position the loading plate of the FWD entirely within the circle mark at each test station. Se-
quence of visiting of the test stations in the reproducibility test should be randomised at least per round
of testing but preferably per test site. Sequence of testing in the second trial should be different from
that used in the first trial.
6.4 Eligibility
Prior to the day of the FWD correlation trial, the organising panel should decide on the eligibility of
the participating FWD equipment with respect to the determination of the reference. Only data col-
lected by the eligible FWDs may serve as input for the determination of the reference deflection bowl.
Only the eligible FWDs that meet the standards set to variation are selected for the actual determina-
tion of the reference deflection. The reference deflection bowl is the mean of the deflections recorded
by the eligible FWDs (see 8.3.6 and 8.3.7).
Annex F-10-6
Page 7
prEN xxxxx-10:2002
7 Procedure
When forming part of the FWD correlation trial, calibration verify the accuracy of the peak value of
the load of each FWD at the first test station using the dynamic reference calibration verification of the
FWD load cell (see the short test procedure specified in Part 8 of this standard).
7.2.1 Verify if the deflection sensors are mounted at the specified offsets. In case on non-
compliance, shift deflection sensors to the correct offset. For the allowed tolerance see Part 2 of this
standard.
7.2.2 Each participating FWD may freely choose the direction in which the deflection sensor beam
should point. The beam may either point forward or backward. Preferably deflection sensor beams
should point in the direction of travel.
7.2.3 Prior to testing and if available, the peak smoothing feature to be applied to the load histories
and deflection histories should be turned on using a low-pass filter with a 60 Hz cut-off frequency.
Testing may also be conducted with the peak smoothing feature turned off. In that case, load and de-
flection signals should be post smoothed before submission of the data for analysis.
NOTE: Various low-pass filters are on the market for peak smoothing purposes. Preferably a
third order Butterworth low-pass filter with a 60 Hz cut-off frequency should be used, al-
though also other types of filters commonly used by the FWD manufacturers may be applied.
All frequencies higher than the cut-off level of 60 Hz should be fully deleted.
7.2.4 Set the target load level at the preselected level at the first test station. When in operational
testing, load targeting features are used, these features should be used in the trial as well.
7.3.1 Each FWD should measure and register the pavement temperature in all pre-drilled holes at
each test site of Set A. Always register the temperature in °C with one digit placed beyond the decimal
point. Clock time of temperature measurement should be recorded in hours (24-hour system) and min-
utes (e.g. 14:35).
7.3.2 The executive officer should conduct independent temperature polling at least three times
between passage of the first and last FWD at each test site of Set A. Temperature should be registered
in °C with one digit placed beyond the decimal point. The accuracy of the temperature recording in-
strument of the executive officer should be ±0.5 °C. Clock time of temperature measurement should be
recorded in hours (24-hour system) and minutes (e.g. 14:35).
7.4.1 Apply one seating drop at each station of Set A for which no data is recorded, followed by
four replicate drops, for which peak values of load and deflection (and whole time histories of the last
drop when possible) is recorded. Only the last four drops will be used in the analysis. Do not raise the
FWD loading plate during the test.
Annex F-10-7
Page 8
prEN xxxxx-10:2002
7.4.2 Apply two seating drops at each station of Set B for which no data is recorded, followed by
ten replicate drops, for which peak values of load and deflection are recorded. Only these last ten
drops will be used in the analysis. Do not raise the FWD loading plate during the test.
7.4.3 Load, deflection and temperature data stored in any format should be copied to diskette and
handed over to the supervisors at the end of the day of testing. These data will serve as backup. Load,
deflection and temperature data stored according to the format specified by the executive officer
should be submitted according to the deadline agreed upon.
8 Analysis
Compute the pattern of change of pavement temperature per drilled hole on the basis of linear interpo-
lation techniques and the independent temperature recordings. These computations are used for esti-
mating pavement temperature for each FWD visit per drilled hole. These temperatures shall supersede
the values recorded by the individual FWDs.
8.2 Outliers
Check the deflection data for outliers. The executive officer should report the outliers with recommen-
dations how to treat these data to the supervisor. The supervisor will take notice of the recommenda-
tions and will decide as considered necessary. The decisions should be reported.
8.3.1 Normalise all deflections per FWD, per test station with the use of linear interpolation tech-
niques to the target load level (Eq. 1). Determine the mean normalised deflection of each deflection
sensor per FWD, per test station for the set of four drops (Eq. 2).
d i, j,k ,m = u i, j,k ,m x FR Fi,k ,m (1)
NK
d j,k ,m = ∑ d i, j,k,m NK (2)
i =1
8.3.2 The centre deflection should not be supplied for analysis at test stations where pavement tem-
perature predicted on the basis of the recordings by the executive officer changed more than 3 °C
among FWDs.
8.3.3 Rank the eligible FWDs per test station and per deflection sensor in terms of mean normalised
deflection. In case of five or more eligible FWDs, remove per test station and deflection sensor the
highest and lowest ranked FWDs temporarily from the population (Eq. 3 and Eq. 4). Use the remain-
ing data for computation of the 'weighted mean deflection' per test station, per deflection sensor
(Eq. 5).
(
dH j,m = max imum d j,k ,1 ; d j,k , 2 ;K; d j,k , NN ) (3)
(
dL j,m = min imum d j,k ,1 ; d j,k , 2 ;K; d j,k , NN ) (4)
Annex F-10-8
Page 9
prEN xxxxx-10:2002
⎡⎛ NN ⎞ ⎤
dW j,k = ⎢ ⎜⎜ ∑ d j,k,m ⎟⎟ − dH j,k − dL j,k ⎥ ( NN − 2) (5)
⎢⎣ ⎝ m =1 ⎠ ⎥⎦
NOTE: In case of four or less eligible FWDs, no FWD data should be removed from the popu-
lation temporarily.
8.3.4 Compute for each eligible FWD (also those temporarily removed in the previous step) per test
station, per deflection sensor the difference between the mean deflection recorded per FWD and the
weighted mean deflection. Determine the ratio of this difference to the weighted mean deflection
(Eq. 6). This ratio will be termed as 'deviation ratio'.
(
D j,k ,m = d j,k ,m − dW j,k dW j,k) (6)
8.3.5 Compute for each eligible FWD per deflection sensor for all test stations the standard devia-
tion of the deviation ratio (Eq. 7). If the standard deviation of the FWD under test exceeds the value of
0,090 for one or more deflection sensors, then this FWD is excluded from the list of eligible FWDs.
⎡ NM ⎛ NM ⎞ ⎤
2
sD j,m = ⎢ NM x ∑ D 2j,k ,m − ⎜ ∑ D j,k ,m ⎟ ⎥
⎜ ⎟ ⎥ [ NM x ( NM − 1 ) ] (7)
⎢ k =1 ⎝ k =1 ⎠ ⎦
⎣
8.3.6 In case of four or less eligible FWDs that comply with the requirement of 8.3.5, compute the
mean deflection per test station per deflection sensor for all eligible FWDs (Eq. 8). Round in presenta-
tions the deflections to the nearest integer, when expressed in µm-units. The deflection bowls obtained
in this way are termed 'reference deflection bowls'.
NR
dR j,k = ∑ d i,k ,n NR (8)
n =1
8.3.7 In case of five or more eligible FWDs that comply with the requirement of 8.3.5, rank the
readings per test station and per deflection sensor in terms of mean normalised deflection. Remove the
highest and lowest ranked readings from the population of each test station and deflection sensor
(Eq. 3 and Eq. 4). Use the remaining data for computation of the mean deflection per test station, per
deflection sensor (Eq. 9). Round in presentations the deflections to the nearest integer, when expressed
in µm-units. The deflection bowls obtained in this way are termed 'reference deflection bowls'.
⎡ ⎛ NR ⎞ ⎤
dR j,k = ⎢ ⎜⎜ ∑ d j,k ,n ⎟⎟ − dH j,k − dL j,k ⎥ ( NR − 2 ) (9)
⎣⎢ ⎝ n =1 ⎠ ⎦⎥
8.3.8 Compute for each FWD (eligible and non-eligible) per test station, per deflection sensor the
ratio of the reference deflection to the mean deflection recorded per FWD (Eq. 10). Determine the
mean (Eq. 11) and standard deviation (Eq. 12) of these ratios per FWD, per deflection sensor. The
means will be termed as 'deflection sensor correlation factor'. Factors less than 0,80 or greater than
1,20 are indicative of an ill-reference calibrated FWD deflection sensor or a damaged deflection sen-
sor, which should be replaced or repaired, and are therefore not accepted.
NOTE: Load pulse durations deviating from those of the reference group may also lead to de-
flection sensor correlation factors outside the tolerated range.
