One of the key points in modern measurement science is the evaluation of measurement uncertainty, according to the definition given by the International Vocabulary of Metrology (VIM) [1] and the procedure recommended by the Guide to the Expression of Uncertainty in Measurement (GUM) and its supplements [2]–[4]. According to these documents, when several sources of uncertainty are present, the related standard uncertainties shall be quadratically combined, and the covariances must be considered, if present [2].

In the previous article, we discussed the importance of uncertainty in measurement results for the metrology teams. In this article, we focus on one step in the calculation of uncertainty: the definition of influencing quantities.

Measurement uncertainty is a value that will quantify the confidence that can be placed in a measurement result. This defines an interval in which the actual quantity has a high probability of being. Without associated uncertainty, a quantity has little meaning. It can not be compared to another or to a reference value. This paper first recalls what measurement uncertainty is. It then summ arises the most widely used method within metrologists’s community to perform the calculations: the GUM (NF ISO/IEC Guide 98-3). Finally, the main uses of uncertainties are discussed.

GUM imposes to take covariances into account in the estimation of measurement uncertainty (8.4). This article provides a practical method to estimate the covariances based on the concept of "high opportunity variance" (HO) and "low opportunity variance" (LO).

Many publications have talked about the measurement uncertainties in recent years. The industry has been dealing with the concept of measuring process capability without any standard for years. Now, the time has come to deal with measurement uncertainties in conformity assessment. Metrologists can now rely on ISO / IEC Guide 98-4 that was released late last year. It provides guidelines and procedures to assess the conformity of a measuring instrument to specifications. Without going through the complex mathematical calculations of the standard, we discuss the "Consumer’s risk," the "Producer’s risk" and discuss new concepts such as "guard banding" of the new standard. The magazine Mesuresinterviewed Jean-Michel Pou, founder and president of Deltamu. He presents the ISO/IEC Guide 98-4 standard and reminds us of some useful truths...

Deltamu has long been promoting the guard bandind strategy in measurement uncertainty consideration for conformity assessment. During the 13th International Congress of Metrology, In 2007, Jean-Michel POU and Dimitri VAISSIERE proposed, a conference on this subject. Earlier in 2003, Jean-Michel POU proposed an article for AFNOR's metrology documentation, which addressed this issue (Article III-20-02: The capability of the measurement process).

Totally convinced that this normative evolution is fundamental for metrology in the business, but can also work against the interests of the industry in the event that it is poorly interpreted, Deltamu decided to make a free Excel application for numerical simulation available for metrologists to calculate:
– The consumer risk and the producer risk, from the characteristics of the manufacturing process, the measurement process and contractual capability
– guard bands to ensure the underlying consumer risk calculation when the previous measurement uncertainty does not respect the contractual capability. This strategy leads to a considerable increase of the producer risk which, in the case of calibration and controls, is also supported by the consumer (see article in Contrôles Essais Mesures -. February 2013)
– Guard bands that minimize the weighted sum of consumer risk plus producer risk which seems the most appropriate way to use this new ISO / IEC Guide 98-4

It should be noted that this free application only takes the normal law into account for Process and Measurement. The strategy used (computer simulation) is perfectly adaptable to any type of laws. Will professionals have the curiosity to go beyond the standard Gaussian case to take full advantage of this strategy? The future will tell. In the meantime, Deltamu offers you the opportunity to exercise at a low cost... Enjoy!

Note: Deltamu has put a training day in place on 2103, May 28th in Clermont -Ferrand. This training program about ISO / IEC Guide 98-4, reference IN05, Title "ISO 98-4:.ISO 98-4 : Risques clients et fournisseurs. Les nouveaux enjeux de la déclaration de conformité is available on our training page. The trainees will learn how to use the application through practical cases.

Everything you need to know before you apply the new ISO/IEC Guide 98-4: 2012. The new ISO/IEC Guide 98-4 standard was published in the end of 2012. Part 4: “Uncertainty of measurement -- Part 4: Role of measurement uncertainty in conformity assessment.” This standard is the result of works on measurement uncertainties during the last two decades. Measurement uncertainty in conformity assessment is not, however, the major concern of the industrial users.

