oiml bulletin july 2003

8/12/2019 Oiml Bulletin July 2003

1/45

ISSN0

473-2812

OIML

BULLETIN

VOLUME XLIV NUMBER 3

JULY 2003

Quarterly Journal

Complete 2020 Seminar Proceedings now available on CD-ROM

What Will LegalMetrology

be

in the

Year 2020?

Organisation Internationale de Mtrologie Lgale


2/45

T H E O I M L B U L L E T I N I S T H E

Q U A R T E R L Y J O U R N A L O F T H E

OR G A N I S A T I O N I N T E R N A T I O N A L ED E M T R O L O G I E L G A L E

The Organisation Internationale de Mtrologie Lgale(OIML), established 12 October 1955, is an inter-governmental organization whose principal aim is toharmonize the regulations and metrological controlsapplied by the national metrology services of itsMembers.

ED I T O R - I N-C HIEF : Jean-Franois Magaa

ED I T O R : Chris Pulham

2003 SU B S C R I P T I O NRAT E60

ISSN 0473-2812

PRIN TE D IN FRA N CE

GRA N DE IMPRIME RIE DE TROYE S

130, RUE G N RA L DE GA ULLE10000 TROYE S

OIML PR E S I D I U MA N D PR E S I D E N T I A L CO U N C I L

PR E S I D E N TGerard J. Faber (NETHERLAND S)

V I C E -P R E S I D E N T SManfred Kochsiek (GERMANY)

Lev K. Issaev (RUSSIAN FEDERATION )

CH A I R P E R S O N, DE V E L O P M E N T CO U N C I LGhaet-El-Mouna Annabi (TUNISIA )

ME M B E R SJudith Bennett (A USTRA LIA )Alan E. Johnston (C A N A DA)

Wang Qinping (C HIN A)Mitsuru Tanaka (J APAN)

Stuart Carstens (S OUTH AF R I C A)Charles D. Ehrlich (U N ITE D STATES)

Jean-Franois Magaa( DIRECTOR OFBIML)

OIML SECRETARIAT

BU R E A U INTERNATIONALD E M T R O L O G I E L G A L E (BIML)

11 R UE TURGOT 75009 PA RIS FRA N CE

TEL : 33 ( 0) 1 487 8 1 282FAX : 33 ( 0) 1 4 28 2 1 727INTE R NE T: www.oiml.org or www.oiml. int

BIML TE C H N I C A L AG E N T S

D I R E C T O R Jean-Franois Magaa ([email protected])

ASSISTANT D I R E C T O R SAtti la Szi lvssy ([email protected])

Ian Dunmil l ([email protected])

ED I T O R Chris Pulham ([email protected])

EN G I N E E R SEdouard Weber ([email protected])

Jean-Christophe Esmiol ([email protected])

ADMINISTRATORPhi l ippe Leclercq ([email protected])

OFFICE MA N A G E R Patricia Saint-Germain ([email protected])

AL B AN I AAL GE R I AAUS T R AL I A

AUS T R I ABE L AR USBE L GI UMBR AZ I LBUL GAR I ACAME R OONCA N A D AP. RE P. OF CH I N ACROATIACU B ACY P R USCZ E CH RE P UB L I CDE N MAR KEGY P T

ET H I OP I AF I N L AN DFR AN CEGE R MAN YGR E E CEHUN GAR YI N DI AI N DON E S I AI S L AMI C RE P U B L IC O F I RA NI R E L AN DI S R AE LI TALY

JAPANKAZ AKH S T ANKENYA

DE M . P. R E P. O F KOR E ARE P. OF KOR E AMACE DON I A , THE F O R M E RY UGOSLA V R E P U B L I C O FMO N A C OMOR OCCONE T H E R L AN DSNORWAYPAKI S T ANPOL AN DPOR T UGALROMAN I ARUS S I AN FE DE R AT I ONSAUDI AR AB I ASE R B I A AN D MON T E N E GR O

SLOVAKIASL OV E N I ASOUT H AF R I CASPAINSRI LAN KASW E DE NSW I T Z E R L AN DTAN Z AN I ATUN I S I AUN I T E D KI N GDOMUN I T E D STATES OF AME R I CAZAMB I A

O I M L

M E M B E R S T A T E S

O I M L C O R R E S P O N D I N G

M E M B E R S

AR GE N T I N ABAH R AI NBAN GL ADE S HBAR B ADOSBE N I NBO S N I A A N D HE R Z E GOV I N ABOT S W AN ABUR KI N A FAS O

CAMB ODI ACOMOR E S , ISL AM IC F ED . REP . OFCOSTA R I CAES T ON I AF I J IGAB ONGH AN AGUAT E MAL AHON G KO N G , CH I N AI CE L AN DJOR DANKUWAITLATVIA

L IBYAL I T H UAN I ALUXE MB UR GMA D A G A S C A RMAL AY S I A

MALTAMAUR I T I USME XI COMOLDOVAMON GOL I AMOZ AMB I QUENEPALNE W ZE AL AN DN I CAR AGUAOMA NPA N A M APA P U A NE W GUI N E APAR AGUAYPE RUSE Y CH E L L E SS I N GAP OR ESY R I ACH I N E S E TAI P E ITH AI L AN DTR I N I DAD AN D TOB AGO

TUR KE YUKR AI N EUR UGUAYUZ B E KI S T ANV I E T N AM

B U L L E T I NVOLU ME XLIV NU M B E R 3

JULY 2003


3/45

t e c h n i q u e

5 Confidence levels of measurement-based decisionsJos G.M. van der Grinten

s e m i n a r 2 0 2 0

12 How will the development of regional and local authorities affectintergovernmental organizations such as the OIML?Jean-Franois Magaa

16 Legal Metrology and the Metre ConventionLev K. Issaev

18 The role of metrology in a cognitive societyThierry Gaudin

20 Towards total approach in legal metrologyBruno Vaucher

u p d a t e

22 OIML Certificate System: Certificates registered by the BIML, 2003.02 2003.04

26 Reports: Towards a Global Legal Metrology System: Speech given byGerard Faber in Maastricht (Milestones in Metrology Conference)

29 OIML TC 12 Meeting

30 WTO/OIML/IEC Seminars

31 13th COOMET Meeting

34 SADC SQAM Meetings

36 Press Release: Cooperation with the Metre Convention,ILAC and other International Organizations

37 Assessment of OIML Activities, 2002

43 Announcements of forthcoming events

44 New CIML Members; OIML Meetings; Committee Drafts received by the BIML

What Will LegalMetrology

be

in theYear 2020?COMPLETE 2020 SEMINAR PROCEEDINGS

NOW AVAILABLE ON CD-ROM

OIML

BULLETINVOLUMEXLIV NUMBER3

JULY2003

Contents


4/45

t e c h n i q u e

5 Niveaux de confiance des dcisions bases sur le mesurageJos G.M. van der Grinten

s e m i n a r 2 0 2 0

12 Quelles sont les consquences du dveloppement des autorits rgionaleset locales sur les organisations internationales telles que lOIML ?Jean-Franois Magaa

16 La Mtrologie Lgale et la Convention du MtreLev K. Issaev

18 Le rle de la mtrologie dans une socit cognitiveThierry Gaudin

20 Vers une approche globale en mtrologie lgaleBruno Vaucher

u p d a t e

22 Systme de Certificats OIML : Certificats enregistrs par le BIML pour la priode 2003.02 2003.04

26 Rapports: Vers un Systme Global de Mtrologie Lgale: Discours prononc par Gerard Faber Maastricht (Confrence Milestones in Metrology)

29 Runion OIML TC 12

30 Sminaires OMC/OIML/CEI

31 13me Runion COOMET

34 Runions SADC SQAM

36 Communiqu de Presse: Coopration avec la Convention du Mtre,ILAC et dautres Organisations Internationales

37 valuation des Activits OIML, 2002

43 Annonces des prochains vnements

44 Nouveaux Membres du CIML; Runions OIML; Projets de Comit reus par le BIML

Sommaire BULLETIN

OIMLVOLUMEXLIV NUMRO3

JUILLET2003


5/45

Ed ito r ia l

Reachi n g con sen sus w i l l lead t o global su ccess

The year 2002 was a period during which we activelylooked into the future directions legal metrology willhave to take, and specifically how the OIML can not

only react but anticipate future needs in line with thechanging tendencies.

The Birch Report has progressed, and very interestingand fruitful discussions were held with John Birch both bye-mail and face to face during the 37th CIML Meeting. The2020 Seminar held in Saint-Jean-de-Luz was a resounding

success and the complete proceedings have just been pub-lished, opening up the way to detailed future research andstudy on the ideas put forward by a wide-ranging group ofinternational experts. A Task Group on DevelopingCountries was also formed last year, much progress wasmade on the Mutual Acceptance Arrangement, and manyother fields of interest continued to develop.

We are witnessing growing interest in increased cooper-ation on the part of other organizations (such as the WTOand UNIDO). A large number of conferences, seminars andworkshops are regularly organized throughout the world onmetrology and legal metrology and we can observe the

effects of the marked increase in awareness of metrology

outside our own community. The OIML must now answerthe needs of its stakeholders over the coming years.