Annex F-10-9
Page 10
prEN xxxxx-10:2002
NM
R j,m = ∑ R j,k ,m NM (11)
k =1
⎡ NM ⎛ NM ⎞ ⎤
2
sR j,m = ⎢ NM x ∑ R j,k ,m − ⎜ ∑ R j,k ,m ⎟ ⎥
2
⎜ ⎟ ⎥ [ NM x ( NM − 1 ) ] (12)
⎢ k =1 ⎝ k =1 ⎠ ⎦
⎣
8.3.9 The presence of a standard deviation of the in-situ deflection sensor harmonisation factor in
excess of 0,090 invalidates the test results for the FWD and deflection sensor under test. Identify the
source of the problem, try to correct it and recompute (if possible) the in-situ deflection sensor har-
monisation factor again. Persistent non-compliance invalidates data collected by the FWD under test
and results into non-compliance with the specifications of the whole procedure described in this proto-
col.
8.3.10 Determine the mean of the deflection sensor correlation factors over the set of deflection sen-
sors used by each FWD (Eq. 13). This mean is defined as 'FWD correlation factor'. Computed factors
between 0,995 and 1,005 inclusive are considered to be equivalent to a factor of 1,000. In other words,
no adjustment is required.
NG
Rm = ∑ R j,m NG (13)
j=1
8.4.1 Normalise all deflections with the use of linear interpolation techniques to the target load level
(Eq. 1). Determine the mean deflection of each deflection sensor for the set of ten drops (Eq. 2).
8.4.2 Determine the standard deviation of all loads (Eq. 14), and the standard deviation of all nor-
malised deflections of each deflection sensor (Eq. 15).
⎡ NK ⎛ NK ⎞ ⎤
2
sFk ,m = ⎢ NK x ∑ Fi2,k ,m − ⎜ ∑ Fi,k ,m ⎟ ⎥
⎜ ⎟ ⎥ [ NK x ( NK − 1 ) ] (14)
⎢ i =1 ⎝ i =1 ⎠ ⎦
⎣
⎡ NK ⎛ NK ⎞ ⎤
2
sd j = ⎢ NK x ∑ d i2, j − ⎜ ∑ d i, j ⎟ ⎥
⎜ ⎟ ⎥ [ NK x ( NK − 1 ) ] (15)
⎢ i =1 ⎝ i =1 ⎠ ⎦
⎣
8.4.3 Per test station, the standard deviation of the load recorded in the series of ten drops shall be
less than, or equal to two percent of the mean of the recorded values. If the actual standard deviation
exceeds the requirement, then the repeatability verification criterion is not complied with at the station
under test.
8.4.4 Per test station, the standard deviation of the normalised deflections, recorded in the series of
ten drops shall be less than, or equal to 2 µm in case the mean of normalised deflections is less than, or
equal to 40 µm. The standard deviation of the normalised deflections, recorded in the series of ten
drops shall be less than, or equal to the sum of 1,5 µm and 1,25 % of the mean of the recorded normal-
ised values, in case this mean is greater than 40 µm. If the actual standard deviation of one or more
deflection sensors exceeds the specified values, then the repeatability verification criterion is not com-
plied with for the station and sensor(s) under test.
Annex F-10-10
Page 11
prEN xxxxx-10:2002
8.4.5 Any FWD complies with the load and deflection repeatability criteria when full compliance
for both load and all deflection sensors is achieved for at least two test stations.
8.5 Verification
8.5.1 The supervisors should compare the data submitted at the day of testing and the data submit-
ted later to the executive officer according to the specified format by means of sampling. In case of
non-similarity, the executive officer will contact the respective FWD user for clarification. Insufficient
clarification may lead to elimination of the analysis data, consequently leading to failing to meet the
specifications set to repeatability and reproducibility.
8.5.2 The supervisors should check by means of sampling whether the analysis procedures followed
and calculations performed by the executive officer are in conformity with the specifications of this
standard. In case of non-compliance, corrective measures should be taken.
9 Precision
The repeatability (= r) and reproducibility (= R) were determined for a group of twelve FWDs partici-
pating in the 1999 FWD correlation trial held in The Netherlands organised by CROW established in
Ede. All FWD equipment used a load pulse duration around the value of 25 ms. Additional data were
analysed for analysis of the error in deflection reading due to inaccuracies in positioning the FWDs
exactly at the centre mark of each test station. The results, which have been interpreted according to
ISO 5725:1986, were as follows:
Table 1 - Repeatability and Reproducibility
Precision Value
Repeatability r 1,1 %
Repeatability r for positioning at selected test station 4,1 %
Reproducibility R 13,3 %
10 Report
Annex F-10-11
Page 12
prEN xxxxx-10:2002
j) analysis results;
Annex F-10-12
Page 13
prEN xxxxx-10:2002
(normative)
The actual planning of the FWD correlation trial should be the task of a panel of experts familiar with
the test method and its application. The planning might also be the task of an organisation endorsing
improvement of accuracy and reproducibility of FWDs under operational conditions, or an organisa-
tion issuing certificates or other documents stating accuracy, repeatability and reproducibility aspects
of FWDs
A.1.2 Supervisor
A staff member of the organising panel should be made responsible for initiation of the FWD correla-
tion trial. He should also review the reported test data and accept the final values of the results of the
trial when found to comply with the procedure described in the standard. He should be accompanied
by a representative of the participating FWDs, also called supervisor.
The actual organisation of the FWD correlation trial should be entrusted to a single laboratory or or-
ganisation familiar with the test method and its application. A member of the staff of that laboratory or
organisation shall take full responsibility; he is called the executive officer. At least one member of the
staff or personnel of the actual organisation should have experience in the analysis of the trial data.
A staff member of each participating FWD should be instructed by the executive officer as to the date
on which the trial will be carried out, and the actual test programme to be followed. He should also be
responsible for reporting and submitting the requested test results.
A.1.5 Operator
A staff member or a team of staff members of each participating FWD selected to perform the actual
test is termed the operator. The operator should be instructed by the contact person as to the date on
which and how the test programme should be carried out.
A.2 Tasks
Annex F-10-13
Page 14
prEN xxxxx-10:2002
- How should the FWDs be recruited and what requirements should they satisfy?
- Which FWDs should be selected as member of the group of eligible FWDs from which the
group of FWDs is formed that establish the reference datum
- What instructions should be issued to the FWD participants concerning the execution of the
test?
- What information should be requested in addition to the numerical test results?
- Who should be appointed to be executive officer?
- Who should be appointed to be supervisors?
- When should the correlation trial be planned?
- Should the dynamic load cell calibration verification procedure form part of the FWD correla-
tion trial?
A.2.2 Supervisor
- to supervise the execution of the test (the supervisors shall not take part in performing the test)
- to take decisions with the assistance of the executive officer on adapting the original test pro-
gramme when difficulties are experienced
- to collect the test results, including any anomalies and difficulties experienced, and to report
them to the executive officer
- to verify whether analysis of the test data is performed according to the specifications in the
standard
- to take decisions on anomalies and problems encountered in the analysis of the test data.
The task of the executive officer is to organise the FWD correlation trial as planned by the organising
panel, in particular:
Annex F-10-14
Page 15
prEN xxxxx-10:2002
- to instruct the operator as to the date of testing and the test settings to be used
- to provide the operator the test programme description.
A.2.5 Operator
- to ensure that the participated FWD complies with the specifications listed in the test pro-
gramme
- to perform the test in accordance to the test programme
- to report any anomalies or difficulties experienced prior to the day of testing or on the day of
testing to the executive officer
- to hand over the raw test data at the end of the day of testing to the supervisor.
Annex F-10-15
DRAFT
EUROPEAN STANDARD prEN xxxxx-11
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-11:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
This standard specifies a method for ensuring that the temperature probe of the falling weight deflec-
tometer (FWD) measures the pavement temperature accurately. The standard describes the determina-
tion of adjustable temperature probe calibration factors in a series of temperature recordings over a
range wider than usually encountered in operational falling weight deflection FWD testing. This stan-
dard applies to thermo-couples and other contact type temperature recording devices and does not
apply to infrared temperature sensors.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
4 Apparatus
Annex F-11-2
Page 3
prEN xxxxx-11:2002
- Reference electronic thermometer with resolution of 0,1 °C and accuracy of ±0,5 °C over the
range from 0 °C to +50 °C
- Bath with contents between 5 dm3 to 10 dm3 filled for more than 50 % with glycol or water
- Stirring device
- Protective cover at container with two holes allowing the temperature recording tips of the
temperature probe under test and reference thermometer to be lowered in the fluid. Cork stub-
bles or equivalent material should be used to reduce heat loss through the holes
- Heater/cooler system with sufficient capacity for increasing the temperature of the fluid
- Clock
5 Procedure
5.1 Periodicity
5.1.1 The procedure described in this standard should be conducted as often as specified in the cali-
bration scheme described in Part 1 of this standard or more frequently as considered necessary by the
FWD user. When the results of application of this standard give reason to further investigation and/or
repair or servicing of components, then this procedure should be repeated after the equipment has been
returned from repair or servicing.
