Metrology in Industry
This page intentionally left blank
Metrology in Industry The Key for Quality
French College of Metrology Series Editor Dominique Placko
First published in Great Britain and the United States in 2006 by ISTE Ltd Translated into English by Jean Barbier Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address: ISTE Ltd 6 Fitzroy Square London W1T 5DX UK
ISTE USA 4308 Patrice Road Newport Beach, CA 92663 USA
www.iste.co.uk © ISTE Ltd, 2006 The rights of the French College of Metrology to be identified as the authors of this work has been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
____________________________________________________________________ Library of Congress Cataloging-in-Publication Data Metrology in industry : the key for quality / edited by French College of Metrology. p. cm. Includes bibliographical references and index. ISBN-13: 978-1-905209-51-4 1. Quality control. 2. Metrology. I. Collège français de métrologie. TS156.M485 2006 620'.0045--dc22 2006003530 British Library Cataloguing-in-Publication Data A CIP record for this book is available from the British Library ISBN 10: 1-905209-51-7 ISBN 13: 978-1-905209-51-4 Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire.
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
Foreword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
Chapter 1. Analysis of the Metrological Requirements Needed to Ensure Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jean-Yves ARRIAT and Klaus-Dieter SCHITTHELM
19
1.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Definition of the objectives . . . . . . . . . . . . . . . . . . . . . . 1.3. Choice of the method of measurement . . . . . . . . . . . . . . . . 1.3.1. Accounting for the selection of the method . . . . . . . . . . 1.3.2. Defining the method and the principle to implement . . . . . 1.4. Choice of the means of measurement . . . . . . . . . . . . . . . . 1.4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.2. Analysis of what is already available . . . . . . . . . . . . . . 1.4.3. Assessment and acquisition of material . . . . . . . . . . . . . 1.4.4. Technical criteria . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.4.1. Basic characteristics . . . . . . . . . . . . . . . . . . . . . . 1.4.4.2. Comportment towards influence quantities. . . . . . . . . 1.4.4.3. Durability of the instruments used . . . . . . . . . . . . . . 1.4.4.4. Homogeneity of the supply of instruments . . . . . . . . . 1.4.4.5. Quality of the supplier’s service . . . . . . . . . . . . . . . 1.4.4.6. Adaptation of the instrument . . . . . . . . . . . . . . . . . 1.4.4.7. Possibility of traceability . . . . . . . . . . . . . . . . . . . 1.4.4.8. Computerization and the speed of taking measurements . 1.4.4.9. Ergonomics . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.4.10. Capability of measuring instruments. . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
19 21 22 22 23 24 24 25 26 27 27 27 27 28 28 28 29 29 29 29
6
Metrology in Industry
1.4.5. Economic criteria . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.6. Grid of the analysis of the choice . . . . . . . . . . . . . . . 1.4.6.1. Stage 1: primary technical requirements (unavoidably necessary) . . . . . . . . . . . . . . . . . . . . . . . . 1.4.6.2. Stage 2: secondary technical requirements (desirable) . 1.4.7. Technical assistance for users of measuring instruments. . 1.4.7.1. The EXERA (France) . . . . . . . . . . . . . . . . . . . . 1.4.7.2. VDI/VDE-GMA (Germany) . . . . . . . . . . . . . . . . 1.5. The traceability of the measurements . . . . . . . . . . . . . . . 1.5.1. The necessity of traceability of the measurements . . . . . 1.5.2. Calibration requirements . . . . . . . . . . . . . . . . . . . . 1.5.3. The selection of standards . . . . . . . . . . . . . . . . . . . . 1.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .
30 31
. . . . . . . . . .
. . . . . . . . . .
31 31 33 33 34 36 36 38 39 42
Chapter 2. Organization of Metrology: Industrial, Scientific, Legal. . . . . Luc ERARD, Jean-François MAGANA, Roberto PERISSI, Patrick REPOSEUR and Jean-Michel VIRIEUX
43
2.1. A metrological organization: why? . . . . . . . . . . . . . . . . . . . . 2.2. Metrology: how?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Scientific and technical metrology . . . . . . . . . . . . . . . . . . . . 2.3.1. The BIPM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2. Results of the international activities . . . . . . . . . . . . . . . . 2.3.3. Regional organizations. . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3.1. EUROMET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3.2. European Cooperation for Accreditaton (EA) . . . . . . . . . 2.3.3.3. Accreditation procedure . . . . . . . . . . . . . . . . . . . . . . 2.3.4. Organization at the national level . . . . . . . . . . . . . . . . . . 2.3.4.1. The Laboratoire National de Métrologie et d’Essais (LNE) . 2.3.4.2. The Italian national calibration system (SNT) . . . . . . . . . 2.3.4.3. The Swiss national calibration system . . . . . . . . . . . . . 2.4. Legal metrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1. Scope of legal metrology . . . . . . . . . . . . . . . . . . . . . . . 2.4.2. The International Organization of Legal Metrology (OIML) . . 2.4.3. The European level . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3.1. European Union harmonization . . . . . . . . . . . . . . . . . 2.4.3.2. WELMEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3.3. Other regional bodies . . . . . . . . . . . . . . . . . . . . . . . 2.4.4. At national level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.4.1. Legal metrology in Italy . . . . . . . . . . . . . . . . . . . . . . 2.4.4.2. Legal metrology in Switzerland . . . . . . . . . . . . . . . . . 2.4.4.3. Legal metrology in France . . . . . . . . . . . . . . . . . . . .