It is now time for the Bureau - and very shortly for therest of the Organization - to embark on our journey into thefuture that we only envisaged and imagined up to now. Thisrequires a profound transformation of the methods of workof the BIML, and we already started this process about oneyear ago.

Change and evolution will also require willingness on

the part of OIML Members to actively contribute. And thesuccess of the OIML will depend on the willingness of ourMembers to reach consensus, to consider that solutions pro-posed by other countries may be as effective as their ownnational systems, to accept that evolutions at internationallevel may cause changes to happen in their nationalapproach towards legal metrology.

Finally, in the development of a global metrology sys-tem, if the OIML were to fail then this would translate intoa failure for each national legal metrology system. Successglobally for the Organization, on the other hand, our pri-mary objective, will hugely benefit Members and pave the

way to a strong and bright future.

J E AN-FRANOIS M AGAA

DIREC TOR, B I M L


6/45

Abstract

M etr ological decision s are based on m easurement s that

have uncertai nt ies. Exam ples are car veloci ty measure-

ments for law enfor cement, tests, veri ficati ons and

in specti ons that lead to the decision to appr ove or reject

an i nstrum ent, an d the signi ficance of differences found

dur in g intercompari sons. Thi s paper show s for each of

th ese exam ples the relati onshi p between acceptan ce

criterion , tolerance, u ncertai nty and confi dence level.

From the discussion of t hese exampl es it can be

concluded that u ncertai nti es must be know n in order to

evalu ate th e ri sk on an erro neous decisio n. Conf idence

levels are associated wi th decisions for w hi ch i t i s

im possib le to achi eve 100 % con fi dence.

Conform ance and non -conform ance are not tw o comple-

mentary n otion s. If t he accepted ri sk on an erroneous

decision is less than 50 % there is a range of

observations for whi ch the in strum ent i s not conform ing

and not n on-conform in g at the same tim e.

For veri ficati ons, an in creasin g number of verif icati on

poin ts leads to an in creased ri sk of makin g an in correct

decision. I n order to appreciate the extra i nform ation of

more observat ion s, a cur ve fi t pr ocedu re described by

Van der Gri nten and Peters [ 1] can be follow ed. I f th ere

are suf fi cient data, i .e. at l east 6 degrees of fr eedom , it is

best to m ake a curv e fit wi th a 95 % confidence envelope.

I n al l of t he exam ples discu ssed above the stati stical

distri but ion of the observed results is not kn own . So the

risk an alysis is based on the assumpti on of a Gaussian

distri but ion of the measurement r esults that i s the worst-

case represent ati on of ou r kn ow ledge. I f oth er

distri but ion s can be demonstrated to describe the

measur ement results thi s wi ll certai nly lead to a higher

degree of con fi dence or acceptan ce criteri a t hat are closer

to t he tolerances.

Introduction

In the daily practice of metrology many decisions are

based on tests or measurements. Instruments may beplaced on the market and put into use after it has beenclearly demonstrated that they meet the applicablemetrological requirements, especially the accuracyrequirements. And if instruments are in use for sometime they may be subject to a re-verification system oran in-field inspection system that is supervised by thegovernment (Market Surveillance). Here an instrumentwill be rejected after it is demonstrated that it isoperating outside its metrological limits. Also in lawenforcement people obtain a ticket if they have exceededthe limits beyond all reasonable doubts. So one may say

that in legal metrology every measurement results in adecision: good - not good, fault - not fault. In otherwords it is decided that the instrument is conforming ornot conforming, or that it is non-conforming or not non-conforming. As will be shown later, there is a differencebetween non-conforming and not conforming. Thedecisions based on doping tests that are carried out inmodern top sports, require the same level of confidenceas speeding tickets.

In scientific and industrial metrology, calibrations ofinstruments do not result in decisions. The deviations ofan instrument are simply reported on a calibrationcertificate without making reference to a tolerance.However, industrial production requires (statistical)process control to monitor the production quality.Adjustments are made if production is no longer withinpreset factory tolerances.

One of the most important decisions in inter-comparison testing is if deviations between laboratoriesare significant or not.

In the above situations it is vital that reliabledecisions are taken. The reliability of a decision isexpressed by the confidence level, which is one minusthe probability (risk) that an erroneous decision istaken. If the measurement value is close to the tolerance,

part of the uncertainty interval is within and anotherpart is outside the tolerances. In other words due to themeasurement uncertainty four possibility arise:

a) The object is within tolerances and is approved.

b) The object is within tolerances and is not approved.

c) The object is outside tolerances and is notapproved.

d) The object is outside tolerances and is approved.

UNCERTAINTY

Confidence levels of

measurement-baseddecisions

JOS G.M. VAN DER GRINTEN,NMi Certin B.V., Dordrecht, The Netherlands

5

technique

O I M L B U L L E T I N V O L U M E X L I V N U M B E R 3 J U LY 2 0 0 3

This Paper was presented by the Author at the 11th Flomeko Conference, May 2003, Groningen, The Netherlands. The Editors of theOIML Bulletin wish to thank both the Author and Flomeko for their kind permission to reprint it.


7/45


8/45

7

technique


In the Netherlands the instruction of the ProsecutionCounsel is to fine the speed excess with a threshold of7 km/h for vtol < 100 km/h, and for vtol 100 km/h thethreshold is 8 km/h [6]. The fines are based on theobserved velocities reduced by 3 km/h for vobs by 3 % for

vobs 100 km/h, respectively [7]. In combination with anin-field MPE of 3 km/h or 3 % this means that the risk oferroneously being given a speeding ticket is limited tothe order of 0,01 %. Since the tariffs for speeding showa step-wise increment, the maximum risk of an incorrectamount on the ticket is 4,2 %.

During type approval and initial verification thevelocity meter has to stay within an MPE of 1 km/h forvobs < 100 km/h or 1 % for vobs 100 km/h. During a fieldtest the velocity meter has to stay within the 3 km/h or 3 % [7]. In addition, the Prosecution Counsel hasordered all speed meters to be re-verified annually. This

guarantees the motorist only a low risk of being given anincorrect speeding ticket.

Testing and verifications, single point case

In the previous section the confidence of exceeding one

tolerance was calculated. This section deals with testingand verification where it needs to be demonstrated thatthe instrument stays within two tolerances or MPEs.The starting hypothesis is that the instrument isperforming within tolerances. Since two tolerances areinvolved, this test is called a two-sided test. Theconfidence level of the test, i.e. the probability that theobserved value is between tolerances, equals:

(4)

information is available on the statistical distributionassociated with the measurement uncertainty. So anassumption needs to be made that corresponds to theworst-case situation: i.e. the Gaussian distributionwhere the standard uncertainty equals the standarddeviation. This distribution function is not a distributionin the statistical meaning but is a knowledge represen-tation. An example is given in Figure 2 for a car that isdriving 2,5 km/h too fast, which corresponds to car (3)

in Figure 1.Now the probability that a motorist is driving faster

than the speed limit P(vobs > vtol) can be calculated fromthe cumulative Gaussian distribution:

(3)

in which z = (v vobs)/us, ztol = (vtol vobs)/us where visthe velocity, vtol the speed limit and us is the standarduncertainty of the observed car velocity vobs. The riskthat the car is driving at less than the speed limit is1 P (vobs > vtol) and is displayed as the shaded area in

Figure 2. Since only one tolerance is involved, this test iscalled a single side test.

The result for different car velocities is shown inTable 2. In the first column the excess of the speed limitis expressed as a multiple of the standard uncertainty.The second column shows the probability that themotorist is exceeding the speed limit. The last column isthe risk of an erroneous decision if the police give themotorist a speeding ticket. From Table 2 it is shown thatif a motorist is exceeding the speed limit by 2 standarduncertainties or more, the risk of an erroneous penaltyis 2,3 % or less. Due to the worst-case character of the

assumed Gaussian distribution the confidence level ofthe decision taken is always higher than calculated.

(vobs vtol)/us P(vobs > vtol) Risk

1,00 84,1 % 15,9 %

1,64 95,0 % 5,0 %

1,96 97,5 % 2,5 %

2,00 97,7 % 2,3 %2,33 99,0 % 1,0 %

3,00 99,9 % 0,1 %

Figure 2 Velocity knowledge distribution corresponding to a cardriving 2,5 km/h too fast and a standard uncertainty of1,73 km/h. The shaded area represents the risk that themotorist is incorrectly given a speeding ticket.

Table 2 Confidence levelsP(vobs >vtol) for a single sided test

depending on the observed relative speed excess(vobs vtol)/us. The risk is 1 P(vobs >vtol).


9/45

8

technique


In line with previous discussion, Sommer andKochsiek [8] propose acceptance criteria obtained by

reduction of the tolerance with the expanded measure-ment uncertainty. They also state that this results in ade-facto reduction of error limits and that common usein legal metrology seems to be unlikely due to thecommercial implications of such a reduction. Moreoverthe acceptance criteria are variable, depending on theuncertainties that are obtained by different laboratories.