5.1.2 The procedure described in this standard does not need to be performed when the FWD tem-
perature probe carries a valid calibration certificate.
5.2 Preparation
Verify that the heater/cooler system and the stirring device are working properly. Place the bath with
the fluid on a stable platform. Tape the FWD temperature probe under test and the reference ther-
mometer together and lower them into the bath to a depth that their temperature sensitive tips are en-
tirely covered by the fluid. Verify that the FWD temperature probe and the reference electronic ther-
mometer are firmly attached to the protective cover of the bath. Fill the gaps in the cover to prevent
loss of heat.
5.3 Test
5.3.1 Heat/cool the fluid to a bath temperature of 5 °C ±2 °C with the stirring device in operation.
Read and record temperature of FWD temperature probe and reference thermometer simultaneously
both with a resolution provided by the device.
5.3.2 Heat the fluid by a step in the range of 5 °C to 8 °C. Read and record temperature of FWD
temperature probe and reference thermometer again. Repeat this procedure until at last a bath tempera-
ture of 45 °C ±2 °C has been reached.
Annex F-11-3
Page 4
prEN xxxxx-11:2002
6 Analysis
6.1.1 Perform a linear least squares regression for all temperature recordings. The result of this re-
gression will be the coefficients for the slope and the intercept of the linear relation, where the de-
pendent variable is the temperature recorded by the FWD temperature probe under test and the inde-
pendent variable is the temperature recorded by the reference thermometer (Eq. 1 and Eq. 2).
⎡ NS NS ⎤ ⎡ ⎞ ⎤
2
2 ⎛
NS NS NS NS
β 0 = ⎢∑ Ti ∑ TR i − ∑ TR i ∑ (TR i x Ti )⎥
2 ⎢ NS x ∑ TR i − ⎜ ∑ TR i ⎟ ⎥
⎜ ⎟ (1)
⎣ i =1 i=1 i =1 i =1 ⎦ ⎢ i =1 ⎝ i =1 ⎠ ⎥
⎣ ⎦
⎡ NS NS NS ⎤ ⎡ NS ⎛ NS ⎞
2 ⎤
β1 = ⎢ NS x ∑ (TR i x Ti ) − ∑ TR i ∑ Ti ⎥ ⎢ NS x ∑ TR i2 − ⎜ ∑ TR i
⎜
⎟
⎟
⎥ (2)
⎣ i =1 i =1 i =1 ⎦ ⎢ i =1 ⎝ i=1 ⎠ ⎥
⎣ ⎦
6.1.2 The presence of data points that depart more than 1 °C from the regression line invalidates the
test results (Eq. 3). Identify the source of the problem, correct it, and repeat the calibration test. Persis-
tent presence of variation in the FWD temperature probe data in excess of the reference data, necessi-
tates shipping the instrument to the manufacturer for removing this problem. When the manufacturer
has corrected the problem, this calibration procedure should be performed again.
Ti − β 0 − β1 x TR i ≤ 1 °C (3)
6.1.3 Enter the computed values of slope and intercept into the FWD temperature recording system,
or enter data derived from the linear regression when specific types of FWD temperature probe need
this information for accurate temperature recording.
7 Report
e) location of calibration;
j) analysis results;
k) declaration whether FWD component under test complies with the specifications
Annex F-11-4
DRAFT
EUROPEAN STANDARD prEN xxxxx-12
2nd draft
NORME EUROPÉENNE
English version
This draft European Standard is submitted to CEN members for enquiry. It has been drawn up by COST Action
336 "Falling weight deflectometer" of the European Commission and the Consortium "SpecifiQ" a project
funded by the European Commission
Warning: This document is not a European Standard. It is distributed for review and comments. It is subject to
change without notice and shall not be referred to as European Standard.
If this draft becomes a European Standard, CEN members are bound to comply with the CEN/CENELEC Inter-
nal Regulations which stipulate the conditions for giving this European Standard the status of a national standard
without any alteration
Page 2
prEN xxxxx-12:2002
Foreword
This draft standard is under the responsibility of FEHRL and has been prepared by Task Group 3 "Fal-
ling weight deflectometer calibration" of COST Action 336 "Falling weight deflectometer" of the
European Commission and the Consortium 'SpecifiQ'.
The drafting of this standard started in 1996 in the framework of COST Action 336 / TG3 'Falling
weight deflectometer calibration' of the European Commission. The work was continued in the project
"SpecifiQ" ("Specifications for a harmonised European calibration station for improved FWD meas-
urement of road quality") funded by the European Commission under the 4FP scheme of "Standards,
Measurements and Testing" under contract number SMT4-CT98-5518.
1 Scope
Reference calibration of the FWD oedometer or distance measuring instrument is applied to ensure
that subsequent comparison with other field data, e.g. results of distress mapping, ground penetrating
radar, can be performed with sufficient accuracy. The calibration is also beneficial to additional field
experiments. These activities, e.g. coring can be performed at the same test position as visited by the
FWD.
This standard specifies a method for the determination of adjustable FWD oedometer or distance
measuring instrument calibration factors in a test over a stretch of smooth, even asphalt road.
2 Normative references
This European Standard incorporates, by dated and undated reference, provisions for other publica-
tions. These normative references are cited at the appropriate places in the text and the publications are
listed hereafter. For dated references, subsequent amendments to or revision of any of these publica-
tions apply to this European Standard only when incorporated in it by amendment or revision. For
undated references the latest edition of the publication referred to applies (including amendments).
Annex F-12-2
Page 3
prEN xxxxx-12:2002
4 Apparatus
- 25 m long measuring tape or running wheel both with a relative accuracy of at least 0,1 %
5 Procedure
5.1 Periodicity
The procedure described in this standard should be conducted as often as specified in the calibration
scheme described in Part 1 of this standard or more frequently as considered necessary by the FWD
user. This procedure should also be conducted after change of tyre of the car, van or trailer to which
the oedometer is mounted. When the results of application of the procedure give reason to further
investigation and/or repair or servicing of components, this procedure should be repeated after the
instrument has been returned from repair or servicing.
5.2 Preparation
5.2.1 Locate a not too trafficked section of smooth, even asphalt road close to the FWD home base.
The pavement section should be level and straight and have a length of at least 500 m. Mark the be-
ginning of the test section preferably both on the pavement and on an indicator close to the pavement.
Set out a section with a length of at least 500 m with a calibrated reference distance measuring device.
Mark the end of the test section preferably both on the pavement and on an indicator close to the pave-
ment.
5.2.2 Check, adjust when necessary and record tyre pressure of each wheel of the FWD car, van or
trailer (in the remainder of this standardl the term FWD car is used even when FWD van or FWD
trailer would be more appropriate). Tyre pressure should equal to the tyre pressure commonly used
under operational conditions.
5.2.3 Stick a mark at the body of the FWD car as reference for positioning the FWD car at the start-
ing and ending points of the calibration test section.
5.3 Test
5.3.1 Drive the FWD car to the starting point of the test section. Position the FWD car exactly at the
starting point. Set the oedometer at zero. Start driving with gentle acceleration at a speed that does not
exceed 15 km/h.
5.3.2 Reduce speed at the end of the test section with gentle deceleration and pass the end of the
section with creep speed. Register the reading of the oedometer when the mark on the body of the
FWD car is in line with the test section end-mark. Do not back up when the end mark has been passed
without proper registration of the oedometer reading. Redo the test in this case.
5.3.3 Label the reading of the test described in the previous steps as 'Trial 1'. Repeat the procedure
described and label these test as 'Trial 2' and 'Trial 3'.
Annex F-12-3
Page 4
prEN xxxxx-12:2002
6 Analysis
6.1.1 Compute the ratio of test section length over the reading produced by the FWD oedometer
(Eq. 1) based on the data of Trial 1. Repeat the procedure for the data of Trial 2. If the difference in
results of the two trials is less than, or equal to 0,003, then the data of Trial 3 should be discarded.
Average the results of the first two trials. If the results of the first two trials do not agree within 0,003,
then the data of Trial 3 should be processed as well.
R m = L R Lm (1)
6.1.2 If three trials are analysed, then compute the mean and the standard deviation of the three
results (Eq. 2 and Eq. 3). If the standard deviation exceeds 0,003, then repeat the entire procedure.
NT
RM = ∑ R m NT (2)
j=1
⎡ NT ⎛ NT ⎞ ⎤
2
sT = ⎢ NT x ∑ R 2m − ⎜ ∑ R m ⎟ ⎥
⎜ ⎟ ⎥ [ NT x ( NT − 1 ) ] (3)
⎢ m =1 ⎝ m =1 ⎠ ⎦
⎣
6.1.3 The computed mean can be termed as the new FWD oedometer adjustment factor to be multi-
plied with the current factor to obtain the new FWD oedometer calibration factor.
7 Report
b) FWD van, car or trailer registration number to which the oedometer is mounted;
j) location of calibration;
Annex F-12-4
Page 5
prEN xxxxx-12:2002
n) analysis results;
o) declaration whether FWD component under test complies with the specifications.