43 45 47 48 50 51 51 54 58 59 59 63 65 67 67 68 71 71 71 73 73 73 74 76
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
Table of Contents
Chapter 3. Mastering Measurement Processes Approach to the Setting up of a Metrology Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marc PRIEL and Patrick REPOSEUR 3.1. What to do at the beginning? . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Goals and role of the measurement management system – metrological function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. The measurement processes . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Conception and development of a new measurement process. . . 3.3.1.1. Analysis of the requirements . . . . . . . . . . . . . . . . . . . . 3.3.1.2. Transcription of the characteristics of the product in “measurand” form or “characteristics to be measured” form . . . . . . 3.3.1.3. The development of a measurement process can be managed as a project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2. Exploitation of a valid process . . . . . . . . . . . . . . . . . . . . . 3.3.3. Continuous improvement of measurement processes . . . . . . . . 3.4. Management of the measuring equipment (metrological confirmation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Analysis of the requirement and selection of the measuring equipments . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1.1. Technical requirements . . . . . . . . . . . . . . . . . . . . . . . 3.4.1.2. Economic and commercial conditions. . . . . . . . . . . . . . . 3.4.1.3. Assessment of the measuring equipment . . . . . . . . . . . . . 3.4.2. Receiving the measuring equipment and putting it into service. . 3.4.2.1. Compliance with the order . . . . . . . . . . . . . . . . . . . . . 3.4.2.2. Identification of the measuring equipment . . . . . . . . . . . . 3.4.2.3. Inventory (description). . . . . . . . . . . . . . . . . . . . . . . . 3.4.2.4. Technical dossier of the equipment . . . . . . . . . . . . . . . . 3.4.2.5. Technical documentation . . . . . . . . . . . . . . . . . . . . . . 3.4.2.6. Basic definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3. Calibration and verification operations . . . . . . . . . . . . . . . . 3.4.3.1. Calibration or verification program . . . . . . . . . . . . . . . . 3.4.3.2. Calibration or verification intervals . . . . . . . . . . . . . . . . 3.4.3.3. Supervision of the measuring equipment . . . . . . . . . . . . . 3.4.4. Fitness for use of measuring equipment. . . . . . . . . . . . . . . . 3.4.4.1. Freedom from bias, repeatability, stability . . . . . . . . . . . . 3.4.4.2. Maximum permissible errors . . . . . . . . . . . . . . . . . . . . 3.4.4.3. Demands for an assurance of the quality . . . . . . . . . . . . . 3.5. Setting up a metrological structure within the firm . . . . . . . . . . . 3.5.1. Analysis of the metrological requirements and setting up standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2. Traceability of the measuring instrument(s) to the firm’s reference standards . . . . . . . . . . . . . . . . . . . . . . . . 3.5.3. Traceability of the firm’s reference standards to the SI. . . . . . .
7
79
.
79
. . . .
80 86 86 86
.
87
. . .
87 88 88
.
89
. . . . . . . . . . . . . . . . . . . .
91 91 93 93 93 94 94 94 94 94 94 97 99 99 100 100 100 101 101 102
.
102
. .
104 104
8
Metrology in Industry
3.6. Suggested approach for setting up a metrology function . . . . . . . . . 3.7. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
105 106
Chapter 4. Handling of a Bank of Measuring Instruments . . . . . . . . . . Jean-Yves ARRIAT
109
4.1. Acquaintance with the bank . . . . . . . . . . . . . . . . . . . 4.1.1. Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2. Identification . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Metrological policy of the firm . . . . . . . . . . . . . . . . . 4.2.1. Objective and commitment of the firm’s management . 4.2.2. Plan of actions to launch. . . . . . . . . . . . . . . . . . . 4.2.3. Awareness, training and vocabulary . . . . . . . . . . . . 4.2.4. Selection of the material to be followed periodically . . 4.3. Drafting of the documents . . . . . . . . . . . . . . . . . . . . 4.3.1. Codification of the documents . . . . . . . . . . . . . . . 4.3.2. Work instructions . . . . . . . . . . . . . . . . . . . . . . . 4.3.3. Result-recording documents . . . . . . . . . . . . . . . . 4.3.4. Other documents . . . . . . . . . . . . . . . . . . . . . . . 4.4. Physical handling of the measuring instruments . . . . . . . 4.4.1. Receipt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2. Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2.1. Traceability . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2.2. Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2.3. Precautions. . . . . . . . . . . . . . . . . . . . . . . . . 4.4.3. Storing and environment. . . . . . . . . . . . . . . . . . . 4.4.4. Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Follow-up of the measuring instruments over time . . . . . 4.5.1. Periodicity of the follow-up . . . . . . . . . . . . . . . . . 4.5.2. Campaign of recall . . . . . . . . . . . . . . . . . . . . . . 4.5.3. Follow-up of the results . . . . . . . . . . . . . . . . . . . 4.6. Software for the handling of the means of measurements .
. . . . . . . . . . . . . . . . . . . . . . . . . .