Although it is at the heart of industrial issues, metrology is not recognized at its real value. Despite the work that has been done at international level to standardize vocabulary, bad habits persist, few of us make the effort to use appropriate words...

During the J'M exhibition in Toulon (France), some workshops were held to highlight problems related to measurement. Many people are used to believe the measurements they make are correct and perfectly mastered. But precautions must be taken to get and interpret correctly an "as close to reality as possible" value whatever the measurement domain.

Measurement uncertainty is at the hearth of metrology, but still very hard to grasp. Its definition in the VIM (International Vocabulary for Metrology) may not help us a lot...

This paper discusses the measurement uncertainties of geometrical defects on a coordinate measuring machine. We made an estimation of measurement uncertainties through an interlaboratory comparison performed in accordance with the standard ISO 5725 [1-6]. However, non-normality of the probability distribution for measuring error makes it necessary to adapt the measurement tests values that are described in the standard. We propose an iterative method to exclude laboratories providing aberrant results. We also provide a measurement uncertainty reporting compatible with the actual distributions.

Part 6: Choosing a Measurement process
This last paper discusses consumer's risk and producer's risk and tools that help defining a pragmatic strategy when a contractual capability is out of reach (the measurement uncertainty is too large as compared to the maximum permissible measurement error)

Deltamu develops a metrology management software and provides training and consulting in the metrology business. Its founder, Jean-Michel POU, is an iconoclast personality in the world of metrology. He likes to shake the GUM, the VIM and conventional thinking. In this paper, he describes an approach that distinguishes measurement uncertainty for the instrument from the "background noise" of the calibration process. Although this solution is still theoretical (but already proven), it attempts to answer real questions: those questions that arise in the industrial area.

Part 5: Confrmity assessment This paper goes through ISO 14253-1 standard principles. These principles can be used in all quality management systems.

Part 4: Some numerical applications (2/2) Water flow speed evaluation and calculation of associated measurement uncertainty.

Part 3: Some numerical applications (1/2) This paper provides examples using traditional calculation methods (standard deviations evaluation, sensitivity coefficient...) but also numerical simulation.

Part 2: the use of statistical properties How to use the statistical properties described in Part 1 to evaluate measurement uncertainties? This paper is a new way of dealing with the GUM (Guide to the expression of uncertainty in measurement). It deals with its implications.

Measurement uncertainty is at the heart of the metrologist challenges. More than just managing the calibration of measuring instruments, the metrologist must now guarantee the capability of the measurement processes, that is to say their ability to perform as they are expected to. This series of 6 article aims to help the reader better understand measurement uncertainty issues and capability of the measurement process. This 6 articles series discuss theoretical aspects and provide examples to help the reader get into using the knowledge in his business.
Part 1: A statistical approach of measurement
This article presents a statistical approach of measurement. Using the dice roll example, it helps to understand the tools used for the evaluation of measurement uncertainties.

"To evaluate measurement uncertainty, following the recommendations of the GUM (NF ENV 13005 standard in France), you need to evaluate standard deviation for all of its components. One part of measurement uncertainty is obtained from type B evaluation for which we must associate distribution models, sometimes arbitrarily. The other part is calculated with a type A evaluation. But do we know the imperfection of type A evaluation? One can find several corrections and formulas for the experimental standard deviation estimation in the case of small sets of results. All publications provide different coefficients. Using numerical simulation, it is possible to set the coefficients and the uncertainties associated with the experimental standard deviations."

If the current practice of calendar-based calibration has eventually convinced us that it is a good thing, a brief look at its historical context leads us to a consideration that sheds a totally different light on this issue. First, we will clarify that legal metrology set this practice in place to grant fairness in commercial transactions for both customers and suppliers.