Annex F-12-5
ANNEX G
OWNERS OF FWD'S
Results of an Inventory
7 March 2000
Annex G-1
G.1 List of FWD owners
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-2
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-3
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-4
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-5
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-6
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-7
Manufacturer
Owner type
Number
Continent Country Owner Type
Europe Sweden National Road and Traffic Research 1991 1 KUAB gov
Institute, Linköping
Europe Sweden RST Sweden AB, Bearing Capacity 1991 1 KUAB gov
Group
Europe Sweden Swedish National Road Administration 1976 1 KUAB gov
Europe Sweden Swedish National Road Administration 1979 2 KUAB gov
Europe Sweden Swedish National Road Administration 1982 1 KUAB gov
Europe Sweden Swedish National Road Administration 1984 1 KUAB gov
Europe Sweden Swedish National Road Administration 1987 1 KUAB gov
Europe Sweden Vägundersökningar AB, Härnösand 1988 1 KUAB gov
Europe Switzerland INFRALAB S.A. FWD 1 Dynatest com
Europe Turkey DHMI - directorate General of the State HWD 1 Carl Bro gov
Airfields Administration
Europe Turkey General Directorate of Highways 1984 1 KUAB gov
Europe Turkey General Directorate of Highways 1992 1 KUAB gov
Europe Yugoslavia Federal Directorate of Supply and 1984 1 KUAB gov
Procurement
Europe Yugoslavia Highway Institute, Belgrade HWD 1 Dynatest com
N. America Canada Dynatest Ltd. HWD 1 Dynatest com
N. America Canada EBA Engineering Consultants Ltd., FWD 2 Dynatest com
Edmonton, Alberta, Canada
N. America Canada J.R.Paine and Associates Ltd. ,Alberta FWD 1 Dynatest com
N. America Canada Ministère des Transports du Quebec FWD 1 Dynatest gov
N. America Canada Province of British Columbia FWD 2 Dynatest gov
N. America Canada Thurber Engineering Ltd. , Alberta FWD 1 Dynatest com
N. America Canada Thurber Engineering Ltd. , Alberta FWD 1 Dynatest gov
(Operated for government)
N. America Canada Trow Ltd. FWD 1 Dynatest com
N. America Dominican Dept. de Estado de ObrasSecr. de Estodo FWD 1 Dynatest gov
Republic de Obras Publicas Y Comm.
N. America Dominican Dept. de Estado de ObrasSecr. de Estodo HWD 1 Dynatest gov
Republic de Obras Publicas Y Comm.
N. America Mexico Geosol, S.Z. De C.V. HWD 1 Dynatest com
N. America Mexico Raul Vicente Orozco y CIA SA de CV HWD 1 Dynatest com
N. America Mexico Raul Vicente Orozco y Cia., S.A. de 1992 1 KUAB gov
C.V., Col. Prado Coapa, C.P.
N. America Puerto Rico Puerto Rico DOT & Public Works FWD 1 Dynatest gov
N. America U.S.A. Alabama Highway Dept. FWD 2 Dynatest gov
N. America U.S.A. Alaska DOT FWD 4 Dynatest gov
N. America U.S.A. ARE Inc. FWD 1 Dynatest com
N. America U.S.A. Arizona DOT FWD 1 Dynatest gov
N. America U.S.A. Arkansas State Highway & Transporta- FWD 2 Dynatest gov
tion Dept.
N. America U.S.A. Austin Testing Engineers, Austin, Texas 1990 1 KUAB com
Annex G-8
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-9
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-10
Manufacturer
Owner type
Number
Continent Country Owner Type
N. America U.S.A. Roy D. McQueen & Assoc., (703)709- JILS-DLS 1 JILS com
2540, Sterling Vashington
N. America U.S.A. SME (Soil & Material Engineers) HWD 1 Dynatest gov
N. America U.S.A. South Carolina Dept of Highways & FWD 2 Dynatest gov
Public Trans.
N. America U.S.A. South Dakota Department of Transporta- FWD 1 Dynatest gov
tion
N. America U.S.A. State of California, DOT, Sacramento 1996 1 KUAB gov
N. America U.S.A. Tennessee DOT FWD 1 Dynatest gov
N. America U.S.A. Terracon, Inc HWD 1 Dynatest com
N. America U.S.A. Texas State Department of Highways FWD 15 Dynatest gov
N. America U.S.A. Texas Transportation Institute FWD 1 Dynatest gov
N. America U.S.A. U.S. Air Force FWD 3 Dynatest gov
N. America U.S.A. U.S. Air Force HWD 2 Dynatest gov
N. America U.S.A. U.S. Army Corps of Engineers (CRREL) FWD 1 Dynatest gov
N. America U.S.A. U.S. Army Corps of Engineers (CRREL) HWD 1 Dynatest gov
N. America U.S.A. U.S. Army Corps of Engineers (WES) FWD 2 Dynatest gov
N. America U.S.A. U.S. Army Corps of Engineers (WES) HWD 1 Dynatest gov
N. America U.S.A. U.S. Navy (LANTDIV) HWD 1 Dynatest gov
N. America U.S.A. U.S. Navy (NCEL) HWD 1 Dynatest gov
N. America U.S.A. U.S. Navy (Southern Division) HWD 1 Dynatest gov
N. America U.S.A. U.S. Navy (Southwest Division) HWD 1 Dynatest gov
N. America U.S.A. University of Kentucky, Clark Graves - JILS-20 1 JILS gov
(606)257-4513
N. America U.S.A. University of Massachusetts FWD 1 Dynatest gov
N. America U.S.A. Utah DOT FWD 1 Dynatest gov
N. America U.S.A. Vermont Agency of Transportation FWD 1 Dynatest gov
N. America U.S.A. Virginia DOT FWD 1 Dynatest gov
N. America U.S.A. Washington DOT FWD 1 Dynatest gov
N. America U.S.A. Wisconsin DOT, Materials Center, 1990 1 KUAB gov
Pavement Management Section
N. America U.S.A. Wyoming State Highway Department, 1991 1 KUAB gov
Cheyenne, WY
S. America Bolivia Servicio Nationale Construcione FWD 1 Carl Bro gov
S. America Brasil Engefoto Engenharia e Aerolevanta- 1995 1 KUAB gov
mento S.A., Jardim Santa Bárbara
S. America Brazil Dynatest do Brasil Ltda. Strata FWD 1 Dynatest com
Engenharia Rodoviaria Ltda.
S. America Brazil Republic of Brazil, Ministry of 1990 1 KUAB gov
Aeronautics
S. America Brazil Consultoria d Projetos Rodoferrovlarios FWD 1 Dynatest com
Ltda.
S. America Brazil COPAVEL FWD 1 Dynatest com
S. America Brazil STE – Canoas RS FWD 1 Dynatest com
Annex G-11
Manufacturer
Owner type
Number
Continent Country Owner Type
Annex G-12
ANNEX H
FWD FOUNDATION TEST
December 2002
Annex H-1
Preface
This annex describes the state-of-the-art of the use of FWD equipment for the determination of the
stiffness of the road base and underlying substructure or the sub-base with underlying substructure. This
test will be labelled 'FWD Foundation Test' in the remainder of this annex. Different testing and analysis
procedures are currently used in Europe. The objective of this annex is to describe the current status and
results, to gain experience and to provide means for harmonisation. The various steps described have not
reached the same degree of maturity yet.
H.1 General
One of the most important sets of parameters in any mechanistic design methodology is the stiffness
modulus of the constituent layers of the pavement structure. Based on these stiffnesses and the layer
thicknesses, critical stresses and strains are computed to investigate whether these values do not exceed
limit values to achieve adequate pavement performance during the design life. It would be useful if test
procedures would be available that could verify in the field whether the assumptions made in the
theoretical design are complied with in practice.
For completed pavement structures, the FWD is a useful tool. Far less experience has been gained with the
use of the FWD on road bases, capping layers and subgrade, although it would be very helpful to collect
field data in terms of layer stiffness during the process of road building. In Quality Control and Quality
Assurance (QC/QA) approaches these results could be useful for validation of the quality of work. A
couple of years ago, the FWD Foundation Test was developed to measure stiffnesses of road bases and
underlying layers during the construction of the road. Unfortunately, different researchers used different
versions of the FWD Foundation Test, obviously leading to different results. This in turn hampers proper
comparison of research findings among various sources. This annex will present some research findings
that are useful towards exchange of experience and knowledge and development of a harmonised testing
and analysis procedure.
