European Specification On Weigh-In-Motion of Road Vehicles (COST323)

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/341911554

EUROPEAN SPECIFICATION ON WEIGH-IN-MOTION OF ROAD VEHICLES


(COST323)

Conference Paper · September 1998

CITATIONS READS

4 809

2 authors:

Basima Jacob Eugene J. OBrien


University of Duhok University College Dublin
23 PUBLICATIONS   130 CITATIONS    396 PUBLICATIONS   5,719 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mobile Automated Rovers Fly-By (MARS-FLY) for Bridge Network Resiliency (Funding Agency: SFI US-Ireland R&D Partnership at NSF) View project

TRUSS (Training in Reducing Uncertainty in Structural Safety, 2015-2019, http://trussITN.eu ) View project

All content following this page was uploaded by Eugene J. OBrien on 04 June 2020.

The user has requested enhancement of the downloaded file.


EUROPEAN SPECIFICATION ON WEIGH-IN-MOTION OF ROAD
VEHICLES (COST323)
A graduate of Ecole Polytechnique Vice-chairperson of the European
and Ecole Nationale des Ponts et research action COST 323 and
Chaussées, started work at the chairperson of the Bridge
SETRA. With the LCPC, in the Applications subcommittee. Prof-
Scientific and Technical Division, is essor & Head of Civil Engineering at
Chairman of the Management University College Dublin, Ireland.
Committee COST-323 and co-
ordinator of the project ‘WAVE’.

Bernard JACOB Eugene J. O’BRIEN


Chair of COST323, Laboratoire Central des Vice-chair of COST323
Ponts et Chaussées, Paris, France University College Dublin, Ireland

Abstract
The European WIM specification, prepared by the COST323 management committee, gives an
indication of what WIM system accuracy might be achievable from sites with particular
characteristics, and what accuracy might be acceptable for various needs. Calibration and testing
procedure for WIM systems are considered. Accuracy classes are defined on the basis of the width of the
confidence interval within which the measured results lie. For each class, confidence interval widths are
specified for gross weights, weights of axles, etc., and the required level of confidence is fixed with
respect to the test conditions and the number of test runs. A simplified requirement with standard test
plans is proposed for common acceptance tests.
Keywords: Specification, standard, load, pavement conditions, vehicle, gross weight, axle load,
Weigh-In-Motion (WIM), sensor, system, calibration.

Résumé
Les spécifications européennes du pesage en marche, préparées par le comité de gestion COST323,
indiquent la précision accessible aux systèmes de pesage selon les sites et leurs caractéristiques, et
celle requise selon les applications visées. Les méthodes d'étalonnage et de contrôle des systèmes sont
décrites. Des classes de précision sont définies par la largeur de l'intervalle de confiance requis pour les
mesures. Pour chaque classe, les largeurs d'intervalle de confiance sont spécifiées pour les poids totaux,
d'essieux, etc., ainsi que les niveaux de confiance exigés en fonction des conditions d'essai et du nombre
de mesures. Des règles simplifiées, comprenant des plans d'expérience standards, sont proposées pour les
essais de réception ordinaires.
Keywords: spécifications, norme, charge, état de chaussée, véhicule, poids total, poids d'essieu,
pesage en marche, capteur, système, étalonnage.

Zusammenfassung
Die europäischen WIM-Spezifikationen sind im Rahmen des Management-Komittees der COST-323
Aktion dargelegt. Diese Spezifikationen ermöglichen Auskunft bezüglich der erreichbare Genauigkeit
der WIM-Systemen gemäss deren Stellen, Anordnung oder vorgesehenen Anwendungen. Die
Eichungs- und Prüfungsmethoden für WIM-Geräte sind beschrieben. Die, für die Messungen
erwartete Vertrauensbereichbreite, definiert die Genauigkeitsklassen. Der Vertauensbereich is gegeben
in jeder Genauigkeitsklasse, für die Gesamtlast, die Achslast, usw. Die, für der Messvorgang oder
Messanzahl erförderlichen Vertrauensniveau, sind auch festgesetzt. Vereinfachte Vorschrifte, sind
angeboten; Sie enthalten standardisierte Prüfverfahren für gewöhnliche Rezeptursproben.
Schlüsselwörter: Spezifikation, Norm, Last, Strassenverhältnisse, Fahrzeug, Gesamtgewicht,
Achslast, Wägen während der Fahrt (WIM), Sensor, Gerät, Eichung.
1. Introduction

At present WIM systems exist in 20 countries in Europe and there are in excess of 350
systems. The reasons are the many applications for which WIM data can be used. There are
also in excess of 15 WIM system/sensor European manufacturers.