Cet article présente deux guides(1) qui traitent du choix des périodicités d’étalonnage : l’ILAC-G24/OIML D10 publié par l’OIML(organisation légale de la métrologie) et le fascicule FD X 07-014 publié en 2006 par l’Afnor. Une première partie présente les principes techniques proposés par ces guides. Une seconde partie comparera les guides et discutera des avantages et inconvénients des méthodes proposées.

Si la pratique a fini par nous convaincre que les étalonnages périodiques calendaires étaient une bonne chose, il suffit de les replacer dans leur contexte historique pour s’imposer une réflexion salvatrice. Il convient donc, dans un premier temps, de rappeler que la métrologie légale les a inventés pour garantir la loyauté des échanges commerciaux, tant pour les consommateurs que pour les commerçants.

La riferibilità dei risultati di misura è, in teoria, la principale preoccupazione di un Metrologo. Taratura e valutazione dell’incertezza di misura consentono di dimostrare questa riferibilità. Tuttavia, la taratura è valida solo al momento dell’operazione medesima. Dato che l’incertezza di misura dipende in parte dai risultati di taratura, essa è realistica solo se il dispositivo non si modifica (non evolve). Così la questione relativa agli intervalli di taratura è una questione essenziale di riferibilità troppo spesso trascurata. Questo articolo si occupa di questo argomento, illustrando tre interessanti metodi per l’ottimizzazione già considerati nella norma francese FD X 07-014 del 2006.

Measure better to produce better. Optimized measurement processes and instruments management can become a factor of competitiveness.

Metrologists have been using the time base "year" in the industry to answer standards' requirements. They often define the calibration interval to 12 months, sometimes 6. In a tense economic context, there are many good reasons to lower costs of operation.

This paper series summerizes a Metrology document published by AFNOR (French Standards Organization): "Measurement instruments management optimization: A user guide to FD X 07-014 standard's implementation" It contains three parts as there are three methods in the AFNOR's documentation: The OPPERET method, drift analysis and measurement uncertainties analysis.

This third paper presents the measurement uncertainties method.

This paper series summerizes a Metrology document published by AFNOR (French Standards Organization): "Measurement instruments management optimization: A user guide to FD X 07-014 standard's implementation" It contains three parts as there are three methods in the AFNOR's documentation: The OPPERET method, drift analysis and measurement uncertainties analysis.

This second paper discusses drifts analysis

This paper series summerizes a Metrology document published by AFNOR (French Standards Organization): "Measurement instruments management optimization: A user guide to FD X 07-014 standard's implementation" It contains three parts as there are three methods in the AFNOR's documentation: The OPPERET method, drift analysis and measurement uncertainties analysis.

This first paper discusses OPPERET method.

This paper presents the results obtained by applying a method inspired by the Six Sigma approach to evaluate the maximum permissible measurement error for a micropipette set in the real world. The approach significantly enhances conformity assessments ratios, optimizes calibration and verification intervals. This method consequently reduces the overall costs of metrology without lowering quality standards.

OPPERET, for "OPtimisation des PÉRriodicités d'ÉTalonnage" (Calibration interval optimization), is a method that helps calculate the best calibration interval for a measurement instrument. Neither too early to avoid additional costs, nor too late to avoid invalid measurements, this method not only monitors the drift, it also takes risk into account, as well as many other factors (risk, cost, environment of use, ...). This is common sense put into mathematical method.

Leroy Somer's electric motors factory in Saint Symphorien d'Ozon (near Lyon) has lowered its metrology budget by 20%. This has been done by optimizing measurement instruments calibrations intervals. Deltamu has had the opportunity to validate its mathematical tools on a large scale real case (more than 5,000 measurement instruments). We calculated a new calibration interval for each of them. Both the customer and us are now studying tools to make good metrology strategies at the motor's design stages, to optimize other parameters such as maximum permissible measurement errors. The aim is still the same: more profit without lowering quality.

If the word is often used in metrology, it is not defined in the VIM (International Vocabulary for Metrology). No standard defines it formally, although some give the principles such as NF E 02-204.