It is very clear that 100 % confidence or zero risk cannever be obtained when taking metrological decisions.For a given confidence level a smaller measurementuncertainty results in acceptance criteria closer to theapplicable tolerances. This is an advantage for manu-

facturers that have their instruments verified with a lowuncertainty. The confidence level or risk is also in-fluenced by the Gaussian knowledge distributionassumed. If it can be proven that another statistical

in which z = (e eobs)/us, ztol = (etol eobs)/us where eisthe deviation of the meter, etol are the + and tolerances

and us is the standard uncertainty of the observed meterdeviation eobs. Of course zobs = 0. The decision to acceptthe instrument has a risk of 1 P(etol < eobs < etol+). Theconfidence level is also influenced by the ratio of thetolerance and uncertainty as is shown in Table 3.

The alternative hypothesis is that the meter isperforming outside the tolerances. The probabilities thatthe observed deviation is above the upper tolerance etol+or below the lower tolerance etol are:

(5)

and

(6)

The risks associated with the decisions based onthese test are 1 P(eobs < etol) and 1 P(eobs > etol+),respectively. A graphical display of the risks associatedwith the acceptance or rejection of the instrument isshown in Figure 3. This figure shows that there is a highrisk associated with the acceptance of an instrument ifobservations are outside tolerances. Likewise, rejection

of an instrument if observations are within toleranceshas a high risk. If the maximum acceptable risk on anerroneous decision is 5 %, the width of the rectangles inFigure 3 represents the range of observations in whichno decision can be taken. Between the shaded areas theinstrument is said to be conforming, outside the shadedareas the instrument is non-conforming. Inside theshaded areas the instrument is not conforming and notnon-conforming at the same time. The width of therectangles is dependent on the risk level accepted: thelarger the risk the smaller the width of the rectangle. Incase of a 50 % risk there will be no rectangle at all. In the

latter case, acceptance or rejection of an instrumentcomes close to tossing coins.

|etol| |eobs|/us |etol| us |etol| = us |etol| = 2us |etol| = 2,5us |etol| = 3us

1 84,134 % 68,269 % 83,999 % 84,131 % 84,134 %1,64 95,000 % 94,074 % 94,960 % 94,999 %

1,96 97,500 % 95,433 % 97,382 % 97,497 %

2 97,725 % 95,450 % 97,590 % 97,722 %

2,33 99,000 % 98,625 % 98,988 %

3 99,865 % 99,730 %

Table 3 Confidence levelsP(etol


10/45

9

technique


of this approach is that despite obtaining a betterimpression of the meter performance, the risk attributedto erroneously approving the meter increases with anincreasing number of observations.

A method that utilizes the additional information ofthe extra verification points is the linear curve fitmethod developed by Van der Grinten and Peters [1]. Asa standard curve fit method does not count with the

individual uncertainties of the data points, it is onlysuitable for type A evaluation of uncertainties. Verifica-tion data, however, contain uncertainties that are theresult of both a type A and a type B evaluation. For thelinear case Van der Grinten and Peters calculate the fit,which is identical to the standard regression method,and an uncertainty envelope in which the uncertaintiesof the individual data are included. All regressionmethods require at least 6 degrees of freedom. Theregression line is:

(7)

where m is the regression coefficient, e0 the deviation atx0 and Qthe flow rate in m

3/h. If x0 and e0 are chosen tobe the arithmetic averages of transformed flow rates andthe deviations, respectively, the uncertainty envelopeUe (k=2) is found from [1]:

(8)

in which U(m) and U(e0) are the combined uncertaintiesof the regression coefficients mand e0. The trend linewill meet the tolerances with a confidence level 1 ifthe bounds e(x) are within the tolerances, where:

(9)

distribution gives a more adequate knowledge represen-tation, this will most certainly result in acceptancecriteria that are closer to the tolerances. The use ofdistributions will be an important research issue in thenear future.

The current practice of initial verification and lawenforcement by in-field inspection of fuel dispensers inThe Netherlands is already aiming at limiting the risksof metrological decisions. Accredited organizations thatperform the initial verification of a fuel dispenser use asan acceptance criterion 0,4 % where the tolerance is 0,5 %. The law enforcement officers close the fuel dis-penser if it is deviating more than + 0,7 %. For devia-tions within 0,5 % no action is required. In all othercases repair by the owner is required. This practice hasworked for more than 10 years to the satisfaction of allparties involved [9] due to a reduced risk that an already

approved fuel dispenser will appear non-conforming.Another example of risk reduction concerns coriolismeters intended for metering gasses, which may beverified with water if reduced tolerances are used. Thereason is that the use of gas introduces additionaluncertainties. This practice has not been established yetfor other metrology areas in The Netherlands. In thelight of these different metrological practices, thequestion of which risks are acceptable should beelevated to the level of the OIML Technical Committeeon verification.

Testing and verifications, multi-point case

In the previous section the confidence level of a singlepoint test was determined. However most instrumentsoperate in a range and the result of a test or verificationis based on observations spread over the range of theinstrument. If one of the observed deviations does notmeet the acceptance criteria, the meter is not approved.Now the risk that the instrument is not conforming, isthe sum of the risks of all individual observations. An

example of a verification of a turbine gas meter is shownin Table 4. For a normal initial verification the deviationis measured at 6 different flow rates. For curve fittingpurposes two additional verification points are added.The right-hand column shows for each observation therisk that the observation leads to an erroneous approvalof the meter. The highest contribution is found at 20m3/h where the difference between the observeddeviation and the tolerance equals the uncertainty. If thedifference between tolerance and observed deviation istwo uncertainties (k=2) or more, its contribution to therisk of erroneously approving the meter can be

neglected. For a better impression of the performance ofthe meter more observations can be made. The paradox

Flow rate DeviationUe (k=2) Tolerance Risk m3/h e

5 1,50 % 0,40 % 2 % 0,62 %

10 0,00 % 0,40 % 2 % 0,00 %

20 0,70 % 0,30 % 1 % 2,28 %

40 0,55 % 0,30 % 1 % 0,13 %

70 0,40 % 0,30 % 1 % 0,00 %

100 0,00 % 0,30 % 1 % 0,00 %

3,03 %

55 0,50% 0,30% 1% 0,04%

85 0,30% 0,30% 1% 0,00%

Table 4 Example of a verification of a turbine gas flow meter. Thethird column represents the uncertainty of the observeddeviation. The last column represents the risk that anobservation leads to an erroneous approval of the meter.

At the 8th row the sum of all risks is displayed. The twobottom lines are additional verification points that are

used in the curve fit.


11/45

10

technique


the absolute value of the difference is greater than theuncertainty of the difference:

(10)

Here R1 U1 is the result by laboratory 1 and R2 U2is the result by laboratory 2. U is the expandedmeasurement uncertainty with k=2. Another quantityused for comparing results is the normalized differenceEn:

(11)

If En is greater than 1 the difference is calledsignificant. What is the degree of confidence associatedwith the above significance criteria? Or in other wordswhat is the risk of the decision that the results of twolaboratories are significantly different. Again the answercan be given by means of statistical testing and again thestatistical distribution of the difference |R1 R2| is notknown. Also here the worst-case approximation is theassumption of a Gaussian distribution with an averageof R = |R1 R2| and a standard deviation sequal to:

s = 1_2(U1

2 + U22 ).

Applying the transformation z = (r R)/sthe con-fidence level is obtained from the standard Gaussiandistribution:

(12)

So the maximum risk of the decision that two labor-atories have different results is 4,6 %.

Conclusions

From the preceding analysis the following conclusionscan be drawn:

From a statistical perspective, confidence levels andrisks are associated with decisions, not with un-certainties.

It is impossible to achieve 100 % confidence for anydecision taken, i.e. there is always a risk of anincorrect decision.

There are two types of tests: one is carried out withthe objective of approving the meter; the other typeis to find instruments that are performing outsidethe metrological tolerances. If the accepted risk on

an erroneous decision is less than 50 % there is arange of observations for which the instrument can

and t1-/ 2 follows from the Student-t distribution for 6degrees of freedom. The results of applying this methodto all the experimental data of Table 4 are displayed inFigure 4, where the results are transformed back to theflow rate domain. Figure 4 shows that the 95 % con-

fidence envelope is within the tolerances everywhereexcept for the lowest flow rate. A better fit with smalleruncertainties will be obtained if higher order curve fitsare utilized for the trend of the meter curve. To this endthe method developed in [1] has to be generalized for themulti-linear regression case.

Figure 4 Linear curve fit (solid line) together with the 95%confidence envelope (dotted lines) of the verification

data of Table 4. The deviation is plotted versus theindicated flow rate in m3/h.

For the moment it seems practical to use thefollowing strategy to determine the risk associated withapproving an instrument. For low numbers of data therisk of erroneously approving an instrument is the sumof the risks that an individual data point leads to anincorrect decision. If there are sufficient data, i.e. atleast 6 degrees of freedom, it is best to make a curve fit

with a 95 % confidence envelope.

Intercomparisons

The last example of taking decisions concerns inter-comparisons and round robins when measurementresults from different laboratories are compared. If twoor more laboratories measure the same specific quantityunder fully comparable conditions they will obtain

different results. According to EA recommendation EA-2/03 [10] two measurement results differ significantly if


12/45

11

technique


[5] Eadie, W.T., D. Drijard, F.E. James, M. Roos,B. Sadoulet (1971): Statistical methods in experi-mental physics, North-Holland PublishingCompany, Amsterdam London.