In order to facilitate the relationships between WIM users and suppliers, to clarify the real
performance and levels of accuracy of WIM systems, and to provide a strong basis for any
call for tender or acceptance test, it has been decided to develop a harmonised European WIM
Specification. The context of European legislation harmonisation, standardisation and
codification (such as with the Eurocodes) is particularly favourable for such work at present.
This task addressed one of the most important objective of the COST323 action.

This technical specification currently fills the lack of WIM standard in Europe, where only a
few national specifications (METT-LCPC 1993, VIEA 1994, NWML 1995) were in use. The
work started with the analysis of existing and emerging specifications (ASTM 1994, OIML
1996). The specification was drafted between May 1996 and January 1997. The content was
discussed with European WIM manufacturers in January 1997. A few comments were
received and incorporated, while the document was complemented following a proposal from
the manufacturers’ with an appendix entitled: ‘Simplified Requirements’. The latter is
designed to meet the needs of common acceptance tests for most of the customers. The latest
draft (COST323, 1997b) was submitted for comments to a large panel of users during the
Summer 1997, which welcomed it, and then has been widely distributed throughout Europe
and elsewhere.

2. Structure of the specification

The specification consists of the main text, ‘Detailed Specification’, and three appendices:
(1) Simplified requirements, (2) Statistical issues on accuracy of WIM systems, and (3)
Implementation and software. The appendix 3 is still incomplete.

The detailed specification is a comprehensive document. There are a number of mandatory


clauses but considerable flexibility is allowed in terms of the testing of systems. In addition,
there are informative clauses which are intended to provide the user with some background
information about WIM and about the reasons for the mandatory clauses. In the present draft,
the two types of clauses are mixed but the informative clauses are marked with a bar in the
margin.

Appendix 1 contains a simplified set of rules for the common calibration and testing of WIM
systems. Five standard test plans are proposed, which may be applied depending on the means
(number and type of test vehicles) and time available. That allows to classify a system into
one of the accuracy classes. Moreover charts are provided to obviate the necessity for long
and complex calculations during the data analysis; thus the analysis can be performed,
possibly on site, using a simple pocket calculator.

The main physical, metrological and statistical background for the specification are presented
in appendix 2, as a separate document, also described by (Jacob, 1997).

The specification gives a detailed description of the means by which a WIM system can be
calibrated, tested and classified in an accuracy class.

172
3. User and Performance Requirements

According to (COST323 1997a), three main types of application with increasing accuracy
levels were identified for road authorities, decision makers and WIM users:
• Statistics: economical, geographical and technical studies of freight transport and traffic
evaluation. Accuracy classes D(25) or D+(20) seem to be suitable for this application.
In some cases lower classes such as E(30) and below, may be accepted for a rough
evaluation, using simple and inexpensive devices or on poor pavements.
• Infrastructure and preselection: detailed traffic analysis, road and bridge design and
maintenance, pavement aggressiveness and fatigue studies, code calibration,
preselection for enforcement. Classes C(15) or B(10) are required.
• Enforcement of legal weight limits: if the legislation allows the use of WIM (Low speed
WIM or eventually High speed WIM), Class A(5) is likely to be required. It is possible
that B+(7) may also be accepted if the vehicle exceeds the legal limit by a sufficient
amount.
The definitions of the accuracy classes are given in section 6 below. These requirements are
only given as an indication and are based on the current state-of-the art. Clients may specify
whatever they need within the framework of the specification.

4. Choice of WIM site

The characteristics of a WIM site can significantly affect the accuracy of results. Road
roughness, in particular, results in an excitation of the dynamics of passing trucks. Increased
dynamic variation in applied force clearly tends to result in a reduction in the accuracy of
static weight estimates.