Whereas FWDs at completed road pavements measure the deflections in the load centre through a hole in
the loading plate and at various offsets from the load centre, different approaches were used in the FWD
Foundation Test. Figure H-1 shows the four set-ups most used. Set-ups A and B have in common that
deflection is measured by three sensors mounted in a triangular layout at the loading plate. The deflection
for analysis is the average of the three recordings. This approach accounts for some unequal support under
the loading plate. The loading plate of set-up B is a typical FWD loading plate with a centre hole for
mounting the centre deflection sensor for routine FWD testing at completed pavement structures. In set-
up A, a solid plate is used instead. The reason of measuring the deflections at 80 percent of the radius of
the plate will be explained later in this paper. Set-ups C and D have in common that the deflection is
Annex H-2
measured in the load centre. In set-up C, the sensor is mounted onto the solid loading plate, whereas in
set-up D, the deflection sensor is spring mounted through a hole in the centre of the loading plate.
Es =
( ) =
( )
S⋅ 1− ν2 ⋅ a ⋅ σ S⋅ 1− ν2 ⋅ F
d π⋅a ⋅d (H.1)
When using set-ups A or B, the deflections entered in equation (H.1) are the mean of the three recordings
on the loading plate. The majority of the test procedures do not submit the first couple of drops for
analysis due to seating problems. Some procedures use the mean of drops 4, 5 and 6 whereas other
procedures use the mean of drops 11 and 12.
Annex H-3
The magnitude of S depends on the stress distribution under the loading plate or in other words on the
ratio of the rigidity of the plate to that of the underlying medium. Under a very rigid plate deflections are
the same over the plate area. Stress peaks can be observed at the rim of the plate. Under a uniform load
distribution, deflections in the load centre will be higher than those at the rim of the plate. Using a very
flexible plate, not only deflections but also stresses will decrease from the centre towards the rim. The
magnitudes of the factor S are p/2 (= 1.571) for a rigid plate and S = 2 for a uniform load distribution.
Figure H-2 discloses that the shape of the stress distribution under the plate becomes irrelevant when
deflections are measured at a radial distance of 80 percent of the plate radius. This explains the basis of
the deflection testing set-ups A and B of Figure H-1. In practice loading will be neither uniform nor 'rigid'.
This means that it is problematic to compare stiffness data collected by the various test set-ups. Figure H-2
is based on the deflection of a semi-infinite medium. The differences are less pronounced when stiffer top
layers are modelled. The factor S of Equation (H-1) and Poisson's ratio are not constants for each
combination of field testing procedure and testing equipment. Table H-1 lists the values that were
recommended by several users/manufacturers.
0.0
Ratio deflection to centre deflection
-0.2
under uniform loading
-0.4
-0.6
-0.8
-1.0
-1.2
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
r/a
From these the most likely combination of Factor S and Poisson's ratio was derived. The factor 1.500
originates from equations originally issued by Burmister [H-1]. The last column shows that for identical
deflections, the PRIMA100 device generates aggregate stiffness moduli that are some 27 percent higher
than those produced by the FWD equation as is also shown in Figure H-2. This figure demonstrates that
by measuring deflections at a radial distance of 80 percent of the plate radius, the relative rigidity of the
loading plate becomes irrelevant.
Annex H-4
H.4 Effect of stress dependency
Not only the stiffness conversion equations, the rigidity of the loading plate and the test set-ups affect the
end results, so do the dead weight of the equipment and the magnitude of the load pulse. Since granular
road bases have a stress dependent resilient behaviour, the stress ratios are bound to vary from site to site.
Figure H-3 attempts to clarify why the stress ratios will vary.
Annex H-5
deflections measured on the loading plate measure an aggregate stiffness modulus. Figure H-5 exhibits
some test results.
140
100
80
60
40
40 50 60 70 80 90 100
Aggregate stiffness (MPa)
The differences between Figure H-6 and H-7 explain why in the case of Figure H-5, very low values of
aggregate stiffness were found when compared with the subgrade stiffness values.
Annex H-6
700
EpML450
EpMH450
600 EpSL450
EpSH450
E0plate, measured (MPa)
500 EpML300
EpMH300
EpSL300
400 EpSH300
Equality
Regr.plate
300
200
100
0
0 100 200 300 400 500 600 700
E0plate, calculated (MPa)
Figure H-6 Measured and predicted aggregate stiffness modulus based on plate deflection
700
EcML450
EcMH450
600 EcSL450
EcSH450
E0centr, measured (MPa)
500 EcML300
EcMH300
EcSL300
400 EcSH300
Equality
Regr.centr.
300
200
100
0
0 100 200 300 400 500 600 700
E0centr, calculated (MPa)
Figure H-7 Measured and predicted aggregate stiffness modulus based on centre deflection
Another reason why low stiffness are measured at the plate is that the deflection is not only the result of
bending of the pavement structure, but also resulting from shear. The outer deflections are not affected by
vertical translation of the loading plate into the road base. The degree of shear depends on the material
used for the road base, the contact pressure and the degree of compaction. The stiffer the road base is, the
less shear will affect the results.
Annex H-7
H.6 Seating factor
Attempts have been made to use the FWD Foundation Test for gathering information on the rate of seating
in a series of successive drops at a number of test points. It is somewhat speculative to correlate this rate
of seating with the degree of compaction. This section presents the findings of a pilot study.
The rate of seating, also termed seating factor, is defined in equation H.2. This equation determines the
decrease in deflection, measured at drops 1 and 2 and those collected at drops 9 and 10 (Notice that the
pre-drops do not count). The less the deflections have changed with number of drops, the better the
compaction has been realised, at least according to the hypothesis.
d 9 − d10
R SF = 1 −
d1 − d 2 (H.2)
Figure H-8 displays the relationship between the seating factor and the degree of compaction determined
by nuclear test devices [H-3]. From the picture it is clear that the correlation is poor. According to the
hypothesis, the data points should basically be positioned around the diagonal running from the NW
corner to the SE corner of the graph. The scatter that can be observed is simply too great. Closer
investigation disclosed that the seating factor is not only related to the degree of compaction, but also to
the overall aggregate stiffness of the test medium.
Annex H-8
H.7 Conclusions
If the FWD Foundation Test is applied for QC/QA purposes, then it is better to use small diameters than
large diameters, at least if the characteristics of the road base need to be assessed. Plate diameters should
not be smaller than 300 mm because of the size of the largest particles of the aggregate. In this case, the
plate may be carried by only a few stones in the layer. Moving the plate a few centimetres will lead to a
different degree of structural support. These differences lead to high values of spatial variability. This
variation is reduced when a 300 mm plate diameter is used.
No appropriate material modelling, pavement structure modelling and FWD Foundation Test procedure is
available to assess the stiffness of the road base and the underlying medium during construction of the
road. This means that fine-tuning of the design of the asphalt layer is hindered. Also quality control during
work is severely hampered by this lack of knowledge. More emphasis should be paid to this aspect in the
near future.
Although no layer moduli can be inferred from the deflections measured by the FWD Foundation Test,
minimum acceptance values for aggregate stiffness may be developed on the basis of the test results and
structural performance of various types of pavement structures and subgrade.
Variation of the peak value of the load pulse may theoretically give insight in the stress dependency
characteristics of the road base and underlying medium. However, the field experiments analysed did not
provide evidence for this assumption. The variation in stiffness moduli appeared to be very small. The
pavement structure modelling used in the tests presented may have had some influence on this result.
Determination of the rate of decrease of deflection in a series of drops is not a satisfactory indicator for
prediction of the degree of compaction of a road base. The rate of deflection is not only dependent on the
degree of compaction, but also on the aggregate stiffness of the road base.
Current experience with the FWD Foundation Test reveals that various plate diameters, deflection sensor
configurations, equipment weights, peak values of load pulse and conversion equations from deflection to
stiffness modulus are exploited. This variation inevitably leads to a wide range of test output for identical
pavement structure conditions. Harmonisation is urgently requested to allow proper exchange of data and
experience worldwide but certainly Europe wide. In summary, it can be stated that it is of no use to
perform FWD foundation tests if the test procedure and apparatus are not specified. The variation of end
results is simply too wide to serve as a tool for QC/QA for road works. Harmonisation steps will
eventually lead to substantial improvement.
H.8 References
[H-1] Burmister, DM, The theory of stresses and displacements in layered systems and application to the
design of airport runways, Proceedings Highway Research Board. 1943.
[H-2] Groenendijk, J, Van Haasteren, CR, Van Niekerk, AA, Comparison of stiffness moduli of secondary
road base materials under laboratory and in-situ conditions, Proceedings 5th International Sympo-
sium Unbound Aggregates in Roads (UNBAR5), University of Nottingham, 2000.
[H-3] Van Gurp, C., Groenendijk, J. and Beuving, E., Experience with various types of foundation tests,
Proceedings 5th International Symposium Unbound Aggregates in Roads (UNBAR5), University of
Nottingham, 2000.