[6] See the website of the Dutch Openbaar Ministerie(Prosecution Counsel):www.om.nl/beleidsregels/dbase/verkfrm.htmdocumentwww.om.nl/beleidsregels/docs/2002a014.htm(in Dutch)

[7] Private communication with Mr. Paul Kok of NMiCertin.

[8] Sommer, K.D. and M. Kochsiek (2002): Role ofmeasurement uncertainty in deciding conform-ance in legal metrology, OIML Bulletin Vol. 18, No.2, pp. 1924.

[9] Private communications with Mr. Aart Kooimanand Mr. Ed van Rmer of Verispect.

[10] European Accreditation (1996): EA Inter-laboratory Comparison (previously EAL-P7),document EA-2/03, rev. 01, March 1996.

Con tac t In fo

Dr. ir. Jos G.M. van der GrintenNMi Certin B.V.P.O. Box 394

3300 AJ DordrechtThe NetherlandsTel.: +31 78 633 2368Fax: +31 78 613 4647E-mail: [email protected]

About the Aut hor

Jos van der Grinten has a background in appliedphysics. Within NMi he has worked in the nationalstandards laboratory on gas and liquid flow standardsand measurement uncertainty. Presently, he works in

the legal metrology section of NMi in the field of utilitymetering and industrial gas metering.

neither be approved nor rejected: the instrument isboth not conforming and not non-conforming at thesame time.

In order to establish a relationship betweenacceptance criteria, tolerance, uncertainty andconfidence level a statistical distribution has to beassumed that represents our knowledge of themeasurement results. For this purpose the worstcase approximation is the Gaussian distribution.This means that in practice the confidence level ofthe decision taken will always be higher thancalculated. If our knowledge can be represented byother statistical distributions this will certainlyresult in acceptance criteria that are closer to thetolerances.

For the moment it seems practical to use thefollowing strategy to determine the risk associatedwith approving an instrument based on a multi-point verification. For low numbers of data the riskof erroneously approving an instrument is the sumof the risks that an individual data point leads to anincorrect decision. If there are sufficient data, i.e. atleast 6 degrees of freedom, it is best to make a curvefit with a 95 % confidence envelope.

In intercomparisons the criterion for a significantdifference is if the difference of two results is biggerthan the uncertainty (k=2) of the difference. Theconfidence level of this decision is 95,4 %.

With respect to the acceptable risks of metrologicaldecisions there are different practises in differentmetrological areas. The juridical and commercialimplications of applying lower risk levels should bediscussed at OIML level.

References

[1] Grinten, J.G.M. van der and Charles M.E.E Peters(1994): Trend analysis, general model and

application to high-pressure gas flow standards,i n :Proceedings of the IMEKO XIII conference,September 1994, Turin, Italy, pp. 11611164a.

[2] BIPM/IEC/IFCC/ISO/IUPAC/IUPAP/OIML (1993):Guide to the expression of uncertainty inmeasurement (GUM), First Edition, ISO 1993.

[3] European Accreditation (1999): Expression of theuncertainty of measurement in calibration,document EA-4/02, December 1999.

[4] Grinten, P.M.E.M. van der and J.M.H. Lenoir(1973): Statistische procesbeheersing (Statistical

process control), Uitgeverij Het Spectrum B.V.Utrecht / Antwerpen.


13/45

12

seminar 2020


Globalization in legal metrology has beenon its way for centuries

Historically in feudal organizations taxation dependedon local authorities, on the definition of measurementunits, and on the systematic prosecution of fraud-relat-ed offences concerning the quality and quantity ofproducts traded. Originally, legal metrology was a con-sistent system locally within each feudality, using thedefinition of measurement units as a basis and extend-ing to that of good measurement practices. But thedownside of this local consistency was that importantdiscrepancies were witnessed from one region to anoth-er. Traders had to travel with their own measures and

instruments, and had to deal with significant differ-ences in units and/or in measurement standards fromone city to another.

The formation of states, which brought these feudal-ities together into merged federations, was accompa-nied by a number of harmonization measures: the lan-guage of the ruling bodies became the national lan-guage, currencies were unified and were managed bycentral government, local taxes on the transit of goodswere progressively abolished, and the measurementunits in use in the central capital city became thenational measurement standards. The prosecution of

fraud-related offences concerning the quality and quan-tity of goods generally remained within the scope of

local regulations and jurisdictions. As the centuriespassed, each country established its own national mea-surement system, but local units sometimes survivedand were used locally as customary units.

By the end of the 18 th Century, the situation of

metrology in most countries had already become quitecomplex. The uniformity and consistency which hadexisted in the feudalities had sometimes given way, atnational level, to the coexistence of national and localunits bearing the same name, but having different val-ues. In France for example, one could have to deal withthe pound of Paris (the national one), but also with thepound of Bordeaux or of other cities. The local jurisdic-tions, in charge of fair trading, practiced legal metrolo-gy at their level, but no authority was in charge of uni-fying measurements and legal metrology regulations atnational level.

During the 19th

Century, the development of energyand technologies resulted in the emergence of industryand in the acceleration of trade. The systems of unitswere unified in each country, in order to answer thenew needs of science, technologies and the economy.These systems were extended to new fields of measure-ments, giving rise to new units. A singular country,whose scientists and philosophers had cooperatedtogether for decades, made a political decision at thevery beginning of the 19th Century to abolish the oldunit systems and to introduce a new scientific-basedsystem. In so doing, France anticipated the future needsof unification and consistency and proposed the metric

system to other nations. However, this new system wasonly generalized when the economy and industry feltthere was a real need, some decades later.

The rapid development of national trade during the19th Century convinced most governments to introducesome degree of consistency into their regulations relat-ed to measuring instruments used in trade, thusrebuilding a national legal metrology system. Howeverin a number of countries, and especially in federalstates, the prosecution of fraud-related offences and theimplementation of legal metrology controls remainedthe responsibility of local authorities. Legal metrology

was then often rebuilt in a bivalent way, where mea-surement units and measurement standards, and mostoften technical requirements, were the responsibility ofthe central authority or government, while the imple-mentation of legal metrology controls was the responsi-bility of local authorities or governments.

Very soon, the necessity to harmonize measurementstandards at international level appeared and thisresulted in the adoption of the Metre Convention in1875. The international situation at the end of the 19 th

Century (and up to the middle of 20 th Century) repro-duced on a larger scale the situation which had pre-

vailed nationally at the end of the 18

th

Century: a fairlygood harmonization of measurement units and of mea-

How will the developmentof regional and localauthorities affectinter-governmentalorganizations such asthe OIML?

JEAN-FRANOIS MAGAABIML DIRECTOR


14/45

13

seminar 2020


surement standards at international level, but divergingnational legal metrology requirements and sometimeseven specific custom-designed units.

At the end of World War II, a number of Inter-Governmental Organizations among which the OIML

were founded. All these Organizations had for objec-tive to set up (by consensus) mechanisms for regulationin fields in which countries previously acted individual-ly: international relations (UNO), health (WHO), ali-mentation (FAO), development (UNDP, OECD), finance(IMF), trade (GATT then WTO), etc. The OIMLs objec-tive is to contribute to setting up a Global M easur ementSystem, as described in the report published by CIMLImmediate Past President Knut Birkeland.

Everything could have continued to progress withinthe OIML, as in other Organizations, in a steady andforeseeable way. Based on the legitimacy of states and

on their competence, the OIML developed model regu-lations on the basis of which Member States would vol-untarily harmonize their national regulations and rec-ognize each others measuring instruments and mea-surement results. In this way, the dialogue betweenstates would have been a simple way to provide theintended regulations, if the end of the 20th Century hadnot brought about a number of new transformationswhich also had to be taken into account.

The construction of new economicaland political blocks

In the second half of the 20 th Century the industrial,commercial and financial structures developed in atransnational way, having developed in a national wayduring the 19th Century. This globalization is sometimesconsidered as a totally new phenomenon, but in fact itis a simple and logical continuation of the globalizationpattern started one century before, which led thesestructures to develop naturally from local to nationallevel. This globalization is of course now considerably

accelerated by the development of information tech-nologies.

During the 19th Century, local governments foundthemselves increasingly unable to regulate their respec-tive economies and to face this growing trend towardsglobalization, and national governments had to takeover this mission. Today, in a similar way, individualstates are no longer able to achieve the required eco-nomic regulation and they are organizing themselvesinto regional structures (political and/or economic): theEuropean Union, APEC, SADC, etc.

This construction is still under development, and in

particular has neither abolished nor politically mergedthe individual states - which are in fact the only entities

which may legally participate in intergovernmentalOrganizations such as the OIML. However, in the fieldsof activity of these Organizations, the Member Statesare also transferring an increasingly significant part oftheir power to the regional structures, which deal with

support to the economy, technical regulations, taxes,social protection, etc., and which are players in thefields covered by the International Organizations, with-out being able to be members thereof.