110 110 110 113 113 113 113 114 115 115 116 117 118 119 119 120 120 120 121 121 122 123 123 124 125 125
Chapter 5. Traceability to National Standards . . . . . . . . . . . . . . . . . . Luc ERARD and Patrick REPOSEUR
127
5.1. Introduction. . . . . 5.2. Definitions . . . . . 5.2.1. Traceability . . 5.2.2. Calibration . . . 5.2.3. Verification . . 5.3. Traceability chains 5.4. Traceability . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . .
127 127 127 128 129 129 131
Table of Contents
5.5. Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1. Calibration in an accredited laboratory . . . . . . . . . . . . . . 5.5.2. Calibration in a non-accredited laboratory . . . . . . . . . . . . 5.6. Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1. Verification in an accredited laboratory and in its accreditation scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.2. Verification in a non-accredited laboratory or out of the accreditation scope. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7. Use of calibration and verification results . . . . . . . . . . . . . . . 5.7.1. Use of the results of a calibration . . . . . . . . . . . . . . . . . 5.7.2. Use of the results of a verification . . . . . . . . . . . . . . . . . 5.8. Particular cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.8.1. “Self-calibrating” or “self-gauging” measuring instruments. . 5.8.2. Complex instruments in which components/equipments and software are narrowly combined and large measurement ranges are covered for complex quantities. . . . . . . . . . . . . . . . . . . . . . . 5.9. Metrology in chemistry and physical methods of chemical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.9.1. Traceabilty in metrology in chemistry. . . . . . . . . . . . . . . 5.9.2. Influence of the principle of the method . . . . . . . . . . . . . 5.9.2.1. Absolute methods. . . . . . . . . . . . . . . . . . . . . . . . . 5.9.2.2. Relative method. . . . . . . . . . . . . . . . . . . . . . . . . . 5.9.2.3. Comparative method . . . . . . . . . . . . . . . . . . . . . . . 5.9.3. “Documentary” traceability . . . . . . . . . . . . . . . . . . . . . 5.9.4. Control of the reference materials . . . . . . . . . . . . . . . . . 5.9.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.10. Assessment of traceability . . . . . . . . . . . . . . . . . . . . . . . 5.11. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
132 132 132 133
. . .
133
. . . . . .
. . . . . .
133 133 134 134 135 135
. . .
136
. . . . . . . . . . .
. . . . . . . . . . .
136 137 139 139 140 140 141 143 145 145 146
Chapter 6. Calibration Intervals and Methods for Monitoring the Measurement Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Patrizia TAVELLA and Marc PRIEL
149
6.1. Normative requirements . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. Methods for monitoring the instruments in use – general criteria . 6.2.1. First method: metrological redundancies . . . . . . . . . . . . . 6.2.2. Second method: checking the coherence of the results . . . . . 6.2.3. Third method: “monitoring standards” and statistical supervision of the measurement processes . . . . . . . . . . . . . . . . 6.2.3.1. Statistical control of the measurement processes . . . . . . 6.2.3.2. Control charts . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.3.3. Use of the monitoring methods. . . . . . . . . . . . . . . . . 6.3. The determination of the calibration intervals . . . . . . . . . . . . 6.4. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
. . . .
9
. . . . . .
. . . . . . . . . . .
. . . .
. . . .
. . . .
149 150 150 151
. . . . . .
. . . . . .
. . . . . .
152 152 154 157 158 161
10
Metrology in Industry
Chapter 7. Measurements and Uncertainties . . . . . . . . . . . . . . . . . . . Marc PRIEL 7.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2. Measurement of physical quantity . . . . . . . . . . . . . . . . . . . . . 7.3. Analysis of the measurement process . . . . . . . . . . . . . . . . . . . 7.3.1. The cause and effect diagram method . . . . . . . . . . . . . . . . . 7.3.2. Using the list published in the GUM (section 3.3.2) . . . . . . . . 7.3.3. Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.4. Cutting down the errors . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.4.1. Cutting down random errors by repeating measurements . . . 7.3.4.2. Cutting down systematic errors by applying corrections . . . . 7.4. Modeling of the measurement process . . . . . . . . . . . . . . . . . . . 7.4.1. Measurement procedure and model of the measurement process . 7.4.2. An essential stage for the assessment of uncertainty: modeling the measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5. Assessment of the uncertainty of the input quantities . . . . . . . . . . 7.5.1. Type A methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.2. Type B methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.3. Comparing the Type A and Type B methods . . . . . . . . . . . . 7.6. Calculating the combined uncertainty on the result . . . . . . . . . . . 7.6.1. Situation when all the input quantities are independent . . . . . . 7.6.1.1. Situation when the input quantities are independent and the model is a sum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.1.2. Situation when the model is a product . . . . . . . . . . . . . . 7.6.2. Situation when the input quantities are dependent . . . . . . . . . 7.6.2.1. Assessment of the covariances by assessing a coefficient of correlation r(xi, x j ) . . . . . . . . . . . . . . . . . . . . . . 7.6.2.2. Assessment of the covariances by calculating the terms of covariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.2.3. Assessment of the covariances by considering the terms common to two input quantities . . . . . . . . . . . . . . . . . . . 7.7. Use of the performances of the method (repeatability and freedom of bias) to assess the uncertainty of the measurement result . . . 7.7.1. Intra- or interlaboratory approaches . . . . . . . . . . . . . . . . . . 7.7.2. Intra-laboratory approach . . . . . . . . . . . . . . . . . . . . . . . . 7.7.3. Interlaboratory approach. . . . . . . . . . . . . . . . . . . . . . . . . 7.7.4. Data processing for intra- and interlaboratory approaches . . . . . 7.7.4.1. Assessment of the repeatability and the reproducibility . . . . 7.7.4.2. Assessment of the freedom of bias (trueness) . . . . . . . . . . 7.7.4.3. Evaluation of the linearity . . . . . . . . . . . . . . . . . . . . . . 7.7.4.4. The terms ∑ c u (x ) . . . . . . . . . . . . . . . . . . . . . . . . . 2
i
i
i
163
. . . . . . . . . . .