In addition to accuracy, site characteristics can significantly affect the durability of WIM
systems. Properties such as deflection and rutting clearly influence the long term integrity of
the system in its surroundings. Deep ruts may lead to brittle sensor failure and to shocks
increasing the dynamic vehicle motions and affecting the accuracy. Large deflections may
lead to sensor failure by fatigue or brittle failure, but can also affect the repeatability of the
measurements. Pavement cracking also generally results in sensor failure. Other site
properties are specified such as road geometry and pavement layer thickness. Some
environmental factors are also considered.

The specification defines three classes of WIM site based on rutting, deflection and evenness
limits as presented in Table 1. An indication is then given in Table 2 as to the accuracy class
that is likely to be achievable at a site of a given class. This is in accordance with the results
of most of the tests carried out in various countries and under various pavement conditions.
For bridge and culvert WIM systems, criteria are given for ‘optimal’ and ‘acceptable’ bridge
forms, spans, skews and surface evenness (Snyder, 1981).

173
Table 1 - Classification of WIM sites
WIM Site Class
I II III
(Excellent) (Good) (Acceptable)
Rutting (3 m - beam) Rut depth max. (mm) 4 7  10
-2
Semi-rigid Mean deflection (10 mm)  15  20  30
Deflection pavements Left/Right difference (10-2 mm) ±3 ±5 ± 10
(quasi-static) All bitumen Mean deflection (10-2 mm)  20  35  50
pavements Left/Right difference (10-2 mm) 4 8  12
(130 kN - axle) Flexible Mean deflection (10-2 mm)  30  50  75
pavements Left/Right difference (10-2 mm) ±7 ± 10 ± 15
Semi-rigid Deflection (10-2 mm)  10  15  20
Deflection pavements Left/Right difference (10-2 mm) ±2 ±4 ±7
(dynamic) -2
All bitumen Mean deflection (10 mm)  15  25  35
pavements Left/Right difference (10-2 mm) 3 6 9
(50 kN - load) Flexible Mean Deflection (10-2 mm)  20  35  55
pavements Left/Right difference (10-2 mm) ±5 ±7 ± 10
Evenness IRI index Index (m/km) 0 - 1.3 1.3 - 2.6 2.6 - 4
APL(1) Rating (SW, MW, LW)(1) 9 - 10 7-8 5-6
(1)
The APL is a car-towed device used to measure longitudinal profile in the short (SW), medium
(MW) and long wavelengths (LW) respectively.

Table 2 - WIM Accuracy Class Likely to be Achievable in given WIM Site Class
Accuracy Class WIM Site Class
I (Excellent) II (Good) III (Acceptable)
A(5) Sufficient Insufficient Insufficient
B+(7) Sufficient May be Sufficient Insufficient
B(10) Sufficient Sufficient Insufficient
C(15) More than Sufficient Sufficient Sufficient
D+(20) - D(25) More than Sufficient More than Sufficient Sufficient

5. Calibration

Various calibration methods are possible for WIM systems, depending on the used sensor and
on the reference weights (static or impact forces). The use of pre-weighed calibration trucks is
the most commonly applied, because it is simple, direct and applicable to all forms of WIM;
in such a case the best static weight estimation is required. Therefore this method in central in
the specification. Nevertheless, some other methods are proposed, such as using calibration
masses in static, or automatic self-calibration procedures (Stanczyk, 1991).
The specification defines four levels of repeatability/reproducibility test conditions as follows:
r1 - Full repeatability conditions: one vehicle passes several times at the same speed, load
and lateral position.
r2 - Extended repeatability conditions: one vehicle passes several times at different speeds,
different loads and with small variations in lateral position (in accordance with typical traffic).

174
R1 - Limited Reproducibility conditions: a small set of vehicles (typically 2 to 10),
representative in weight and silhouette of typical traffic, is used. Each vehicle passes several
times, at different combinations of speed and load and with small variations in lateral
position.
R2 - Full Reproducibility conditions: a large sample of vehicles (some tens to a few
hundred), taken from the traffic flow and representative of it, pass over the system, each of
them passing only once.
Calibration under full repeatability conditions is not recommended, because the relative
variability is much lower than under normal weighing conditions. Calibration under extended
repeatability conditions is allowed if the client approves. The vehicle should be typical of the
traffic being weighed and should be driven in fully loaded, half loaded and eventually
unloaded conditions, and at speed levels representative of the site. Calibration under limited
reproducibility (or extended reproducibility if possible) conditions, is recommended. The set
of calibration vehicles and loading cases should be representative of the traffic pattern. The
use of instrumented vehicles is discussed if the WIM system is to be used to estimate dynamic
rather than static forces.