It is possible to come to a consensus on a model reg-ulation within the OIML, while a diverging model regu-lation would be adopted by consensus in a regionalstructure. As regional structures are not necessarilybound by the OIML Convention, they may issue diverg-ing regional regulations and make them binding fortheir Member States. Those OIML Member Stateswhich are also Members of a regional structure may

therefore lose a part of their autonomy and scope ofresponsibility, and may not be able to fulfill all theirobligations towards the OIML. This power transferfrom individual states to regional structures is a loss forthe OIML, if the regional structures do not themselvesparticipate in the OIML.

The fragmentation of states

When the United Nations was founded in 1945, therewere initially 51 UN Member states. Today there are189.

From the middle of the 20 th Century onwards, anexplosion was observed in the number of independentstates, sometimes of a small size. This evolution result-ed from a considerable demand for a return to specificcultural identities. A number of states which existedbefore the middle of the 20th Century were split intoseveral smaller states corresponding to these culturalidentities. Other states evolved towards a decentralizedorganization, in which a large autonomy was granted to

local authorities. Local parliaments were sometimesinstalled, with quite far-reaching powers. Many statesevolved towards a more federal organization, or split upinto different states.

A question may be raised when states are fragment-ed into several smaller independent states: will techni-cal structures be viable in each of these independentstates? Is it appropriate - and possible - to developMetrology Institutes and Legal Metrology Institutes ineach of the smaller states which are similar to thosewhich existed in the original country? What is the mini-mum population or gross national product necessary to

be able to afford such institutes, and what capacity maybe envisaged for them?


15/45

14

seminar 2020


tion, bulk quantity certification and quality systems cer-tification, and who already provide measurement andcalibration services, could quite rightly wish to play aspecific role in national and international legal metrolo-gy.

Such evolutions, which are simply the continuationof the ongoing increasing tendency towards globaliza-tion, raise the crucial question for the OIML of the rele-vance of having formal relations with private transna-tional or international bodies, and having such bodiesplay a specific role in the global legal metrology systemthat the OIML has to develop.

Which evolutions can the OIML expectin this context?

The above considerations do not question the utility ofthe OIML. The need for regulation mechanisms (atinternational level), compatible metrology systems, anda Global Measurement System, become more and moreevident as globalization progresses.

A possibility was conceived some years ago byobservers from outside the OIML: to consider the OIMLas a plain international standardization body and totransfer most of the OIMLs work to the general interna-tional standardization bodies. But this would be anerror. Indeed, the OIML deals with technical issues

using methods close to those employed by standardiza-tion bodies, but the essential purpose is to harmonizeregulations and legal requirements, and the legalaspects - the issues related to law implementation - areof major importance in the Organizations work, includ-ing that which seems to be of a purely technical nature.In addition, such an evolution would be contrary to thegoals and efficiency of the OIML, as the commitment ofMember States would disappear. The strength of theOIML, as a harmonization body, directly stems fromthe legal authorities of the member countries.

The Members of the OIML are states, and can legal-

ly only be states. In the future it will be necessary toimprove the implementation of the obligations specifiedin the OIML Convention, and to make sure that theseobligations are taken into account by the RegionalOrganizations as well as by the local authorities.

This requires a constant dialog between the OIMLand the Regional structures in order to take account oftheir policies, to answer their needs and to encouragethem to make use of the OIML in their policies. It is notforeseeable under the present Convention that regionsbecome members of the OIML and participate in theformal process of decision making, nor in the adoption

of Recommendations. On the other hand, regions couldbe more formally associated in the preparation of the

Federal organization raises a number of questionsto Organizations such as the OIML. In the same way asthe regional structures mentioned above do not havethe status of a state and are not Members of theOrganization, neither the local structures in a decen-

tralized state nor the states of a federation can individu-ally participate in the OIML, while at the same timetheir increasing power may raise new technical barriersto trade.

The development of these federal or decentralizedschemes transfers power to the local structures. Doesthis transfer, added to the transfer of power to theregional structures mentioned above, contribute todecreasing the power of the states? Shall we in theyears to come, see most regulatory activities disappearat the level of states and be transferred partly to region-al structures, partly to local authorities? What would

then be the meaning of intergovernmental treaties suchas the OIML Convention?

The trend towards privatization

Another evolution affects the role of the states in legalmetrology: the present trend to privatize or to delegatethe technical tasks of legal metrology to private bodies.Other lectures in the 2020 Seminar present the conse-quences of this evolution on the role of the states, but

the consequences on the international activity of legalmetrology may also be important.

A number of bodies in charge of important legalmetrology tasks such as type approval and initial verifi-cation, are already private bodies. The technical compe-tence required for OIML work for the most part lies inthese private bodies and they play an increasing role inthe Member State representations in the OIML struc-tures. Is the OIML moving towards a more specializedallocation of competences and work, where theMember States would be essentially present in theConference and where the Committee would essentially

be composed of increasingly private technical bodies?Considering the perspective of several states sharing

resources, any institute that owns costly equipmentused by several states will enjoy, de factoif not de ju re,competence in legal metrology in each of these states.

In this evolution towards privatization or delegationto private bodies, it could happen that a given privatebody be designated for type approval by several coun-tries, that several private bodies from different coun-tries merge or take on mutual shares, or that a privatebody becomes a major shareholder in other countriesbodies.

The international technical control bodies, who areactive in the fields of security control, product certifica-


16/45

15

seminar 2020


OIML Action Plan and priorities. To accomplish this, itis essential that those regions that are already struc-tured become or continue to be partners of the OIML,and that the OIML encourage the development of struc-tures in those regions which are not yet organized.

An example of relations between regional structuresand Intergovernmental Organizations must be noted. Inthe World Trade Organization, the members are states.However, the members of the European Union decidedto delegate their powers in negotiations as well as theirvotes, to the European Commission. This is a very effi-cient way to better involve a Regional Organization ininternational work, and benefits at the same time theRegional Organization, its Members and theInternational Organization. This shows that establish-ing links between an International Organization and aregional structure is not only the task of the

International Organization, but also that of the partici-pating states. The development of relations between theOIML and Regional Organizations will not be doneagainst the Member States will, but in harmony withthem.

To prevent subnational authorities from drawing uplocal regulations which diverge from OIMLRecommendations is a difficult task for the OIML andcan only be done by each Member state. The role of theOIML may only be one of monitoring, communicatinginformation, and maintaining updated databases onnational and local regulations. This function is anextension of the role of the OIML DocumentationCenter mentioned in the OIML Convention. Thisrequires a very important reform of the principles ofthis Documentation Center, in particular using newinformation technologies.

The information society

A phenomenon which has appeared over the last fewyears may play a prominent role in the political and

social evolution at international level, and in the futureof International Organizations.

Globalization advances using the communicationtools that technology and the economy provide it with.

In the 19th Century, such communication tools were therailways, newspapers and telegraph. In the 20th

Century, airplanes, radio, television and telephone wereused and now - since the last few years - the Internet.These are the tools for the globalization of economies,

trade and political organizations. They have differentgeographical ranges and have successively permittedglobalization at the level of countries, then continents,and now worldwide. However, the use of these tools isnot restricted to industry, banks and governments, nowthey are readily available to the general public. After ashort period of diffusion and appropriation, these toolsallow public opinion to be globalized, i.e. they allow theemergence of public opinion within their specific geo-graphical range: a country, a continent, or the world.

Today we can observe the beginnings of an interna-tional public opinion, whose expression is just starting.

International associations are expressing general con-cern about environment protection, durable develop-ment, food safety, and the need for mechanisms to reg-ulate the globalization process. This international pub-lic opinion is still anarchic, it has no clear representa-tion, it may not yet be democratic, but it is appearingand growing, it has a notable influence on national pub-lic opinions, and it will probably be a major politicalfact in the coming years.

This international public opinion needs counter-parts to dialog with. Political counterparts are govern-ments, collectively (G8 summits) or individually. But it

also needs to have a dialog with IntergovernmentalOrganizations, who work on specialized issues onbehalf of governments. It will be essential in the futurethat International Organizations be as transparent aspossible for public opinion, that they provide all neces-sary information about their objectives and their work,and that they listen to the needs and concerns of thisinternational public opinion.

Until now, the OIML did not have any direct com-munication with the public, all dialog went via theCIML Members. In the future, some direct communica-tion on the part of the OIML with the public has to beenvisaged, and a policy must be developed by the CIML

for this. The awareness of governments on metrologyand legal metrology will depend on the awareness ofthe public, and the OIML must help governments toanswer the needs of the public in metrology.


17/45

16

seminar 2020


One can say that OIML was established initially in 1937because the First International Conference on PracticalMetrology which had been convened that year by theFrench government had created a ProvisionalCommittee of Legal Metrology in place of the intendedPermanent International Consultative Committee forPractical Metrology acting as an advisory body to theCGPM (Confrence Gnral e des Poi ds et M esures). Thiswas the proof that at that time it was considered appro-priate to create a new international body, independentfrom the Metre Convention, to deal with legal metrolo-gy. This new body was in fact finally established in 1955and in 2005 we will celebrate the fifty-year anniversaryof the establishment of the OIML.