Annex H-9
ANNEX I
GLOSSARY OF TERMS
Annex I-1
Commercial vehicle
Vehicle intended for transportation of goods. In many countries defined as a vehicle with a maximum
gross mass of more than 3500 kg
Commercial vehicle conversion factor
Factor to convert the combination of actual axle loads of a truck to the equivalent number of standard axle
load repetitions
Crack initiation
The development of diffuse micro damage in critical locations of the pavement into discrete cracks
Crack propagation
The growth of initiated cracks in a pavement over the depth (upwards or downwards) and/or in the
horizontal plane
Crazing
A form of cracking where a network of longitudinal and transversal cracks can be discerned; often called
alligator cracking
Critical asphalt strain
Maximum horizontal relative elastic elongation of the asphalt material due to a design load and/or
temperature
Critical subgrade compression
Maximum vertical elastic relative compression at the top of the subgrade due to a design load
Cumulative sum method
Method for dividing road into homogeneous subsections based on e.g. measured condition variables
Damage ratio
Ratio of the number of equivalent standard axle loads that has been carried and the initial bearing capacity
of the pavement
Deflection
The peak value of the vertical elastic displacement of a given position of the pavement surface under a
dynamic load
Deflection bowl
Line drawn through all deflections measured at various offsets (including zero) from the load centre
Deflection bowl parameter
A mathematical formula combining one or more deflections to reflect the properties of different pavement
layers
Deflection sensor
Instrument developed to measure the deflection at the pavement surface directly or indirectly
Design criterion
Requirement on a parameter, governing a deterioration mode that is considered relevant for the pavement,
as a basis for pavement design and/or evaluation
- Strain or stress criterion
Direct or indirect (via a related performance law) requirement to the stress or strains in critical
locations of the pavement as a basis for pavement evaluation
- Stiffness moduli criterion
Requirement to the stiffness moduli of pavement layers as a basis for pavement evaluation
Annex I-2
- Deflection criterion
Requirement to the deflections or deflection parameters (e.g. surface curvature index) as a basis for
pavement evaluation
Deterioration mode
Phenomenon, such as cracking or deformation, which has a negative influence on relevant pavement
properties such as load distributing capacity or evenness
Drilling core
Cylindrical test specimen cored from the pavement by a hollow drill
Dynamic cone penetrometer (DCP) testing
Test to measure the resistance of a granular material or soil against penetration of a cone driven into the
soil by repetitive droppings of a mass on an anvil
Equivalent layer thickness
Thickness that should be assigned to a layer, if the stiffness modulus of this layer is replaced by another
stiffness modulus, to produce similar stresses and strains in the pavement at the same equivalent depth
Equivalent depth
The depth that a point in a pavement structure would have if the thickness of one or more layers is
changed into another, equivalent, layer thickness
Falling weight deflectometer (FWD)
Trailer or van mounted equipment that applies an impact load on a pavement structure by means of a mass
dropping on a set of buffers mounted on a loading plate resting on the pavement surface, and measures the
resulting deflections of the pavement
Fatigue damage
Deterioration of the integrity of a material as a result of frequently/repeated stresses or strains, each
smaller than the failure or yielding stress or strain of the material
Fatigue resistance
Allowable number of stress or strain repetitions until a specified deterioration of material integrity is
reached
Flexible pavement
Pavement composed from materials that have enough deformability to follow gradual subgrade deforma-
tions without developing significant tensile stresses
Ground penetrating radar (GPR)
Portable of vehicle mounted equipment to measure pavement layer thicknesses by means of measuring the
electric conductivity of the pavement structure as a result of emitting electro-magnetic waves into the
structure
Healing
Ability of asphalt to recover some of the deterioration of its integrity between load changes as a
consequence of kneading effects of the traffic loading and temperature effects, by means of which small
cracks will disappear
Homogeneous subsection
Road section that cannot be further subdivided into sections with significantly different properties
Inlay
New layer which is inserted in a partially milled out existing pavement structure
Annex I-3
Intervention level
Level of deterioration at which a pavement rehabilitation action should be taken
Lateral wander
Transversal distribution of the positions of the wheel loads over a carriageway or traffic lane
Load
Peak value of the load pulse
Load pulse
The part of the force-time graph from the moment the force on the pavement increases from zero until the
moment the force has again decreased to zero
Load pulse duration
Duration of the load pulse in seconds
Load pulse rise time
The period of time between the onset of the FWD load pulse and the moment the peak load is reached
Maintenance section
Part of a road for which the same maintenance measure is proposed
Structural maintenance measure
Measure that appreciably improves the bearing capacity of a road
Milling
Mechanical process in which asphalt layers are partly or totally cut away from the top
Miner calculation
Calculation of the total damage ratio of a pavement as the sum of the damage ratios per service life stage
of the pavement
Nearside wheelpath
Wheelpath lying closest to the pavement edge
Network level
Level of survey comprising an entire system of road sections or a large part thereof
Network level analysis
Evaluation of a group of road sections in order to enable budget estimates, budget allocations to different
parts of the network, selection of candidate projects and project scheduling
Normalisation
Adaptation of measured values or of parameter values derived from these measured values, from the
conditions determined by the system and moment of measurement, to other, usually standard conditions
such as standard loads, design speeds and temperatures
Oedometer
(From the Greek word “odos” = road); an instrument which serves to measure travelled distance
Off-side wheelpath
Wheelpath adjacent to the nearside wheelpath with the same traffic direction
Overlay
Extra asphalt layer laid on top of the existing pavement to strengthen the entire pavement structure
Annex I-4
Pavement evaluation
Assessment of the structural and functional condition of a pavement on the basis of measurements and/or
calculations
Pavement management
Comprehensive body of activities to maintain the safety, driving comfort and quality of the highway
system at an optimal cost-benefit ratio
Pavement performance model
Model that predicts the trend of a specific characteristic of a road pavement over time
Performance law
Model that predicts the resistance of a material against failure as a function of the occurring stresses or
strains, their number of repetitions and sometimes other parameters
- Fatigue law
Performance law that predicts the resistance of a material against failure due to disintegration
- Permanent deformation law
Performance law that predicts the resistance of a material against failure due to permanent
deformation
Permanent deformation
Irreversible deformation of pavement and/or subgrade layers after application of one or more traffic loads
- Primary deformation
Permanent deformation of the asphalt concrete layers
- Secondary deformation
Permanent deformation of unbound pavement or subgrade layers
Poisson’s ratio
Ratio of the transversal relative contraction and longitudinal relative elongation of a material specimen in
one - directional loading conditions
Prioritisation
Determining the priority order of rehabilitation of the pavement sections that have been selected as
requiring maintenance
Project level
Pavement evaluation level dealing with the condition of a selected pavement section and the type and
timing of maintenance
Reference calibration
Type of calibration where the instrument or device under test is compared and if necessary adjusted to the
output of reference instrumentation, which may not form part of the instrument or device under test
Relative calibration
Type of calibration where the instrument or device under test is compared and if necessary adjusted to the
output of similar equipment
Reflection crack
Crack that occurs in a pavement layer as a result of the propagation of an existing crack in the underlying
layer of the pavement
Reliability factor
See safety factor
Annex I-5
Repeatability of FWD, short-term
Ability of FWD to produce consistent results on a specific test site in a sequence of multiple drops,
provided that the pavement is in sound condition, that the support by the subgrade is not too poor and that
temperature or weather factors do not change significantly during the test.