Different methods are presented and discussed to calculate the calibration coefficient(s) of a
WIM system, from the calibration test results.

6. Accuracy Classification

The WIM system accuracy classification is based on tests of measured results against
reference values which, it is anticipated, would generally be determined by statically
weighing trucks. To comply with a given accuracy class, the calculated probability that results
are within the interval [Ws (1-), Ws (1+)], where Ws is the accepted reference value, must
exceed a specified minimum, o. The confidence interval width, , is a function of the
accuracy class and the type of result as given in Table 3 and illustrated in Figure 1;
interpolation and extrapolation of accuracy classes is allowable. The minimum probability o
is a function of the test conditions (repeatability, reproducibility, duration, etc.) and the
sample size, as outlined in section 8. The statistical background for this procedure has been
reported by (Jacob, 1997).

Table 3 - Tolerances of the accuracy classes ( in %)


Criteria (type of Domain of use Accuracy Classes:
measurement) Confidence interval width  (%)
A (5) B+ (7) B (10) C (15) D+(20) D (25) E
1. gross weight Gross weight > 35 kN 5 7 10 15 20 25 > 25
axle load Axle load > 20 kN
2. group of axles 7 10 13 18 23 28 > 28
3. single axle 8 11 15 20 25 30 > 30
4. axle of a group 10 14 20 25 30 35 > 35

175
Tolerances of Accuracy Classes
70

60

50 E...
40 D
 (%)

D+
30 C Gross weight
B
20 Group of axles
B+
Single axle
10 A
Axle of Group

0
0 5 10 15 20 25 30 35 40 45 50
Accuracy Classes

Figure 1 - Variation of the tolerances with respect to the accuracy classes

7. Initial and In-Service Verifications of a WIM System

Accuracy classification of a WIM system may be carried out either as an initial verification or
in-service. The first case applies after installation, or after some modifications, repair or part
replacement. The second case applies when an operational WIM system is to be tested.

7.1 Initial verification


In most cases (except when an automatic self-calibration procedure is used) the calibration
procedure provides data which may be used for an accuracy evaluation. In such a case, the
same sample is used for calibration and for accuracy assessment. This is an initial
verification.

If the WIM system is calibrated using static calibration masses, then according to the OIML
recommendation (OIML 1996), all the relative errors must be in the interval [- /2 ; + /2 ] of
the relevant accuracy class and for the criterion being considered.

If the WIM system is calibrated using repeated runs of pre-weighed vehicles or instrumented
vehicles, the confidence intervals given in Table 3 are used, but the tolerance  is reduced by
a multiplicative factor, k, which is provisionally fixed at 0.80. The required level of
confidence of this interval [-k. ; +k. ] is given in section 8.3, and its calculation is explained
in Section 8.4.

7.2 In-service verification


An in-service verification may be done at any time of the lifetime of a WIM system. It should
be done periodically, but particularly if conditions change (traffic conditions, environmental
conditions, etc.), or if there is any doubt about the data accuracy. In such a verification, the
data used for the accuracy assessment must not have been used for any calibration or
recalibration of the system.

If the WIM system is checked using static calibration masses, all the relative errors must be
within the tolerance interval of the relevant accuracy class [- ; + ] for the criterion
considered.

176
If the WIM system is checked using repeated runs of pre-weighed vehicles, the confidence
intervals, given in Table 3 are used. The required level of confidence of this interval [- ; + ]
is again given in section 8.3, and its calculation is as outlined in section 8.4.

8. Procedure to check the accuracy of a WIM system

8.1 General rules


The assessment of the accuracy of a WIM system requires a test, using either repeated runs of
pre-weighed test vehicles, and/or the use of single runs of pre- or post-weighed vehicles from
the traffic flow. The more extensive the test plan, the longer the test period, but a higher
number of vehicle types and runs ultimately leads to a higher confidence in the conclusion.
This means that the customer risk (i.e., the risk of deeming a system to be in a higher class
than is actually appropriate) decreases as the test becomes more comprehensive. In this
analysis, the supplier risk, linked to the statistical estimation of the mean bias, is fixed at 5%.