Some years ago (March 1995) there was a proposalfrom the French government to study the possibility ofmerging the two intergovernmental metrology bodieswhich are located in or on the outskirts of the same city,Paris. After much discussion it was decided that amerger was not appropriate (at least not for the timebeing) but that regular contacts between the two orga-nizations should exist. A joint Metre Convention/OIML

working group was established, and meets annually inFebruary. This group has recently been enlarged inorder to associate ILAC.

The participants in the Metre Convention activitiesare the National Metrology Institutes (NMIs) and themain focus is on national measurement standards. Forlegal metrology, these national measurement standardsare important but are not the primary concern becauselegal metrology is related to other activities. Therefore,the participants in the activities of the two intergovern-mental metrology bodies are quite different, with theexception or perhaps five or six countries for which the

OIML representatives are the Directors (or theirDeputies) of NMIs. I suppose that it is quite clear that

the two organizations have different and well-definedfields of activities. In addition, I would like to repeat myopinion that metrology is not only the science of mea-surements: it also includes specific activities related tomeasurements, this second aspect of the definition of

metrology being close to our legal metrology activitieswhich include type approval testing and verification, aswell as procedures related to metrological supervisionand control.

The Metre Convention bodies (including theInternational Committee of Weights and Measures -CIPM - of which I am a Member) are mainly responsi-ble for the highest level of accuracy and for traceabilityat the level of the national measurement standards,whereas legal metrology is close to the measuringinstruments, their usage and the requirements applyingto such instruments. In fact there is a gap between mat-

ters of traceability and matters of usage of measuringinstruments with no specific international body respon-sible for this part of metrology. So I understand thatthis gap is covered by bodies which are not explicitlyrelated to metrology, e.g. bodies which are close to stan-dardization, certification, accreditation, etc., whichmeans that we are gradually losing our metrologicalposition in this field. Sometimes we are trying to saythat type approval testing is some kind of conformityassessment, and that verification is not a very impor-tant procedure because it is close to calibration - ormaybe it is some type of certification. This is a danger-ous and unacceptable situation for us.

Some years ago, Prof. Kind (who was at that timePresident of the International Committee of Weightsand Measures) made a classification of our activitieswith the following three groups of activities: measure-ment standards, measurement-related regulations, andapplications by users. The widely recognized need forquality of products and services is closer to the applica-tion by users. The classical scheme comprises severalelements:

NMIs, which are responsible for establishing andmaintaining national measurement standards, for

disseminating the size of these units, and for actingas centers for expertise in measurements;

Calibration networks, calibration laboratories andlaboratory accreditation;

Regulations and specifications, including governmen-tal regulations, legal metrology, and voluntary andregulatory standards; and

Users of metrology (including metrological informa-tion, measuring instruments, etc.): these are manu-facturers and other industries, bodies involved in

trade and commerce, health and safety, environmen-tal protection, science, communication, transporta-

Legal Metrology and theMetre Convention

LEV K. ISSAEVCIML SECOND VICE-PRESIDENT,DEPUTYDIRECTOR, VNIIMS, RUSSIAN FEDERATION


18/45

17

seminar 2020


tion, enforcement of government regulations, produc-tion and distribution of energy, military services, etc.

For certain of these activities there exist internation-al bodies: the Metre Convention bodies for units andcalibration, including the CIPM MRA; ILAC is active for

laboratory accreditation, the OIML for type testing andverif ication laboratories, etc. However it is not clearwhere the responsibilities of NMIs stop. It is possiblefor the OIML to be between the NMIs and the users ofmetrology since this field of work may be empty inmany countries and since it involves regional bodieswith which the OIML has good relations. The OIMLshould increase its membership so that all UN countriesparticipate, directly or indirectly, in its activities. Thiscould be achieved through the increased participationof regional organizations in the OIML so that the OIML

might increase its influence worldwide.When comparing the situation of the OIML with

that of the Metre Convention especially as concerns cer-tain trends in our modern world, in my opinion theOIML is in a better position especially considering its

relations with the WTO, since the OIML has observerstatus within the WTO. Therefore, the OIML is closer tothe UN family, closer to the WTO and closer to practicallife.

As a conclusion, it seems to me that its is not possi-ble to envisage a merger between the two internationalmetrology organizations even in the future, since bothhave very well-defined and clear responsibilities.However, it is necessary that the OIML activity fills thegap between users of metrology at international level,thus establishing a worldwide measurement system.


19/45

18

seminar 2020


Most of the work I will present was carried out for theFrench Ministry for Research in order to gain a fore-sight into the 21st century. The first point I will stress isthe speed of change. But we do not believe in the accel-eration of history - let me give an example.

At the time of the French Revolution, for the firsttime hot air balloons could transport people. At lastman could fly, and thus century-long dreams becamereality. One hundred and twenty years later airshipswere invented, and it was only after another hundredyears that the first useful application was conceived inGermany - a cargo-lifter, which was in fact an airship

used as a crane which transformed rescue operationsand was also of use in the construction industry.

So the lead time for concrete technical changes tohappen in this example was over two centuries.Transformations in technical systems do take this long,and when you examine patterns of progression in thepast, this was also the case even 12 or 6 centuries B.C.

During the industrial revolution (which startedaround 1750) there were four major technical poles:materials, energy, time scale and the man-biosphererelationship. Industry began to evolve: combustion wasused to generate energy, the second or 1/10 second

could be measured and Pasteur developed microbiologytechniques at the end of the XIX century.

But this industrial revolution is not yet over on aplanetary scale and there are signs of a new technicalsystem revolution which we call the cognitive revolu-tion. The four poles are changing: for example the timescale is now the nanosecond and will become thefemto-second in some ten or twenty years time. Andbiotechnology is delving deeper and deeper into theexploration of living matter.

So this is a global change - which is a change notonly in technology but also in civilization itself, and

which is evolving from a materialistic mentality (a rem-nant from the industrial age dominated by the belief inscience) to a cognitive civilization which is a relation-ship between subjects and also the recognition process,not merely the knowledge system. It is not an informa-tion society, it will be a recognition society, which isquite different. Of course this leads to the emergence ofvery small enterprises with values based on autonomyand recognition, and of course the infrastructure isgeared around telecommunications.

Networking this civilization requires a much greaterquantity of information than the former one. Spoken

language normally consists of some 60 000 words. Butto describe modern science and technology, some 6 mil-lion references are required, which is one hundredtimes a language. So no expert can totally dominatemodern science and technology. Intelligence at all levelsis necessary and of course there is also the phe-nomenon of the Tower of Babel. We are not living in aninformation society but in a disinformation society,because no one brain can handle all the knowledge andso everyone is a victim of the disinformation processes!

The role of metrologyin a cognitive society

THIERRYGAUDIN*

PRESIDENT OF THE ASSOCIATION PROSPECTIVE 2100

* To contact the Author: http://gaudin.org/


20/45

19

seminar 2020


The classical economic theory is no longer validbecause this theory assumes that there is a need for per-fect information and, in this ocean of information,information is mostly imperfect. The basis affects dayto day life; this is of course a characteristic.

On the Internet one can find medical images model-ing the brain, showing the amount of knowledge whichis immediately accessible. But of course technology alsocreates barriers between human beings. All sciencesboth now and in the past have relied on metrology. Butnow, we have the femto-second system.

The recent Nobel prize-winner Cohen Tanuggidemonstrated that everything is now under the modelof vibration. This changes the way in which we considerthe universe and the way we look at ourselves too. Theother difference is that at the time of the industrial rev-olution, mines and crude oil were the basis on which

industry was built but nowadays, in a cognitive society,measurements are the basic input.

This is the core of my message: industry relies onmines and crude oil, and the cognitive society relies onmeasurements.

Measurements are also necessary for nature becauseconcern over nature is increasing. The second worldsummit in Johannesburg recently illustrated how thingsare developing right now.

Let us also talk about globalization. If we look at aworld map drawn ten years after ChristopherColumbus trip to Cuba and other islands, it shows that

the desire at that time was to make a world map inorder to organize world trade. But the first signs ofglobalization date back to longer ago than that: it wasthe silk way from the Mediterranean region to China.The silk way started during the 6th century B.C. and wasstill operational in the 2nd century B.C. The center wasSamarkand in Uzbekistan.

The second globalization trend concerned the mar-itime field, with Vasco de Gama, Columbus, etc. and thethird trend relates to present-day electronics. But the

idea of globalization is long-standing and dates back tothe Mesopotamian civilization which created metrologyfor trade. They invented trade, accountancy, schools,courts and business, and the first recorded measure-ment inspector lived in 2700 B.C. in Ur in the center of

Mesopotamia.In the agricultural civilization, the territory is land.

In the industrial civilization, it is capital and owner-ship, and in the cognitive civilization, it is intellectualproperty, i.e. patents and copyright. This is an accelera-tion of competition with the rule the winner takes all.You have or you do not have the pattern. This is anacceleration of capitalist concentration in the first stagewhich firms up the forecasts for the next twenty years.

There is another phenomenon: when a new techni-cal system comes along, it marginalizes the work forceof the previous system. A slow period of exclusion start-

ed in the 1980s and worldwide there now exists a veryimportant phenomenon which causes all kinds of disor-der.

The responses can be multiple. The first one wouldbe to create local moneys instead of global moneys(such as exists in Argentina due to the recent crisis).