The 13th International Metrology Congress has just ended up. Deltamu has, again, opened new doors: we discuss measurement uncertainties use in conformity assessment. We show the limits of existing thinking and methodologies (ISO 14253-1 and capabilities) by asking the question as it should be on a pragmatical point of view: "What is the risk that my customer accepts an non-conform product?"
Deltamu proposes to manage processes rather than limit metrology to control through measurements of finished goods.
After Bordeaux's (1999) and Saint-Louis's (2001) conferences dealing with calibration intervals that were the starting points to the AFNOR X 07-014 standard (2006), we may have opened a road for metrology to become a tool for productivity and sustainable development. To be followed up...

The mechanical engineering industry, as a huge user of dimensional measurements, has provided a lot of standards and documentation discussing measurement uncertainty and measurement processes capabilities. This paper presents and analyzes these 'dimensional oriented' tools. It is important to note that the considerations herein are applicable to all measurement domains. The reader can find tools adapted to his specific environment. Note : Please contact AFNOR. to get this paper.

Se da una parte la necessità e la rilevanza delle operazioni di monitoraggio sono sottolineate dai requisiti di Qualità (in particolare dalla ISO 9001, ma anche dalla ISO 10012), dall’altra è curioso notare come la definizione di monitoraggio non sia neanche fornita nel VIM... Forse si tratta semplicemente di una svista, o può anche essere il suggello che questa operazione non viene in realtà considerata “Metrologia”.

Monitoring is supported by quality standards, ISO 9001 among others. It is a recognized method, both necessary and pertinent in metrology. We find it strange that it does not appear in the VIM. Maybe they forgot it, maybe this means that this action is not a "metrologist" one since it does not require complex equipment and techniques?

French Metrology College work group

Calibrations are essential and they are supported by all quality standards. Are calibrations sufficient? They provide in-depth analysis of one measurement instrument in mastered conditions, but the limitation of calibrations is that they make a statement but they rarely anticipate.

Measurement monitoring is also supported by quality standards. It allows monitoring measurement quality on a day by day basis. Many professional describe their practices in the area, from the “stone in the balance“ to the statistical detection of suspicious data.

The French Metrology College proposes its work in a manual with many examples: Order the manual

French Metrology College work group

Calibrations are essential and they are supported by all quality standards. Are calibrations sufficient? They provide in-depth analysis of one measurement instrument in mastered conditions, but the limitation of calibrations is that they make a statement but they rarely anticipate.

Measurement monitoring is also supported by quality standards. It allows monitoring measurement quality on a day by day basis. Many professional describe their practices in the area, from the “stone in the balance“ to the statistical detection of suspicious data.

The French Metrology College proposes its work in a manual with many examples: Order the manual

French Metrology College work group

Calibrations are essential and they are supported by all quality standards. Are calibrations sufficient? They provide in-depth analysis of one measurement instrument in mastered conditions, but the limitation of calibrations is that they make a statement but they rarely anticipate.

Measurement monitoring is also supported by quality standards. It allows monitoring measurement quality on a day by day basis. Many professional describe their practices in the area, from the “stone on the balance“ to the statistical detection of suspicious data.

The French Metrology College proposes its work in a manual with many examples: Order the manual

In a measurement process, the less one instrument impacts the measurment process, the less it is useful to monitor it. When the instrument really impacts the measurement process, it should be calibrated more often. This obvious observation was used leading to decisions to lower calibration frequency by a factor 4 for micropipettes at SANOFI PASTEUR. In the meantime, we reduced non-conformity by 80%. Deltamu has set up a model with numerical simulation to evaluate micropipettes’ influence for each application in the production processes. What we call "tailor-made metrology".

"Well known for its expertise in metrology and statistics, Deltamu enlarges its range of services by offering services in manufacturing processes and control processes management."

Sometimes as a prerequisite, sometimes as part of a project of continuous improvement, interlaboratories comparisons are at the heart of the life of a laboratory. Interlaboratories comparisons seem to be an appropriate tool for improving our practices and know how in guaranteeing the quality of measurements made in a laboratory. Four experts describe the process and discuss its uses.