In this procedure, the customer risk is governed by the probability of an individual error (with
respect to the static load or weight) lying outside of the specified confidence interval
(tolerance). An upper bound of this risk is fixed by specified values (1-0), where 0 is the
minimum required confidence level. This risk (1-0), or the confidence level 0 , may be
chosen by the customer (see section 8.3).

Depending on the objectives, the means and the requirements of the customer, the test may be
carried out during various time periods. This defines another type of repeatability or
reproducibility, called "environmental repeatability or reproducibility":
(I) environmental repeatability: the test time period is limited to a couple of hours within a
day or spread over a few consecutive days, such that the temperature, climatic and
environmental conditions do not vary significantly during the measurements;
(II) limited environmental reproducibility: the test time period extends at least over a full
week or several days spread over a month, such that the temperature, climatic and
environmental conditions vary during the measurements, but no seasonal effect is
considered;
(III) full environmental reproducibility: the test time period extends over a whole year or
more, or at least over several days throughout a year, such that the temperature,
climatic and environmental conditions vary during the measurements and all the
seasonal conditions for the site are encountered.
No recalibration or any manipulation, software adaptation or part replacement can be
conducted on the WIM system during the test period. Only in the case of a long term test (III),
if some part of the system fails, the supplier of the system may be authorised to repair it or to
replace the broken part, under the control of the test organiser. A detailed report about the
failure, its causes and the repair done must be provided.

Test vehicles are vehicles which are pre-weighed on an approved static scale or weigh-bridge,
and perform repeated runs over the system. The static weighing operation must be carried out
carefully on a flat and horizontal static weighing area. The results of lorry weighing during an
enforcement operation by the police may introduce a bias. Therefore, it is recommended that
lorries specially retained for the test for one or two consecutive days are used.

177
8.2 Test Plans
The definition of a test plan consists of the choice of a sample of vehicles, their number and
the conditions of the vehicle runs. These vehicles may be either test vehicles provided by the
organiser, and/or vehicles taken from the traffic flow and pre- or post-weighed. In this latter
case only one run per vehicle is considered.

It is recommended to carry out the test in reproducibility conditions (R1) or (R2). Condition
(r2) is also possible but with enough loading cases distributed within the range of axle/gross
weights to be weighed, and runs per loading case. It is not recommended to perform the test in
condition (r1), except by special agreement of the user (customer).

The specification leaves all the flexibility to the customer or user, so that the most suitable test
plan can be chosen according to its requirements and means. The data analysis procedure
described in section 8.3 can deal with any test plan, for all cases, the level of confidence of the
results is calculated. Alternatively, the customer may choose the appropriate level of
confidence (or the highest acceptable risk),and design a test plan which complies with it.
However, for common checks of WIM systems, some standard simplified test plans are
provided in appendix 1 (simplified rules). These are described briefly below, but more
information about the speed and load cases is given in the specification itself (COST323,
1997b). All of the standard test plans are in environmental repeatability conditions (I), and are
carried out in between one and three consecutive days. The choice of test vehicles should be
based on the most common types of vehicle in the traffic flow.

Test plan N° 1
One 2-axle lorry (T2) which only permits the checking of 2 criteria from among the 4 (single
axle and gross weight), or an articulated tractor with semi-trailer T2R3 (or T2R2) or some
other vehicle types which allow the checking of all 4 criteria. The confidence level is 0 =
95%. This test plan is divided into two sub-plans:
• Test plan N° 1.1: one load, 10 runs, full repeatability conditions (r1).This very short test
is mainly recommended for periodical checks, carried out several times per year, or if only
one type of vehicles is to be weighed by the system.
• Test plan N° 1.2: two loads, 30 runs, extended repeatability conditions (r2). This short test
is recommended for a annual checks of a WIM system.