But in the nineteenth century, when you had thiscrisis and the European revolution of 1848, whatoccurred is that the ruling class started a new policy, avery tough and voluntary policy, with education andpublic works like the Suez Canal and urban develop-ment such as Haussmanns work.

So we can guess that the following years will be ofthat kind. The first stage between now and 2020 will bethe disarray of the show-business society. The secondstage will be education and a public works society. Andmaybe, the third stage will be a creation society at theend of the XXI century.

To sum up my presentation, I would say that theresult is that the transition to a cognitive society will beone from hom o cocacolansisthat we have now to homosapiens.


21/45

20

seminar 2020


In Switzerland we have decided to totally renew ourlegal metrology system and I think that other decisionmakers, metrologists and experts should maybe consid-er the reflections and solutions that we are about toimplement.

The first step when thinking about the future is toknow exactly what our objectives will be. In broadterms, these may be summarized as prot ecti ng peopleagainst the effects of i naccurate measur ements an d elim i-

nati ng techni cal barri ers to trade. I feel sure that thesesame objectives will be still valid in the year 2020. Whatis going to change are the ways and means to reachthem.

The means are adequate legislation and effectiveenforcement by an efficient infrastructure. What pro-tection measures and which level of protection will bedecided, remain a political question. If we study thepresent situation, we cannot avoid the conclusion thatthe existing system has many strong points, but it also

has several weak features, some of which will be brieflymentioned here.

Legal metrology today suffers from outdated regula-tions in the field of trade consisting of too many detailsand requirements which are too rigid and which focussolely on measuring instruments. On the other hand inother fields like health, safety or environmental protec-tion, metrological legislation is either non-existent orhas many large loopholes.

Since legal metrology has expanded or is in processof expanding into many new fields other than trade, itis of paramount importance that the various state

authorities responsible for these areas coordinate theiractions. This coordination is largely lacking today.

Another failure in the present situation is the lack ofdata security. Rough data is increasingly being trans-mitted and evaluated through complex and extendednetworks. This is fine so long as nobody can tamperwith it. But since this is the subject of another paper of

the seminar, I will not expand on it.About the means, with one exception, today we still

only have the procedure to ensure the continuing mea-surement reliability of measuring instruments: this pro-cedure is pattern approval coupled with verification.Although quite adequate for measurements in trade, itis hopelessly inadequate for other areas in which thepeople performing the measurements and the proce-dures are much more important than the instrumentitself. Take for example non-ionizing radiations emittedby antennae of mobile phone networks. The measuredquantity is vectorial, depends on reflections, on mobile

reflecting objects, number of channels used at the timeof measurement, etc. The procedures and experience ofthe staff is much more relevant to correct measure-ments than the instruments themselves.

From that starting point, the question is: what arethe ways and the means to overcome these failures?

We have decided:

to use all the existing competence of state authoritiesand private bodies as soon as their competence canbe proved;

not to restrict legal metrology to the classical field;

to set up a national coordination committee in whichevery state authority having metrological responsibil-ities is represented;

to introduce performance-oriented requirements formeasuring instruments and methods, fully harmo-nized with those of our main trade partners, whichincludes legislation on prepackages; and

to take all necessary measures in order to have all ourmetrological certificates recognized worldwide and torecognize certificates of other countries.

We have also decided to add to the traditional

scheme, type examination and product verification, thenew features of the European Union as laid down in thenew and global approach, and to complement this sys-tem which covers only the production and putting onthe market, by the necessary ways and means to main-tain measurement reliability in all stages of measure-ment activities.

Unlike the classical system in which only one possi-bility of conformity assessment is offered, namely pat-tern approval and verification, the new system offers amodular solution at two phases of the life of measuringinstruments.

Firstly, the manufacturer has the choice of differentmodules in order to establish the conformity of its

Towards total approach inlegal metrology

BRUNO VAUCHER, CIML MEMBER,DEPUTYDIRECTOR, METAS, SWITZERLAND


22/45

21

seminar 2020


instruments before the long-term placing on the mar-ket; these modules are described in the EU Directive onthe global approach. It also has the choice of competentbodies, state or private, which will perform the neces-sary tests and evaluations for it to prove conformity. In

this system, the manufacturer is responsible for confor-mity and this shift from the current preventive systemto a more or less repressive system requires that wehave market surveillance in order to ensure the protec-tion of both people and the environment.

Secondly, there is a choice of different ways tomaintain measurement reliability depending on the fea-tures of the measuring instrument. This includes peri-odical verification and also remote calibration and veri-fication, and a combination of these. The competentauthorities will prescribe which modules are valid forwhich type of instrument used in their field of responsi-

bility.At all steps of this scheme, the severity of the activi-ty required will depend on the risks associated witherroneous measurement results. If the risk is small, theconformity assessment procedure and surveillance willbe simple. If the risk is high, for instance for medicaldosimetry or radiation protection in nuclear powerplants, the procedure will be much more demanding.For that reason, not every module will be available forevery type of instrument. The specific ordinances willprescribe what modules or conformity assessment andwhich level of measurement reliability assurance willapply to a given type of instrument. The scheme is alsoapplicable to measurement methods and proceduressuch as for non-ionizing radiations. In this case themeasurement procedure must be examined andapproved and compulsory comparisons must be per-formed. Moreover, the testing laboratories have to beassessed and/or accredited.

A new surveillance concept will be introduced tocontrol that the new system is correctly enforced at allsteps of measurement activities; this surveillance hasseveral elements. First there is the surveillance of con-formity assessment bodies. The state authority shall notonly assess and notify them, but also control that they

maintain their competence and correctly perform thetasks they have been mandated to carry out. For that itcan rely on accreditation. The surveillance authorityshall check by means of random controls that theinstruments declared to be conform really do complywith the legal requirements at the time they are put onthe market. To achieve that, a centralized informationsystem is required to avoid multiple controls. We alsohave surveillance on the enforcement, whereby wemonitor that the procedures prescribed for maintainingmeasurement reliability are really and completely per-formed in the prescribed time spans. The user is

responsible for this.

The last factor is that the authority shall controlwhether the instruments and the measurement proce-dures are adequate for use and whether they are usedand perform correctly.

A very important feature for surveillance is mea-

surement, which in my opinion is more important thanfastidious checks of documents and certificates. Themain points will be to actively check that the instru-ments measure within their maximum permissibleerrors and that the measurements are reliable. I callthis scheme measuring surveillance.

As of now the state authority will be responsible forthis surveillance. According to the level of risk associat-ed with erroneous results, the authority may delegateall or part of the surveillance to competent third par-ties.

In the above explanations I have tried to show yousteps involved towards total approach. We shall startwith the introduction of new means not only for thecontrol of measuring instruments (including softwareof course), but also measurement methods and, if nec-essary, the measurement actors in order to ensure mea-surement reliability, and this not only for trade but alsoin the new fields. For that, coordination between thestate authorities is a must. We will try to achieve this bysetting up the coordination committee already men-tioned.

I think it is clear for everyone involved that ourtasks and activities will become even much more com-plex and demanding in the future considering the ongo-ing technical developments and the new field of legalmetrology. Therefore, it is a must that all partiesinvolved maintain and develop their competence andcollaborate closely together to reach a transparent, uni-versal and global measurement system and conformityassessment system. Only this will allow us to attain themain objectives of legal metrology outlined at thebeginning of my presentation and I do hope to see oneday, and this before 2020, the merger of the internation-al organizations involved. This will also solve the dis-pute about names we had just a few minutes before:whether it is legal metrology or metrology, or trademetrology.

A final remark: total approach does not mean totalsurveillance or over-regulation. It means appropriate,effective and efficient measures to protect people andthe environment where, and as much, as is needed. Andhere I agree with a statement made yesterday: when itis not necessary to regulate something, then it is forbid-den to do so. For that we need to monitor the outcomeof our activity both with the public and in line with thefeedback of this controlling system, with a view to

increasing, maintaining or reducing our efforts.


23/45

22

update


Issuing Authority / Au to r i tde dli vr an ce

Netherlands Measurement Institute (NMi)Certin B.V., The Netherlands

R60/2000-NL1-02.02Type 0765 (Class C)

Mettler-Toledo Inc., 150 Accurate Way,Inman, SC 29349, USA

This list is classified by IssuingAuthority; updated informationon these Authorities may beobtained from the BIML.

Cette li ste est cl asse par Au to ri tde dli vr an ce; les in for m ati on s

jour relat iv es ces Au to ri ts son tdispon ibl es auprs du BIM L.

OIML Recommendation ap-plicable within the System /

Year of publicationRecommandation OIML ap-plicable dans le cadre duSystme / An ne d'di ti on

Certified pattern(s)

Modle( s) cer ti fi(s)Applicant

Demandeur

The code (ISO) of the Member State in

which the certificate was issued, withthe Issuing Authoritys serial number ifthere is more than one in that MemberState.

Le code (I SO) in dicatif de l'tat M embr eayan t dli vrle cert i fi cat , av ec le numro desri e de lAu to ri tde Dli vr ance si l en ex is-te plus du ne dans cet tat M embre.