The NF ENV 13005 standard is now the reference in measurement uncertainties evaluation. It imposes to answer some questions: What causes should we consider? What distribution law should we choose for "B-type" evaluations? Which covariances should we take into account, and how?

By comparing results obtained by applying the GUM and those obtained running an interlaboratories comparison (relying on experimental reality rather than hypotheses), some issues arise...

We published a paper “May the VIM be wrong?” for the 10th International Metrology Congress in Saint Louis in 2001. This paper highlighted that when making a calibration, we do not see the actual instrument error but the measurement error for a particular measurement process which is the calibration process. Then we must try to evaluate the instrument error from this information, as far as it is possible.

L’étalonnage périodique s’est imposé comme le Graal de la métrologie. Par un accord quasi tacite entre auditeurs et audités, le respect des dates de ré-étalonnage est devenu, curieusement et avec le temps, la pierre angulaire des audits. Gare à celui ou celle qui n’aurait pu étalonner/vérifier un instrument dans le délai qu’il s’est lui-même imposé ! S’il est pris en faute, il sera puni d’une remarque, voire d’une non-conformité… L’auditeur, convaincu qu’il oeuvre ainsi dans l’intérêt des clients actuels et « à venir » de son audité, n’hésitera probablement pas à sanctionner le fautif !

La Smart Metrology di Deltamu è una metrologia in grado di adattarsi a tutti i tipi di strutture industriali, dalla PMI ai gruppi internazionali, un’opportunità per passare gradualmente dalla Metrologia degli strumenti alla Metrologia dei processi.

La metrologia sempre più fondamentale per l'azienda competitiva : parlano i protagonisti dell'11a edizione di A&T, la manifestazione italiana delle Prove, Misure e Controlli qualità.

Le mot "métrologie" est connoté, plutôt négativement. À force de la limiter aux instruments, la "métrologie" est devenue synonyme de "gestion d'instruments". À l'heure de l'Industrie du Futur, les auteurs souhaitent redonner un peu de faste à cette discipline, notamment avec les liens Métrologie/Bigdata.

En invitant plus d'une centaine de personnes dans le berceau de la Smart Metrology, à Clermont-Ferrand, Jean-Michel Pou et Laurent Leblond ont officiellement initié le changement dans le monde de la métrologie. Avec comme support les concepts développés dans le livre, Smart Metrology de la métrologie des instruments à la métrologie des décisions, de nombreux industriels ont témoigné de l'intérêt de la Smart Metrology pour leurs organisations.

Si l’on s’en réfère à l’étymologie, «métrologie» signifie «science de la mesure». L’approche des mesures en entreprise devrait donc se situer dans un domaine où la rigueur et la rationalité règnent en maîtres. Et pourtant, force est de constater que,encore aujourd’hui, la métrologie fait l’objet d’un certain nombre de croyances qui ont la vie dure et qui, de fait, perturbent la performance industrielle, tant sur le plan technique que sur le plan économique.

In the last paper, we discussed the definition of influencing quantities, one of the steps of uncertainty calculation. In this paper, we focus on metrological confirmation.

In this paper, we discuss measurement traceability, which is more complex than it seems. Experts explain the basics of traceability by sharing their field experience. Traceability makes sense only if one can have confidence in the measures, and this aspect is throughly discussed here. Our experts also give some ideas to reduce costs.

Measurement is not a standalone need in industry. Measurement is a way to describe what reality of the industry is to control and manage production in order to reach conformity assessment.

Sensors have invaded our daily lives. They also appear more and more in the industry, thanks to wireless communications that greatly facilitate the development of new applications, such as online monitoring. Therefore, the process data volumes increase exponentially. Welcome to a Big Data era. Big Data has been massively discussed and used in other areas like marketing for instance, as to study the behavior of Internet consumers. In metrology where it is a new tool, Big data opens new perspectives and allows us to rethink how to exploit the measured values in metrology area.

With Big Data and Data Mining, measurement and metrology finally have their place at the heart of industrial productivity.