Test plan N° 2
Two lorries, including a 2-axle rigid lorry (T2), and an articulated tractor with semi-trailer
T2R3 or a rigid lorry with a trailer and a tandem or tridem axle, if the traffic flow contains a
high proportion of this type. This test plan is again divided into two sub-plans:
• Test plan N° 2.1: one load per lorry, 2 x 10 runs, limited reproducibility conditions (R1).
This test can be conducted for a newly installed WIM system, or after repair or
modification of the system, if customer resources and time are limited. The confidence
level is 90%, lower than for the other tests.
• Test plan N° 2.2: two loads per lorry, 110 runs, limited reproducibility conditions (R1).
This test is recommended for a newly installed WIM system, or after repair or
modification of the system. The confidence level is 95%.
Test plan N° 3
Four lorries including a 2-axle rigid (T2), a 3- or 4-axle rigid (T3 or T4), an articulated tractor
with semi-trailer (T2R3) and a road train. 110 runs are required in limited reproducibility
conditions (R1). This test plan is the most representative. However, it is quite an extensive

178
test, which may require excessive resources for a common acceptance test. Nevertheless, it is
recommended for new types of system or if several systems are tested together at the same
test site. The confidence level is 95%.

8.3 Confidence levels


When a test is performed with a non-standard test plan, the confidence level  to get an
individual error within the confidence intervals specified in Table 3 may be estimated from
the test results and statistics. The individual errors are assumed to be random, independent of
each other and Normally distributed.

The mean error estimation is affected by a statistical uncertainty, which depends on the
sample size n. This uncertainty is taken into account in the specified values of Tables 4 to 6
and in the formulas of section 8.4, assuming that the samples have Normal distributions (this
may be checked by testing if required).

Table 4 - Minimum levels of confidence o (in %), "environmental repeatability" (I)


Sample size (n) 10 20 30 60 120 
Test conditions
Full repeatability (r1) 95 97.2 97.9 98.4 98.7 99.2
Extended repeatability (r2) 90 94.1 95.3 96.4 97.1 98.2
Limited reproducibility (R1) 85 90.8 92.5 94.2 95.2 97.0
Full reproducibility (R2) 80 87.4 89.6 91.8 93.1 95.4

Table 5 - Minimum levels of confidence o (in %), "limited environmental reproducibility" (II)
Sample size (n) 10 20 30 60 120 
Test conditions
Full repeatability (r1) 93.3 96.2 97.0 97.8 98.2 98.9
Extended repeatability (r2) 87.5 92.5 93.9 95.3 96.1 97.5
Limited reproducibility (R1) 81.9 88.7 90.7 92.7 93.9 96.0
Full reproducibility (R2) 76.6 84.9 87.4 90.0 91.5 94.3

Table 6 - Minimum levels of confidence o (in %), "full environmental reproducibility" (III)
Sample size (n) 10 20 30 60 120 
Test conditions
Full repeatability (r1) 91.4 95.0 96.0 97.0 97.6 98.5
Extended repeatability (r2) 84.7 90.7 92.4 94.1 95.1 96.8
Limited reproducibility (R1) 78.6 86.4 88.7 91.1 92.5 95.0
Full reproducibility (R2) 73.0 82.3 85.1 88.1 89.8 93.1

Depending on the test plan repeatability or reproducibility conditions (r1) to (R2), and on the
environmental repeatability/reproducibility conditions (I) to (III), the minimum values of the
required level of confidence for the centred confidence intervals specified in Table 3. 0 are
given in Tables 4 to 6. The minimum value 0 increases with the size of the test data sample
n. For sample size n not mentioned in these tables, the figures may be interpolated. It is
recommended to require - through the choice of the test plan - a confidence level greater than

179
or equal to 90% in reproducibility conditions (R1) and (R2), and greater than or equal to 95%
in repeatability conditions (r1) and (r2).