Repeatability of FWD, long-term
Ability of FWD to produce consistent results on a specific test site in tests with time intervals of weeks or
months, under identical temperature, weather and other climatic conditions
Reproducibility of FWD
Ability of FWD to reproduce readings at specific test sites under identical test conditions, even when the
instruments or devices are of different makes and/or types, and/or are operated by different crews
Residual life
Period of time from the moment of pavement evaluation until the moment the pavement no longer meets
specified intervention levels for relevant deterioration modes
Residual structural life
Residual life of the pavement on the basis of structural deterioration modes
Road base
Pavement layer located below the surface and binder layers, fulfilling a load carrying and distributing
function. The road base can be an asphalt concrete road base, a granular road base or a cement bound road
base
Rut depth
Distance between the lowest point of a wheel path and the imaginary straight line drawn between those
parts at the surface of the cross-section immediately on either side of the wheel path
Rutting
The development of concave shaped, continuous permanent deformation of the cross-section of a
pavement; transverse profile of a road in the wheel paths
Safety factor
Factor used in design, residual life and overlay analyses, serving to create some margin between the
objective outcome of the analyses and the results as accepted for design and/or maintenance decisions, for
reasons of the various uncertainties that influence the analyses
Selection
Selection of pavement sections that are in need of maintenance
Semi-rigid pavements
Pavements featuring both (a) bituminous layer(s) and (a) cement bound layer(s) (usually road base layers)
Service life stage
Time period between construction and/or consecutive maintenance stages
Standard axle load
Selected axle load that is utilised in the design and structural evaluation analyses of pavement structures
and which therefore serves as a reference load to which the actual traffic is converted
Statistical sampling
Pavement monitoring approach where a road network is monitored by measuring relatively small, but
statistically representative, samples per network category, using sections of limited length (e.g. 1 km),
randomly distributed over each cluster or link of the network
Annex I-6
Stiffness modulus (Young's modulus)
Ratio between stress and relative elastic deformation of a material caused by this stress
Structural condition
Condition of the pavement with respect to structural deterioration
Structural cracking
Cracking of a pavement which is the result of traffic loading and which has a negative influence on its
load distributing capacity
Structural deterioration
Deterioration of a pavement in terms of structural cracking and/or loss of stiffness, which have a negative
influence on the load distributing capacity, and/or deterioration in terms of permanent deformation
Structural maintenance
Maintenance that serves to improve the residual bearing capacity of a pavement
Subgrade modulus
Stiffness modulus of the subgrade
Surface cracking
Cracking which originates in the topmost asphalt concrete layers
Surface curvature index (SCI)
Difference between the deflection recorded at the centre of the dynamic load and the deflection recorded
at a nearby offset (usually up to 900 mm maximum)
Surface modulus
Stiffness modulus of an equivalent half-space that would produce the same deflection at a given offset to
the load as actually measured on the layered structure at this offset
Systematic monitoring
Pavement monitoring approach where each section of the road network is monitored continuously at
regular length intervals
Test point
Location on the pavement where the centre of the loading plate has been positioned during an FWD
measurement
Traffic lane
Marked part of the carriageway offering sufficient room to any class of motor vehicles with more than
three wheels
Traffic load
Equivalent number of standard axle load repetitions passing per lane per unit of time
Unbound road base
Road base in which the degree of binding of the aggregates does not allow development of significant
tensile stresses under loading
Uncertainty
Lack of knowledge about e.g. the local value, mean value and/or variance of a parameter over a pavement
section
Annex I-7
Variance
Natural variations in the local value of a specific parameter within the same pavement section
Visual inspection
Method of identification of distress visible at the pavement surface for the quantification of type, extent
and degree of distress
Annex I-8
ANNEX J
BIBLIOGRAPHY
Austria
Fuchs, Michael
Neues Meßgerät auf Österreichs Straßen - Fallgewichtsdeflektometer des Geotechnischen Institutes
erleichtert Fahrbahnsanierung (New measurement device on Austria’s roads - Falling Weight
Deflectometer of the Geotechnical Institute makes pavement maintanance easier). Arsenal Aktuell,
Informationen 7.Jg., Bundesversuchs- und Forschungsanstalt Arsenal, Wien 1990.
Croatia
Sršen, M.
Development of road maintenance management system for Croatia. Proceedings of the International Road
and Traffic Conference “Roads and Traffic 2000”, Volume 2/1, Berlin, 6-9 September 1988, FGSV,
PIARC, IRF, pp.37-42.
Sršen, M.
Das HDM III-Modell und seine Bedeutung für das Management der Straßenerhaltung (The HDM-III
model and its significance for the management of road maintenance). Die Asphaltstraße 3/93, Ausgabe
Marz/Mai 1993, pp. 14-20.
Sršen, M.
HDM III Model - Appropriate Pavement Maintenance and Rehabilitation Programs under Conditions
prevailing in Central European Countries. Proceedings of the Third International Conference on Managing
Pavements, San Antonio, Texas, USA, May 21-26, 1994, pp. 246-256. National research Council, TRB,
Washington D.C.
Sršen, M.
Introducing Pavement management System in Croatia. Proceedings of the Fourth International Conference
on Managing Pavements, Durban, South Africa, May 17-21, 1998, Volume 1, pp.. 169-183.
Sršen, M., Juric, M., Keind, V., Lamer, M.
Road Infrastrucure Assets Management Performance. Cross linking Session KL 1, National Report of
Croatia, XXIst World Congress, PIARC, Kuala Lumpur, October 3-9, 1999, pp.1-16.
Czech Republic
Annex J-1
Denmark
Ullidtz, Per
Modelling flexible pavement response and performance. Lyngby Polyteknisk Forlag 1998 - 199 p. -
ISBN: 87-502-0801-2.
Ullidtz, Per
Pavement analysis. Amsterdam Elsevier 1987 - 318 p. - ISBN: 0444428178.
Ertman Larsen, Hans Jørgen
Bassinmåling med faldlod detailundersøgelse baseret på bassinmåling med faldlod (Deflection bowl
measurement with Falling Weight Deflectometer: detailed investigation based upon deflection bowl
measurement with Falling Weight Deflectometer). Vejdirektoratet Statens Vejlaboratorium 1983 - 23 p -
Danish language.
Ullidtz, Per and Ertman Larsen, Hans Jørgen
State-of-the-art : stress, strain and deflection measurements. Vejdirektoratet Statens Vejlaboratorium 1989
- 14 p.
Jansen, Jan M.
Strengthening overlay design as routine procedure, a crucial star in the PMS implementation. Vejdirektor-
atet Statens Vejlaboratorium 1991 - SV notat. 230, p. 49-59.
Schmidt, Bjarne
A study of the relationship between temperature and stiffness of full depth asphalt pavements. Vejdirek-
toratet Statens Vejlaboratorium 1991 - SV notat. 215, 89 p.
Schmidt, Bjarne
Experiences in using falling weight deflectometers as routine equipment. Vejdirektoratet Statens
Vejlaboratorium 1989 - SV notat. 220, 19 p.
Krarup, Jørgen
Bearing capacity and water : part 1: materials, construction and instrumentation. Vejdirektoratet Statens
Vejlaboratorium 1992 - SV notat. 238, 75 p.
Krarup, Jørgen
Bearing capacity and water : part II: measured response. Vejdirektoratet Vejteknisk Institut 1994 - SV
notat. 249, 90 p.
Krarup, Jørgen
Bearing capacity and water : part III: measured pavement performance. Vejdirektoratet Vejteknisk Institut
1995 - VI notat. 253, 163 p.
Finland
Liimatta, L.
Taipumamittausten käyttö tien rakenteellisen kunnon arvioimisessa (The use of deflection measurements
for estimating pavement structural condition). Licentiate thesis, University of Oulu, Department of Civil
Engineering. Oulu, 1996.
Annex J-2
Finnish Road Administration
Tienpinnan taipuman mittaus pudotuspainolaitteella (Pavement deflection measurements using the falling
weight deflectometer). Pori 1998. ISBN 951-726-050-4.
Ruotoistenmäki, A. and Spoof, H.
Tien rakenteellinen kunto (Pavement structural condition index). VTT Yhdyskuntatekniikka (Communi-
ties and Infrastructure), Tutkimusraportti (Research report) 435. Espoo,December 1997.
Ruotoistenmäki, A. and Spoof, H.
Tien rakenteellinen kunto (Pavement structural condition - FWD temperature correction). VTT Yhdyskun-
tatekniikka (Communities and Infrastructure), Tutkimusraportti (Research report) 480. Espoo, January
1999.
Ruotoistenmäki, A.
Determination of layer moduli from falling weight deflectometer measurements. VTT Research notes
1552. Espoo 1994. ISBN 951-38-4486-2.
France
Bretonniere, S.
Etude du deflectometre à boulet (Falling weight deflectometer studies). Bulletin de Liaison des Laboratoi-
res Routiers P. et Ch. , no 2, July - August 1963, pp. 43.1-43.16.
Lepert, Ph. and Simonin, J.M.
Methode LPC d'execution et d'exploitation de la mésure de deformabilite de surface (LPC method to
conduct and run surface deformability tests). Bulletin de Liaison des LPC no 208, March - April 1997, pp.
31-38.
Lepert, Ph., Simonin, J.M., Kobisch, R.
Le FWD : Performances, utilisation en France et en Europe (The performance of the FWD and its use in
France and Europe). Bulletin de Liaison des LPC no 209, May - June 1997, pp. 19-20.
Lepert, Ph. and Joubert, P.
Recent developments in the PMS in France. 4th International Conference on Managing Pavement,
Durban, May 1998, vol. 1.
Germany
Annex J-3
Freund, H.J., Großmann, A., Stocker, M.
Langzeitverhalten von dicken Betondecken auf einer Tragschicht ohne Bindmittel und von Betondecken
auf einer Geotextilzwischenlag (A3 bei Hilden). (Long Term Behaviour of thick Concrete Layers on an
unbound Road Base and of Concrete Layers on a Geotextile Interlayer (A3 near Hilden). FE 08.134,
Karlsruhe, 1999.
Grätz, B.
Einfluß der Temperatur, der Belastungsfrequenz und der Impulskraft beim Falling Weight Deflectometer
(FWD) auf die Größe der effektiven Schicht-E-moduli (Influence of temperature, Loading Frequency and
Pulse Force of the Falling Weight Deflectometer (FWD) on the Value of the Effective Layer Moduli). FE
04.174, Darmstadt, 1999.
Wellner, F.
Untersuchung korrelativer Zusammenhänge zwischen den Auswerte-ergebnissen vier verschiedener
Tragfähigkeitsmeßsysteme (Investigation of the Correlative Dependencies between four Measuring
Systems for Bearing Capacity). FE 04.176, Hannover, 1999.