This article first describes the context in which sub-contracting may occur in legal metrology and industrial metrology. It then presents different tools for managing outsourced activities.
Outsourcing of metrology operations. Metrological confirmations, measuring instruments management, measurement standards chaining may seem a simple and open choice. Nevertheless, examining the standards and regulations, we can see that this choice is neither that free, nor that simple...

In the previous paper, we discussed the role and place for documentation in metrology. Among all possible procedures, one must explain the method(s) for defining of the measurement instruments metrological characteristics.

Measurement instruments characteristics are the technical specifications that metrology guarantees through its measurement management system. The most usual characteristics are measurement span and maximum permissible measurement error.

This paper describes methods available to define maximum permissible measurement errors.

Grail Quest will take place in the hyperspace... and metrologists will finally have their word to say...

The purpose of the industry has been to produce (or to copy) objects to sell. The value of a product is defined by the customer. He must want to buy the product (marketing), at the right price (cost competitiveness) and be satisfied (quality). How to succeed?

Fourth paper on metrology in the Industry In the previous paper, we focused on metrology standards. Now we go deeper into metrology documentation and its interrelation with the companies' issues.

Third paper of our series related to metrology in the industry. In previous paper, we discussed the different approaches to set up metrology in a company: the Legal approach using Legal metrology, and the "responsability" approach from industrial metrology. In this latest case, the company must freely determine its own rules to guarantee the critical measurement for a desired product quality. This is critical for product quality. This paper delves further into international texts and standards useful for metrologists.

Second paper on metrology in the industry. In the previous paper, we discussed metrology aim and definition. We summurized the job as "manage a system that guarantees measurement results for being confident". But how has it been incorporated into the standards ? This is what this paper analyzes, with associated approaches and their evolution in time.

This paper defines metrology in the business today in 2013. What is metrology useful for? What are metrology's various activities? How should the metrologist organize it? This paper gives a definition and a meaning to metrologists' business and provides guidance and methods to properly fulfill this function and put it into actions.

Everyone recognizes standardization advantages without being fully aware of it. It simplifies everyday life by standardizing requirements. Let us take the example of power plugs, even if the standards are not the same all over the world: it helps if they are the same when we move. It can also help in the industrial area since the standards benefit from a particular status thanks to quality certification.

This word sounds like the metrologist's leitmotiv, as guided by the auditor's requirements. The auditor sees it as a key element in the world of metrology. Let us remind everyone that neither ISO 9001 nor ISO 17025 give a focus on that topic. Periodicity is not defined in the VIM.

When many models flicker (fossil energy, finance, growth by consumption, services economy…) and when it becomes necessary, even imperative, to regain competitiveness and industrial dynamism, there’s a great opportunity for metrology to play a role.

"Conform": the magic word universally accepted by the auditor and the audited to prove that everything is ok in measurement management.

Calibration: The word has been used many times in the metrology area since quality certification has become a standard. But is there really a consensus on its meaning?

This paper from M. Jean-Michel POU and Thierry CORREVITS is published in the AFNOR toolbox for metrologists. Please contact AFNOR to get a copy.

If limited to administrative management of calibrations and to dogmatic non-conformities as it is often the case today, metrology is far away from answering to the industrial needs. We open the measurement uncertainties debate and discuss functional needs and the role of metrology in the future.

Metrology has grown significantly during the last twenty years, in particular due to quality management systems standards that require it in quality control processes. During the last ten years, the organizations have adapted themselves and we are now in a mature era with more stable and clearly designated metrology function. The COFRAC (French Accreditation Organization) and, more widely, the EA (European Accreditation), have established the rules and put a well-known and efficient laboratories' competencies evaluation system in place. The COFRAC Accreditation is an actual quality guarantee (in France), but does the customer really realize all the benefits from the information in the accreditation documents? In the meantime, measurement instruments' verification market has grown with the creation of many laboratories and services companies in the area. How should the metrology services provider develop its activities to answer the industrial demand, which is produce the required product quality at the best price?

Please contact AFNOR to get this paper.