8.4 Analysis of test results


General procedure
The general procedure for the analysis of test results consists of the following steps:
1. Report on the system failures or malfunctions, including statistics about the time of
operation, the time interval between failures, etc..
2. Statistics about the number of vehicles properly recorded by the WIM system; the
percentage of missing vehicles (not including the vehicles recorded with an error code)
must be lower than the values indicated in the last line of Table 3.
3. Specific analysis of the detected outliers.
4. In the case of a sufficiently large data sample, it is recommended to check the Normality
of the results when only independent random errors are expected to provide the population
variance; this is mostly the case and non-Normality often reveals some dysfunction.
Moreover this assumption is made for calculation of the level of confidence.
5. The relative errors with respect to the static weights and loads (or any other accepted
reference values) are calculated, for each measurement of the different sub-populations,
i.e., the axles, axle groups, axles of groups and gross weights, as:
(Wd i − Wsi )
xi =
Wsi
where Wdi and Wsi are the in-motion measured value and the reference (static) value
respectively of the same unit. Then the mean m and the standard deviation s of each sub-
population sample are calculated.
6. The confidence level  may be either estimated by a theoretical method using the sample
statistics of the test, or, in some cases, by a sample proportion. The theoretical approach is
described briefly here (see Jacob 1997). An individual value is considered of a relative
error, taken randomly from a Normally distributed sample of size n, with a sample mean
m and standard deviation s. A lower bound , on the probability for that individual value
to be in the centred confidence interval [- ; + ], is given at the confidence level (1-) by:
 = (u1 )-(u2 ) (1)
where u1 = ( -m) /s - t,1-/2 /n1/2 , u2=(- -m) /s + t,1-/2 /n1/2,  is the cumulative
distribution function of a Student variable, and t,1-/2 is a Student variable with  = n-1
degrees of freedom. The parameter  is taken to equal 0.05. Then the estimated level of
confidence , for each sample (and criterion) is calculated using Equation (1). For the case
of an initial verification (same data used for calibration and checking),  is replaced by k.
in Equation (1).
7. The accuracy level of the WIM system is assessed using one of two methods:
(i) For each sub-population (sample) corresponding to a criterion from Table 3, and for the
proposed (required) accuracy class defined by  , the acceptance test is:
• if   0 , the system is accepted in that class;
• if  < 0 , the system cannot be accepted in the proposed accuracy class, and the
acceptance test is repeated with a lower accuracy class (a greater ).  should be
recalculated using equation 1.

180
(ii) An alternative way is to calculate, using equation 1, the (lowest) value of  which
provides:  = 0 , and then to check that  is less than the value specified in Table 3 for
the proposed accuracy class and criterion. This approach allows a system to be classified
in any accuracy class defined by an arbitrary -value (the best one). It is mainly
recommended for detailed research and testing studies.

Simplified procedure
In the case of a standard test plan, a simplified procedure, based on the use of a graphical
chart for the acceptance test replaces the calculations of steps 6 and 7. This simplified
procedure is very easy to implement and only requires a pocket calculator. Steps 1 and 2 as
described above are applied. From that point on, the procedure is as follows:
3. The individual relative errors (in %) with respect to the static load (weight) or the
accepted reference values are calculated: x’i = 100*xi.
4. The mean m’ and the standard deviation s’ of each sub-population of x’i (same entity) are
calculated.
5. For the accuracy class considered, m’/s’ and  /s’ are calculated ( is given in Table 3); in
the case of an initial verification,  is replaced by k..
6. In the diagrams of Figure 2, for each test plan and each sub-population of size n, one
curve delimits the "acceptance domain" and the "rejection domain".
For each entity (sub-population), if the point with co-ordinates (m/s ; d/s) in the relevant
diagram is in the "acceptance domain", then the considered accuracy class is accepted;
otherwise, the considered accuracy class is rejected. In the latter case, a lower class is
considered, and the process is repeated from Step 5.

Test plans N° 1.1 - 1.2 (confidence level 95%) Test plan N° 2.1 (confidence level 90%)
1 lorry, 1- 2 loads, 10 - 30 runs 2 lorries, 1 load/lorry, 2x10=20 runs
4,1 3,7
4 3,6
3,9 3,5
3,8
3,7
Acceptance
3,4
3,3
Acceptance
3,6
3,2
3,5
3,4 3,1

3,3 3
3,2 2,9
/s

/s

3,1 2,8
3 2,7
2,9 2,6
2,8 n=10 n=10
2,5
2,7 n=20 n=20
2,4
2,6
2,5
n=30 2,3 n=30
2,2
2,4 n=60 n=40
2,3 Rejection n=90
2,1
Rejection n=50
2,2 2
2,1 1,9
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5

|m|/s |m|/s

Test plan N° 2.2 (confidence level 95%) Test plan N° 3 (confidence level 95%)
2 lorries, 2 loads/lorry, 110 runs 4 lorries, 110 runs
3,5 3,4