163 164 166 166 167 168 169 170 171 172 172
. . . . . . .
173 174 175 176 179 180 180
. . .
180 181 181
.
181
.
181
.
181
. . . . . . . . .
183 184 185 186 187 187 188 189 189
Table of Contents
11
7.8. Reporting of the measurement result . . . . . . . . . . . . . . . . . . . . . 7.9. Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.10. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
189 190 193
Chapter 8.The Environment of Measuring . . . . . . . . . . . . . . . . . . . . Jean-Yves ARRIAT and Marc PRIEL
195
8.1. The premises . . . . . . . . . . . . . . . . . . . . . . . 8.1.1. Ambient temperature . . . . . . . . . . . . . . . . 8.1.2. Relative humidity . . . . . . . . . . . . . . . . . . 8.1.3. Handling of the air conditioning systems . . . . 8.1.4. Power network . . . . . . . . . . . . . . . . . . . 8.1.5. Radioelectric disturbances. . . . . . . . . . . . . 8.1.6. Measurements on-site . . . . . . . . . . . . . . . 8.2. The personnel . . . . . . . . . . . . . . . . . . . . . . 8.2.1. The connection of metrology function . . . . . 8.2.2. Staff involved in the metrology function . . . . 8.2.3. The qualification of the personnel . . . . . . . . 8.3. The documentation . . . . . . . . . . . . . . . . . . . 8.3.1. Filing of the documents . . . . . . . . . . . . . . 8.3.1.1. Documents dealing with the quality system 8.3.1.2. Records regarding quality . . . . . . . . . . . 8.3.2. Management of the documents . . . . . . . . . . 8.4. Bibliography . . . . . . . . . . . . . . . . . . . . . . . 8.5. Appendix . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .
196 197 198 199 199 199 200 200 200 201 202 202 202 202 203 204 205 206
Chapter 9. About Measuring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Claude KOCH
209
9.1. Preliminary information . . . . . . . . . . . . . . . 9.1.1. Physical quantity . . . . . . . . . . . . . . . . . 9.1.2. The object to be measured. . . . . . . . . . . . 9.1.3. Field of measurement . . . . . . . . . . . . . . 9.1.4. Four types of uses of measuring instruments. 9.1.5. Influencing quantities . . . . . . . . . . . . . . 9.2. Choice of a measuring principle. . . . . . . . . . . 9.2.1. Differential measurement . . . . . . . . . . . . 9.2.2. Direct measurement . . . . . . . . . . . . . . . 9.2.3. Indirect measurement . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
209 209 210 210 211 212 213 214 214 214
12
Metrology in Industry
9.3. Practicing in metrology . . . . . . . . . . . . 9.3.1. Implementing the instruments . . . . . 9.3.2. Precautions before measuring. . . . . . 9.3.3. Measurements . . . . . . . . . . . . . . . 9.3.4. Variations and their sign. . . . . . . . . 9.3.5. The time factor . . . . . . . . . . . . . . 9.4. Expression of the results . . . . . . . . . . . 9.4.1. Graphs . . . . . . . . . . . . . . . . . . . 9.4.2. Histograms . . . . . . . . . . . . . . . . . 9.5. What qualities does a metrologist require? 9.5.1. Be inquisitive . . . . . . . . . . . . . . . 9.5.2. Be tidy and methodical . . . . . . . . . 9.5.3. Be open to doubt . . . . . . . . . . . . . 9.5.4. Be observant . . . . . . . . . . . . . . . . 9.5.5. Be honest. . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
215 216 216 216 217 218 218 220 220 221 222 222 222 222 223
Chapter 10. Organization of Metrology at Solvay Research and Technology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . José MONTES
225
10.1. Presentation of the company . . . . . . . . . . . . . . . . 10.2. Organization of the metrology sector . . . . . . . . . . . 10.2.1. Creation . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.2. Missions . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.3. Organization . . . . . . . . . . . . . . . . . . . . . . . 10.2.4. Geographic localization of the activities . . . . . . 10.2.5. Composition of the bank of measuring equipment. 10.3. Metrology . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.1. Identification . . . . . . . . . . . . . . . . . . . . . . . 10.3.2. Connection of the standards . . . . . . . . . . . . . . 10.3.3. Periodicity of the calibrations . . . . . . . . . . . . . 10.3.4. Calibration operations . . . . . . . . . . . . . . . . . 10.3.5. Documentation of the calibration results . . . . . . 10.3.6. Verdict of the metrological confirmation . . . . . . 10.3.7. Indication of the state of the calibrations . . . . . . 10.3.8. Personnel and subcontracting . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
225 226 226 226 226 227 227 228 228 228 229 229 230 231 231 232
Table of Contents
Chapter 11. Metrology within the Scope of the ISO 9001 Standard . . . . . Philippe LANNEAU and Patrick REPOSEUR 11.1. Introduction . . . . . . . . . . . . . . . . . . . . 11.2. Introduction to the evolution of the standard 11.2.1. The concept of continuous improvement 11.2.2. The process approach. . . . . . . . . . . . 11.3. Measurement control process . . . . . . . . . 11.4. The ISO 9001 (2000) standard step-by-step . 11.5. Conclusion . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
233 234 234 235 236 238 245
Chapter 12. Training for the Metrology Professions in France . . . . . . . . Bernard LARQUIER
247
. . . . . . . . . . . . . .