For each Member State, cer-tificates are numbered in theorder of their issue (renum-bered annually).

Pour chaqu e tat M embre, lescert i fi cat s son t n umro ts paror dr e de dli vr an ce (cett enu mrot ati on est ann uelle).

Year of issue

An ne de dlivran ce

The OIM L Cert i f icate System for M easurin g Instrum entswas introducedin 1991 to facilitate administrative procedures and lower costs asso-ciated with the international trade of measuring instruments subject tolegal requirements.

The System provides the possibility for a manufacturer to obtain an OIMLCertificate and a test report indicating that a given instrument patterncomplies with the requirements of relevant OIML International Recom-mendations.

Certificates are delivered by OIML Member States that have establishedone or several Issuing Authorities responsible for processing applications

by manufacturers wishing to have their instrument patterns certified.

The rules and conditions for the application, issuing and use of OIMLCertificates are included in the 2003 edition of OIML P 1 OIML Cert i f icateSystem for Measuri ng Instru ments.

OIML Certificates are accepted by national metrology services on a volun-tary basis, and as the climate for mutual confidence and recognition of testresults develops between OIML Members, the OIML Certificate Systemserves to simplify the pattern approval process for manufacturers andmetrology authorities by eliminating costly duplication of application andtest procedures.

Le Systme de Certifi cats OIM L pou r les Instru m ents de Mesur ea tintroduit en 1991 afin de faciliter les procdures administratives etdabaisser les cots lis au commerce international des instruments demesure soumis aux exigences lgales.

Le Systme permet un constructeur dobtenir un certificat OIML et unrapport dessai indiquant quun modle dinstrument satisfait aux exi-gences des Recommandations OIML applicables.

Les certificats sont dlivrs par les tats Membres de lOIML, qui ont tabli

une ou plusieurs autorits de dlivrance responsables du traitement desdemandes prsentes par des constructeurs souhaitant voir certifier leurs

modles dinstruments.

Les rgles et conditions pour la demande, la dlivrance et lutilisation deCertificats OIML sont dfinies dans ldition 2003 de la Publication P 1Systme de Certi ficat s OIM L pou r les Instr um ents de M esur e.

Les services nationaux de mtrologie lgale peuvent accepter les certificatssur une base volontaire; avec le dveloppement entre Membres OIML dunclimat de confiance mutuelle et de reconnaissance des rsultats dessais, leSystme simplifie les processus dapprobation de modle pour les

constructeurs et les autorits mtrologiques par llimination des rpti-tions coteuses dans les procdures de demande et dessai.

Systm e de Cer tifica ts O IM L:

Certificats enregistrs 2 0 0 3 . 0 2 2 0 0 3 . 0 4

Pour des informations jour: www.oiml .org

O IM L Cer tifica te System :Cer tifica tes reg istered 2 0 0 3 . 0 2 2 0 0 3 . 0 4

For up to d a te informa tion: www.oiml .org


24/45


Netherlands Measurement Institute (NMi) Certin B.V.,The Netherlands

R31/1995-NL1-03.01Types NPL12 / N PA 12

Nuovo Pignone S.p.A, Via Roma 32,I-23018 Talamona (SO), Italy



R50/1997-NL1-03.01Type SFB 23e (accuracy cl ass 1 and 2)

Whwa Waagenbau Josef Whrl GmbH & Co.,

hringer Strae 6, D-74629 Pfedelbach, Germany


Physikalisch-Technische Bundesanstalt (PTB),Germany

R51/1996-DE-03.02Type KWD-2 (cl asses X(1) and Y(a))

Wolf Verpackungsmaschinen GmbH, Bettenhuser Str. 3,

D-35423 Lich-Birklar, Germany


National Weights and Measures Laboratory (NWML),United Kingdom

R51/1996-GB1-01.01 Rev. 3Type 8060 (Classes X(1) and Y(a))

Delford Sortaweigh Ltd, Main Road, Dovercourt,Harwich, Essex CO12 4LP, United Kingdom


Centro Espaol de Metrologia, Spain

R60/2000-ES-03.01Type AW514 /02000C (Class C)

Applied Weighing International Ltd., Unit 5, SouthviewPark, Caversham, Reading, Berkshire, United Kingdom

23

update


INSTRUMENT CATEGORYCATGORI E D I NSTRUM ENT

Continuous totalizing automaticweighing instruments (belt weighers)I nstru ment s de pesage totalisateur s conti nu s

fonct i on nement autom at iqu e (peseuses su r ban de)

R 50 (1997)


Diaphragm gas metersCom pt eur s de gaz paroi s dformables

R 31 (1995)


Metrological regulation for load cells (appli-cable to analog and/or digital load cells)Rglem ent at i on mtrol ogi qu e des cell u les de pese(appl i cabl e aux cellu les de pese aff ichage anal o-

giqu e et/ou numri qu e)

R 60 (2000)


Automatic catchweighing instrumentsI nst ru ment s de pesage tri eu rs-ti queteur s

fonct i on nement autom at iqu eR 51 (1996)


25/45

24

update




R60/2000-NL1-03.04Type SCL ... ( Class C)

Precia-Molen, Teteringsedijk 53, NL-4817 MA Breda,The Netherlands

R60/2000-NL1-03.05Type FT1A (Class C)

AEP Technology S.r.l., Via Bottego 33,I-41010 Cognento (Modena), Italy

R60/2000-NL1-03.06TypeOSBH seri es (Class C)

Bongshin Loadcell Co., Ltd., 148, Sangdaewon-dong,Jungwon-ku, Seongnam-city, Geonggi-do, Rep. of Korea

R60/2000-NL1-03.07Type CP seri es (Cl ass C)

MASTER-K, 38, avenue des Frres Montgolfier,B.P. 186, F-69686 Chassieu Cedex, France

R60/2000-NL1-03.08Type FT ( Class C)

Laumas S.r.l., via 1 Maggio n.6,

I-43030 Basilcanova Parma, Italy

R60/2000-NL1-03.09Type 1242 (Cl ass C)

Vishay Tedea Huntleigh International Ltd.,5a Hatzoran St., New Industrial Zone,Netanya 42506, Isral



R61/1996-DE-02.01 Rev. 1Siw arex A, Siw arex AWS for accu racy class Ref (0.2)

Siemens AG Frth, Wrzburger Str. 121,D-90766 Frth, Germany



R76/1992-DE-02.09Types BPW ... and BPWD ... (Class I I I )

Bizerba GmbH & Co. KG, Wilhelm-Kraut-Strae 65,D-72336 Balingen, Germany

R76/1992-DE-03.02Types NT ... (Classes II , II I an d II II )

Bizerba GmbH & Co. KG, Wilhelm-Kraut-Strae 65,D-72336 Balingen, Germany

INSTRUMENT CATEGORYCATGORI E DIN STRUMENT

Automatic gravimetric filling instrumentsDoseuses pon dral es fon cti onnement au tomati que

R 61 (1996)

INSTRUMENT CATEGORYCATGORI E DIN STRUM ENT

Nonautomatic weighing instrumentsI nstr ument s de pesage foncti onn ementnon automatique

R 76-1 (1992), R 76-2 (1993)

Updated informationon OIM L cert ificate s:

www.oiml.org


26/45



R76/1992-NL1-02.38 Rev. 1Type SM -300... (Class I I I )

Teraoka Seiko Co., Ltd., 13-12 Kugahara, 5-Chome,Ohta-ku, Tokyo 146-8580, Japan

R76/1992-NL1-03.02Type DB-I IF ( Class I II )

CAS Corporation, CAS Factory # 19 Kanap-ri,Kwangjeok-myon, Yangju-kun Kyungki-do, Rep. of Korea

R76/1992-NL1-03.06Type DS-500 (Class II I )

Shanghai Teraoka Electronic Co., Ltd.,Tinglin Industry Developmental Zone, Jinshan District,Shanghai 201505, China

R76/1992-NL1-03.07Type SM -500... (Class I I I )

Teraoka Seiko Co., Ltd., 13-12 Kugahara, 5-Chome,Ohta-ku, Tokyo 146-8580, Japan

R76/1992-NL1-03.08Type AD-1 (Classes II I an d II II )

A&D Instruments Ltd., 24 Blacklands Way,Abingdon Business Park, Abingdon,Oxfordshire OX14 1DY, United Kingdom

R76/1992-NL1-03.10Types PS Panda 7 (Class II )

Mettler-Toledo A.G., Im Langacher,CH-8606 Greifensee, Switzerland

25

update


The OIMLCertificate System

for Measuring Instruments

Le Systme d e C ertificats O IML

pou r les Instrume nts de M esure

OIML

P1

Edition2003(E)

OIM L P 1

Edition 2003 (E)

ORGANISATION INTERNATIONALEDE MTROLOGIE LGALE

INTERNATIONAL ORGANIZATIONOF LEGAL METROLOGY

INTERNATIONAL

PUBLICATION

Download the completely

redrafte d 20 03 Edit ion

o f OIM L P 1 f ro m our w e b

site free of charge

OIM L Cert i f icate System for Measuring Instruments

Original Publ icat ion dated 1991

8/12/2019 Oiml Bul

oiml bulletin july 2003

Documents