3,4 3,3
n=66-72
3,3
Acceptance 3,2
n=80-94 Acceptance
3,2
3,1 n=100-125
3,1

3
3 n=132-172
2,9 2,9 n=188-260
2,8 2,8
/s

/s

2,7 2,7
n=60
2,6 2,6
n=110-120
2,5
n=180-220 2,5
2,4
n=240-280 2,4
2,3

2,2
2,3 Rejection
2,1
Rejection 2,2

2 2,1
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5

|m|/s |m|/s

Figure 2 - Acceptance charts for standardised test plans

181
For Test Plans 1.1 to 2.1, all possible values for n are given in the relevant diagrams (one
curve for each value). For Test Plans 2.2 and 3, the number of possible values for n becomes
too high, and each curve corresponds to an interval for n. For a modified test plan which
provides a number n outside of these intervals, either:
- the curve applicable to the closest interval below n should be considered (i.e., for test plan
N°2.2, if n = 150, the interval [110-120] should be selected in preference to [180-220]), or
- the general procedure described above should be applied.

9. Conclusions

This paper provides a summary of the current European WIM specification. This document
meets the requirements of most users and customers and will facilitate their relationships with
the manufacturers and suppliers. The specification is not in conflict but complementary with
the currently drafted OIML recommendations concern legal applications and fully controlled
low-speed WIM.

Similar accuracy classes are defined in both documents. The OIML recommendation only
refers to the common metrological basis. On the other hand, it can be seen that the
classification system presented in this paper is based on elaborate statistical concepts, which
is appropriate for the specific nature of WIM. While this may at first appear rather complex, it
is quite comprehensive and allows great flexibility for the user. The complexity may be
avoided for common tests through the use of the proposed standard test plans and a very
simple procedure using charts of acceptance. In general cases, some simple software tools
such as spreadsheets may automatically perform all the required calculations and accuracy
assessment.

The proposed tools and framework are detailed. They are intended to be sufficiently general
and flexible to evolve in the future and to be adapted to the expected technical progress in
WIM systems.

This specification was used as the rules and procedure for the European Test Programme
(CET and CMT) of the COST323. According to the preliminary results, it seems to be very
efficient, and the figures seem to lead to consistent results. After the end of this large test
programme, the final version of the specification will be edited, taking into account the output
of these tests (but the rules will not be modified for the test report).
Acknowledgements
The COST323 Management Committee members who contributed to the drafting and
improvement of the European Specification on WIM provided a significant contribution to
this paper.

References
• ASTM (1994), Standard Specification for Highway Weigh-In-Motion (WIM) Systems with
User Requirements and test Method, ASTM Committee E-17, E-1318-94.
• COST 323 (1997a), Collection and Analysis of Requirements as regards Weighing
Vehicles in Motion, EUCO-COST/323/2E/97, EC/DGVII COST Transport Secretariat,
Brussels, January, 64 pp..
• COST 323 (1997b), European Specification on Weigh-In-Motion of Road Vehicles, Draft
2.2, EUCO-COST/323/5/97, EC/DGVII COST Transport Secretariat, Brussels, June.

182
• Jacob, B. (1997), "Assessment of the accuracy of WIM systems", submitted to Heavy
Vehicle Systems, International Journal for Vehicle Design, February/March.
• METT-LCPC (1993), Recommendations for the choice of sites and the acceptance of
weighing stations, National French WIM Project, vol. 2, LCPC, Paris, November.
• NWML (1995), "In-motion road vehicle weighing instruments", First committee draft
Recommendation, National Weights and Measures Laboratory (NWML) - UK.
• OIML (1996), "In-motion road vehicle weighing instruments", Second committee draft
Recommendation, International Organisation for Legal Metrology (OIML), November.
• Snyder, R.E. (1992), 'Field Trials of Low-Cost Bridge Weigh-in-Motion', Federal
Highways Administration Publication No. FHWA-SA-92-014, June.
• Stanczyk, D. (1991), "Etalonnage automatique des stations de pesage ", dans Actes de la
Journée Nationale Pesage en Marche, LCPC, Paris, pp. 37-40.
• VIEA (1994), "Vehicle Weigh in Motion Sensor detailed user requirements", Vehicle
Inspectorate Executive Agency, Bristol, UK, September.

183

View publication stats

You might also like