. . . . . . .
. . . . . . . . . . . . . .
. . . . . . .
233
. . . . . . .
12.1. The metrology function in a firm’s strategy . . . . . . . . . . . . 12.2. Metrology profession . . . . . . . . . . . . . . . . . . . . . . . . . 12.2.1. Metrological engineer . . . . . . . . . . . . . . . . . . . . . . 12.2.2. Metrological technician . . . . . . . . . . . . . . . . . . . . . 12.2.3. Metrological operator. . . . . . . . . . . . . . . . . . . . . . . 12.3. Initial training. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.1. Schools for engineers . . . . . . . . . . . . . . . . . . . . . . . 12.3.2. Courses for higher level technicians . . . . . . . . . . . . . . 12.3.3. Vocational high schools . . . . . . . . . . . . . . . . . . . . . 12.4. Continuing education . . . . . . . . . . . . . . . . . . . . . . . . . 12.5. Long-lasting training courses . . . . . . . . . . . . . . . . . . . . 12.6. The teaching of metrology in secondary schools . . . . . . . . . 12.7. Prospects for the development of long-lasting training courses 12.8. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
13
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
247 248 249 249 250 250 250 251 251 251 253 265 265 266
The Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
267
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
269
This page intentionally left blank
Preface
Metrology is an essential part of the infrastructure of today’s world. It enters into our lives in a multitude of ways, some direct and some indirect. National and international trade increasingly require demonstrated conformity to written standards and specifications and mutual recognition of measurements and tests. The economic success of most manufacturing industries is critically dependent on how well its products are made, a requirement in which measurement plays a key role. Navigation and telecommunications require the most accurate time and frequency standards. Human health and safety depend on reliable measurements in diagnosis and therapy and in the production and trade in food and food products. The protection of the environment from the short-term and long-term destructive effects of industrial activity can only be assured on the basis of accurate and reliable measurements. Global climate studies depend on reliable and consistent data from many disciplines often over long periods of time and this can be assured only on the basis of measurements traceable to measurement standards that are themselves linked to fundamental and atomic constants. Metrology is not an activity that is only carried out in specialized institutes or calibration laboratories. In order to meet the needs of society for accurate and reliable measurements in all its many applications, a strong spirit of metrology must also exist in companies and enterprises that make the instruments and that use them to make measurements. For this reason I welcome this book. It gives a clear outline of the basic ideas of metrology, why we need it and how, in an enterprise it can be practiced. I wish it every success. T.J. Quinn, Director of BIPM
This page intentionally left blank
Foreword
Technically, economically, commercially and, sometimes, statutorily speaking, having relevant and reliable results of measurements, analyses and tests is a real asset for a firm which wishes to make efficacious decisions. You cannot achieve such an end if you do not have firm control over the processes of measurement, analysis and testing. Nowadays, however, the measuring techniques, the normative and statutory requirements, the methods of measurement uncertainty assessment or those to secure the traceability of measurements are all complex and it is more necessary than ever to integrate them into a network of competent bodies so as to exchange experience and information. It is on this fundamental principle that the Metrology College was created in 1986, which became the French College of Metrology in 2002. The purpose of this association is obviously much wider: – to identify which firms and organisms’ needs are to be met from the angle of metrology; – to spread metrological culture and knowledge through the industrial, scientific and economic fabric; – to be a form of exchange between people involved in metrology; – to contribute to make the collective national and regional actions coherent in this sphere; – to perform any action likely to contribute to the development and promotion of metrology. The permanent evolution of metrology, together with the willingness to impart all the knowledge acquired so far, have led a working party of the French College of Metrology to write a second edition of the book Metrology in the Firm. Metrologists from various callings (national metrology laboratories, accrediting organisms,
18
Metrology in Industry
industrial concerns and consulting firms) and from different nationalities make up this working party. This broad range of authors gives the book a pragmatic characteristic and enables it to answer the questions and concerns of organizations, whether they be principals, small or medium firms, laboratories, etc. The contribution from foreign authors gives the book an unquestionable international aspect which accurately reflects the current reality. More than ever, as a matter of fact, metrology contributes to the free circulation of goods between countries, thanks to the international organization of metrology and thanks to the international agreements between national metrology laboratories and between accrediting organisms. Moreover, most of the authors belong to different national or international standardization committees. As a result, the latest normative evolutions are to be found in this book, whether it is the concept of firm certification developed in the 2000 version of standard ISO 9001, or the approach concerning the competence of activities of measurement, testing or analysis as expounded in standard ISO 17025. Whether you are involved in your firm’s metrology function, or are simply interested in a concrete matter of measurement, analysis or testing, I am confident you will find here some clues which will help you progress and improve your processes. The growing interest you have shown in this book has encouraged us in our intention of producing this English version. It is my sincere wish that whatever your need and country may be, you can get as much out of it as our French colleagues do. May you enjoy reading it. P. LEBLOIS, President of the French College of Metrology
Chapter 1
Analysis of the Metrological Requirements Needed to Ensure Quality
Anybody with a mind to implement (or improve) a metrology function might feel a bit panicky at the thought of all the work to be done if they read this book unwarned, and more particularly this chapter. Let the reader’s mind be put at ease first. All the content is not, fortunately, to be carried out literally. All we want to do is to offer as broad as possible a survey of the subject by pointing out practically all the items that require consideration. And then, is it not normal to start wondering what one really needs? Experience has taught us, too often alas, that this is not a natural process. Many industrial difficulties, or many costs, grow out of the inadequacy “means of measurement/real need”.