Hungary
Boromisza, T.
Útpályaszerkezetek dinamikai teherbírásmérésének bevezetése (Implementation of the dynamic pavement
bearing capacity measurement). Közlekedésépítés- és Mélyépítéstudományi Szemle (Civil Engineering
Review) 1993/9.
Kubányi, Z.
Dinamikus teherbíráshoz hõmérsékleti korrekció és évszaki tényezõ meghatározása (Temperature
correction and seasonal factors for dynamic bearing capacity measurements). KTI Rt (Institute for
Transport Sciences Ltd) research report no. 161.1.3.4, 1995.
Adorjányi, K.-Baksay, J.
Determination of Pavement Parameters by Falling Weight Deflectometer Measurements. Rehabilitation of
Roads and Motorways. 7th Budapest International Road Conference, 1996.
Gáspár, L.
Compilation of First Hungarian Network-Level Pavement Management System
Transportation Research Records 1455. Pavement Management Systems. Washington D.C 1994.
Iceland
Jon Helgason, Haraldur Sigursteinsson, Jon Skulason, Rögnvaldur Jonsson, Thorir Ingason.
Bearing Capacity of Roads with thin Pavements. Public Roads Adminstration 1992.
Annex J-4
Haraldur Sigursteinsson, Unnur Helga Kristjánsdóttir
Single Overload on Roads. Public Roads Adminstration, BUSL, 1999.
Jón Skúlason, Valtýr Thorisson
Investigation of base course in streets and roads with thin surface course. Public Roads Adminstration,
BUSL, 1999.
Ireland
----
Guidelines on the Depth of Overlay to be used on Rural Non National Roads. Department of the
Environment and Local Government, May 1999.
Netherlands
CROW
Deflection profile - not a pitfall anymore. CROW Record no.17, Ede, 1998.
CROW
Wegverhardingen op termijn bekeken (Road pavements observed on the long term). Technical Report
SHRP-NL research project, period 1990 - 1995. Ede, The Netherlands, 1996. ISBN 90-6628-207-X.
CROW
Uniformering evaluatiemethodiek cementbetonverhardingen (Specific evaluation procedure for jointed
concrete airfield pavements). CROW Publication no.136, Ede, 1999.
DWW
Handleiding Ontwerp- en Herontwerpsysteem Asfaltbetoverhardingen CARE (Manual for Asphalt
Concrete Pavement Design and Redesign System CARE) DWW-publicatie P-DWW-93-001, DWW,
Delft, 1993.
Gurp, C.A.P.M. van
Characterization of seasonal influences on asphalt pavements with the use of falling weight deflectometers
Ph.D. Thesis, Delft University of Technology , June 1995. ISBN 90-9008036-8.
Norway
Mork, H.
Analyse av lastresponsar for vegkonstruksjonar (Load Response Analysis for Road Pavements). Dr. Eng.
Thesis, NTH, Trondheim 1990, ISBN 82-7119-160-8.
Annex J-5
Mork, H.
Optimal sensorplassering for fallodd og Dynaflect (Optimum Sensor Configuration for Falling Weight
Deflectometer and Dynaflect). SINTEF-rapport STF61 F93005, SINTEF Vegteknikk, Trondheim 1993.
Mork, H.
The Effect of Sensor Configuration on the Backcalculation of Layer Moduli. Proceedings 5th International
Conference on the Bearing Capacity of Roads and Airfields, Trondheim 1998, ISBN 82-519-1346-2.
Portugal
de Almeida, J.R.
Analytical Techniques for the Structural Evaluation of Pavements. Ph.D. Dissertation, University of
Nottingham, 1993.
Antunes, M.L.
Avaliaçao Capacidade de Carga de Pavimentos utilizando ensaios dinamicos (Pavement Bearing Capacity
Evaluation Using Dynamic Non Destructive Tests). Ph.D.Dissertation, Technical Uiversity of Lisbon,
1993.
Romania
IPTANA-SEARCH
Instructiuni tehnice departamentale privind utilizarea deflectometrului Dynatest 8000 FWD pentru
investigarea structurilor rutiere suple si semirigide (Instructions for testing of flexible and semirigid
pavements using Dynatest 8000 FWD). Document nr. 0014/91/45, Bucuresti, Romania, 1992.
Hărătău, S.
Dispozitive de investigare nedistructiva pentru evaluarea structurala a sistemelor rutiere (Nondestructive
investigation equipment for structural evaluation of pavement). Proceedings of OCDE Workshop on Road
Rehabilitation Works. Arad, Romania, 1994.
Hărătău, S., Fodor, G., Capitanu, C., Cioca, S.
Preoccupations of IPTANA-SEARCH in pursuance of overlay systems behavior. Reflective Cracking in
Pavements. Proceedings of the Third RILEM Conference, Maastricht, The Netherlands, 1996.
Hărătău, S. and Capitanu, C.
Metoda moderna de diagnosticare a capacitatii portante pe baza masurarilor efectuate in situ (Modern
method for bearing capacity diagnosis based on in-situ testing). Proceedings of National Conference for
In-situ Behavior of Constructions, Buzias, Romania, 1998.
Slovenia
Hoçevar, A.
Use of Falling Weight Deflectometer at Network Level. Zbornik referatov, 4. slovenski kongres o cestah
in prometu (4th Slovenian Congress on Roads and Transport), Portoroz 1998 ( ISBN 961-90496-5-9).
Annex J-6
Hoçevar, A.
Slovenian Experience in Using the Falling Weight Deflectometer for Determining the Structural
Adequacy of Roads. Zbornik referatov, 4. slovenski kongres o cestah in prometu (4th Slovenian Congress
on Roads and Transport), Portoroz 1998 ( ISBN 961-90496-5-9).
Spain
CEDEX
Norma de Ensayo del Centro de Estudios de Carreteras NLT-338/98. Medida de Deflexiones con el
Deflectómetro de Impacto. (Road Research Center Test Standard NLT-338/98. Measurement of
Deflections with the Falling Weight Deflectometer). Centro de Estudios y Experimentación de Obras
Públicas (CEDEX). Madrid 1998
Sweden
Lenngren, C.A.
Backcalculation of Modulus on Loading Time Dependent Clays. Proceedings from the Fifth International
Conference on the Bearing Capacity of Roads and Airfields, vol.3, pp.1827-1836. Trondheim, Norway, 6-
8 July 1998.
Burton, W.
Analysis of Falling Weight Deflectometer Tests for Various Subgrade Moduli - Case Study and
Numerical Simulation. Department of Geotechnical Engineering, Chalmers University of Technology,
Göteborg. Report B 1998:9
Swedish National Road Administration
Deflektionsmätning vid provbelasting med fallviktsapparat (Deflection measurement under test load with
falling weight deflectometer). Metodbeskrivning 112:1998, Vägverket, Publikation 1998:80.
Djärf, Lennart
Evaluation of in situ soil moduli based on falling weight deflectometer (FWD) loading. Flexible
pavements. European symposium Euroflex 1993. Lisbon, Portugal, 20–22 September, 1999.
Jansson, Håkan
A simple structural index based on FWD measurement. Statens Väg- och Transportforskningsinstitut. VTI
särtryck 223. Linköping. 1994.
Tholén, O.
Falling weight deflectometer: A device for bearing capacity measurement: Properties and performance.
Stockholm. Kungliga Tekniska Högskolan. Vägbyggnad. Bulletin 1990:1.
Tholén, O.
Prov med olika stöttider vid fallviktsbelastning (Experiment with different load rtaising times at falling
weight testing). Stockholm. Kungliga Tekniska Högskolan. Vägbyggnad. Bulletin 1980:7–8.
Tholén, O.
Bärighetsmätning med fjädrande fallvikt vid VTI 1971–1973 (Bearing capacity measurement with falling
weight at VTI 1971-1973.) Statens Väg och Trafikinstitut. Internrapport nr 173. Stockholm. 1974.
Annex J-7
Jansson, Håkan and Wiman, Leif G.
Pavement analysis based on measured in-depth deflection data. Statens Väg- och Transport-
forskningsinstitut. VTI särtryck 222. Linköping. 1994.
Jansson, Håkan and Wiman, Leif G.
Mätta och beräknade deformationer i vägen vid fallviktsmätning: En jämförelse på tre instrumenterade
vägar (Measured and calculated deformations of roads under falling weight measurement; an investigation
on three instrumented roads). Statens Väg- och Transportforskningsinstitut. VTI meddelande 738.
Linköping. 1994.
International
World Bank
HDM III Model - Appropriate Pavement Maintenance and Rehabilitation Programs under Conditions
prevailing in Central European Countries. Proceedings of the Third International Conference on Managing
Pavements, San Antonio, Texas, USA, May 21-26, 1994, pp. 246-256. National research Council, TRB,
Washington D.C.
Annex J-8