1.1. Introduction Before we start any concrete action, it is primordial to analyze the metrological needs carefully. There are two kinds: – The organizational needs for the management of metrology. Are those needs great enough to require the introduction of full-scale metrology? Are premises or qualified personnel needed permanently? What possibilities are there in the region? Chapter written by Jean-Yves ARRIAT – Ascent Consulting – and Klaus-Dieter SCHITTHELM – Expert in Metrology, Germany.
20
Metrology in Industry
Does someone want to manage metrology on his or her own, with the help of a someone else, or to handle it to a subcontractor? – The material needs for the realization of the measurements. In order to realize measurements correctly, it is necessary to have appropriate means; these means are found after analysis of the objectives and the possibilities of the instruments and the connection. In order to define the firm’s needs, it is necessary to answer the following questions: 1. What are my industrial needs? – What do I have to measure and what accuracy shall I expect? 2. How can I meet my needs? – What are the possible measuring methods? – Which method and principle will be used? 3. Which measuring instruments can be used? – Which instrument shall I use? – Can the selected instrument ensure the required accuracy? 4. How is to be used the selected instrument? – What assembly is to be set up and what procedure is to be followed? – What technical competence do you have to have to use it? Then a question of a very different magnitude arises: how am I going to guarantee the quality of my measurements? Setting up a metrological function The three key components of a metrological function have to be under control (see Chapter 4): – adequacy of means to needs; – traceability of the means of measurement to international standards; – administrative management of the equipment (measuring instruments, standards, etc.). The preliminary analysis of the needs will produce a first set of specifications. There is a good chance that these analysis are going to be a bit theoretical and take little heed of the notions of profitability. You have to accept the principle which says that the specifications will evolve and obtain agreement from the major actors taking part in the drafting of the specifications. For a new measuring instrument, all the stages from conception to utilization must be taken into account by the specifications. This is fundamentally the concern
Analysis of the Metrological Requirements Needed to Ensure Quality
21
of the manufacturers, but potential users may sometimes take part in the elaboration of the specifications. The specifications for a new measuring laboratory must ignore all of the environmental characteristics of the measurement (see Chapter 8), and take into consideration the problems of maintainability (for instance, the maintenance of air conditioning), of access to the personnel, of user-friendliness, etc. However big or small the problem is, one must always begin by analyzing one’s real meterological need.
1.2. Definition of the objectives The metrological function must be approached as soon as you start thinking about problems of measurement. Its role may depend on each particular firm (see Chapter 3), but its chief role is to act as a consultant. It examines the need in a logical process based on: – the functional analysis of the measurement (drafting of specifications); – the analysis of the achievement of the measurement results (and of the level of accuracy reached); – the analysis of the risks related to the selected means; – the analysis of the non-conformities which could be encountered. This process makes it possible to identify and quantify the means (personnel and material) to be implemented to take the intended measurements. It is during these phases that the “tools of quality” will be used. Let us point out that the analysis of the value (fundamental at the outset) is among the most useful tools. In order to clearly define the objective, we strongly recommend to use “brainstorming”, cause/effect diagrams, Pareto, etc., which make analysis and collective participation easier. So as to guarantee the quality of its measurements (i.e. a process of management by quality), the firm sets up a real management of the means of measurement. For this purpose, the metrological function conducts the management of these means according to needs that are clearly defined and regularly updated. This involves examining a large number of actions in order to start up and maintain the supply of measuring instruments necessary to meet the firm’s needs.
22
Metrology in Industry
The first thing to do regarding the analysis of the supply of material is to work out: – the list of physical quantities (e.g. temperature, length, electric resistance, etc.); – the ranges which need to be covered for each physical quantity (e.g. length from 0.1 mm to 1,000 mm); – the permissible uncertainty for each quantity and each range (the uncertainty in the 0.1 mm to 0.5 mm range will be different from the one which is expected between 100 mm and 1,000 mm). Then, for each separate case, it will be necessary to consider and define: – the analysis of the needs and the choice of the means of measurement; – the acquisition, the reception and the implementation of these means; – the traceability of the material of measurement (in the case where materials of measurement are assigned); – the traceability of the measurements (which material do they come from?); – the calibration or the verification of the means and the decisions they entail; – the exploitation of the calibration results; – the operations related to the moving of these means (protection, authorization, etc.); – the updating of the inventory of these means. The outcome of this is that the intended objectives must not be mixed up to satisfy: – the needs for the management of metrology with; – the needs for the realization of the measurements.
1.3. Choice of the method of measurement 1.3.1. Accounting for the selection of the method You have to justify the choice of the selected method. It is to be understood by this that the criteria have to possess as little subjectivity as possible. This choice must take possible restraints of qualification into consideration. The fact is that within the scope of some contracts (notably related to safety, public security, health, etc.) you may have to qualify the method of measurement. This means it must be subjected to an authenticated description, officially certified tests, etc., in accordance with the relevant program and by a very precise process. Besides, the ISO/QS 9000 or TS 16 949 certification process also involves a description of the selected method.
Analysis of the Metrological Requirements Needed to Ensure Quality
23
Fortunately, it is often possible to hang on to the methods which are known and officially accepted. You must not forget that the great metrology laboratories can be a great help in this area. In France, for example, these are the laboratories of the LNE (Laboratoire National de Métrologie et d’Essais), and in Germany, those of the PTB (Physikalisch-Technische-Bundesanstalt), or calibration laboratories accredited by the DKD (Deutscher Kalibrierdienst). Whether the method is qualified or not, it is important, after the metrological objectives have been set, to make the methodology of the measurement explicit. The different stages, the conditions of the material and the environment, the operations that make it possible to get the measurement, i.e. everything related to the carrying out of these measurements, must be written in a document and will be taken into account particularly when choosing the operators. One of the very first principles of quality assurance is to write down what is being done. This process is simple and allows people to think further about the choice of the method. There must be a clear distinction between chosing a method and chosing a measuring instrument. For example, you may want to measure a dimension on a rubber part: you happen to be close to a three-dimensional measuring machine and your instant reaction may be to go to this machine without thinking whether there may be a more suitable method than this one. 1.3.2. Defining the method and the principle to implement When there are several methods of measurement, it is often difficult to determine which one will best fit your need if you are not able to classify them. Our advice is to keep only the two (maybe three) most important criteria in mind and to draw a table. Let us consider the example of Table 1.1. It makes it possible to analyze the different methods of measurement that lead to the assessment of the characteristics of industrial robots. Two criteria have been selected: – the principle of measurement (two groups of them here); – the characteristics measured (two families of them here). As a rule, there are in metrology three great principles of measurement; the three of them have advantages and drawbacks. They are: – differential measurement; – direct measurement; – indirect measurement. See Chapter 9 for more details.
24
Metrology in Industry
Local methods
Big base methods
Positioning characteristics Trajectory characteristics Measurement terminal with cubes (Peugeot SA and LNE) Measurement terminal with Measurement terminal on materialized trajectories measuring machine (IPA) (rule and circle) (LNE) Different realizations based on the Measurement terminal with trajectory same principles have been (broken line) (Peugeot SA) developed (IBM, General Motors, etc.) Devices with three sensors and wire (Peugeot) Sweep of two laser beams (University of Surrey, England) Selspine system Robotest (Polytech, FRG) Stroboscoped photogrammetry (University of Dresden, NEL and SETP-LNE) IPA: Institute for Production techniques and Automation, Germany LNE: National Testing Laboratory NEL: National Engineering Laboratory, England SETP: Photogrammetric Studies and Works Society Method of the two theodolites (Renault) Theodolites with automatic data (LNE) Selspine system Photogrammetry (University of Dresden, NEL and SETP-LNE)
Table 1.1. “Classification of the methods of measurement” (Reproduced with the kind permission of Techniques de l'ingénieur – France)
1.4. Choice of the means of measurement 1.4.1. Introduction The choice of the material and/or the equipment must be based on specifications. To make this choice, you must take into consideration: – the technical needs; – the possibilities of calibration; – the assessments already made; – the economic conditions (last, for the technical specifications have to be seen first).
Analysis of the Metrological Requirements Needed to Ensure Quality
25
Practicing metrology is not simply doing plain measurements. To begin with, a special material has to be used, which means that you do not simply use any dimensional comparator lying about on a shelf, you do not borrow a frequency meter from a colleague and you do not hire a “lowborn” multimeter. On the contrary, you use instruments which are well-known and well-regarded, which come with documents and certificates, so as to be sure of their traceability and, therefore, to better guarantee the quality of the measurements. These instruments (said to be “reference instruments”) have to be acquired after you have seriously studied the criteria of choice. It is known that: – the ideal instrument does not exist; – the instrument closest to what is ideal is too expensive; – each buyer limits the claims of technical applicants. Moreover, the choice of an instrument depends on its type of use. Four types of utilization can be distinguished: – for a study (you must look for an instrument that can evolve); – for a site (robustness ought to be favored); – in manufacturing (the “cost” factor will probably prevail); – for a laboratory (your preference will go to a very reliable, strong and proven instrument). For further information, see Chapter 9.
1.4.2. Analysis of what is already available The first thing to do will be to see if there is not already in the firm some available material which can meet your needs. This requires: – good communication between the various parties concerned with the measurements; and – a good knowledge of the material available. The latter point is all the more important when there is a risk of technological obsolescence (using a state-of-the-art instrument to its maximum capacity justifies its acquisition and it makes it easier to get new ones), or when the material is very expensive (when you increase the duration of its productive use, you make its amortization easier).
26
Metrology in Industry
1.4.3. Assessment and acquisition of material Speaking of compromise about the choice was actually slightly simplistic. Of course, the economic requirements are obviously taken into account; few are the cases when the material is selected without the price being considered (either before or after the purchase!). As for the assessments which are otherwise made, they quite simply depend on the competence and professionalism of the person in charge of the metrological function. He must indeed be on a permanent technological watch. Furthermore, he must make an inventory of what is in store (material and tested material), in order not to have to repeat work endlessly. The companies which take the trouble to check all the electric and electronic material they buy admit that a far from negligible proportion of the instruments delivered is partly defective or does not comply with tolerances on delivery. A few years ago a survey showed that the percentage of rejected instruments could reach 50%. This is partly explained by the fact that the stated characteristics are obtained by the manufacturers, in a laboratory and in ideal conditions of use; and this situation is very remote from the user’s reality. Tests of assessment preliminary to purchase would be greatly recommended. However, in frequent cases, the instruments that can perform the same function are many in number, the parameters of each of them are numerous and, consequently, the tests are long and expensive. So, before launching into testing, any person who is interested in purchasing an instrument is entitled to ask the salesman the following questions: – Have any tests been done? If the answer is yes, when? Where? By whom? In which domain? Is a report of the tests available? – How long has the instrument been manufactured? How many copies of it have been produced? – Has stopping its production been considered? – Who has bought it? Is it possible to consult users? Once you have got this information, and if tests seem necessary, you have to choose between doing them yourself or subcontracting them to a better-equipped organization whose results cannot be questioned. A distinction must be made between learning about a instrument which is presented by a salesman and having its characteristics verified by a specialized laboratory. Once again, evidence arises of the importance of good relationships (partnership even) with the manufacturers of the instrument and of their obligation to pass on information in a transparent and unrestricted way. However, the role of the buyer is not simple. He must estimate whether the supplier is capable of keeping to the agreed times in general: time of delivery, time of assistance after the sale. Besides, it
Analysis of the Metrological Requirements Needed to Ensure Quality
27
seems to be of paramount importance that the team responsible for maintaining the instruments, as well as the users, should be involved in choosing the instruments they need for their activities. In essence there are three reasons for this: – Because of their experience, the user and the maintenance team know the little details, which make all the difference (and those which mostly “hinder” the smooth progress of their work). – They get used more easily to equipment they have helped to choose (working and utilizing conditions are improved: that is what is called communicating without demagogy!). – They are not so easily influenced by attractive advertising, or by purely economic criteria, which makes the overall analysis more objective. So, economic conditions and assessments generally being what they are, we find ourselves left with technical criteria. The following are those that seem to be the most important.
1.4.4. Technical criteria 1.4.4.1. Basic characteristics For a measuring instrument (whether used as a standard or not) this most often means that its necessary accuracy is in one certain domain of the studied quantity in ideal conditions, said to be reference conditions: a temperature of 20°C or 23°C, 230V/50Hz power from the mains, no mechanical and electrical perturbations, etc. 1.4.4.2. Comportment towards influence quantities This concerns the way the basic characteristics change with time according to external constraints: variation of the temperature or the electric power, electromagnetic perturbations, vibrations, etc. The way instruments react over a period of time is often undetermined. As a rule, on-off cycles are more harmful than a long, uninterrupted, working period. Contrary to a widespread opinion, all instruments (even the very accurate ones, the expensive ones, etc.) are liable to drift in time. They have to be recalibrated or reset regularly. 1.4.4.3. Durability of the instruments used The durability is the interval of time during which the instrument remains capable of meeting your normal need of it. It must not be mistaken for the longevity, which defines the lifespan, generally speaking, of the instrument, even if the instrument no longer meets your need.
28
Metrology in Industry
A material is durable if it is both reliable (few breakdowns) and maintainable (easy to repair). The information provided by the maintenance teams allows us to have good facts upon which to make a decision. The most accurate metrological instruments are expensive and, as such, you have to be able to use them for a sufficient length of time. So, you should prefer the makes with good durability; higher investments having sometimes to be considered. You have to estimate how much longer the instrument will be manufactured or maintained. In addition, is this instrument “open” to future evolution? Is there any assurance that it will be compatible with the next generation of equipment? 1.4.4.4. Homogeneity of the supply of instruments You must avoid having too many different types of equipment and material: if you have equipment of similar types, maintenance will be less costly, you will know your material better, the supplies of spare parts will be cheaper, there will be a possibility of interchangeability in case of a breakdown, periodic calibrations or verifications can be automated, etc. 1.4.4.5. Quality of the supplier’s service Your relationship with the supplier of instrument must not stop with the purchase. You must analyze the technical assistance the supplier can provide. Have provisions been made for the setting up of the instrument, for clear explanatory documents (utilization, maintenance, intervention, etc.), in the language of the country where it will be used, or at least in English? How is the supplier able to help if problems occur, and how long, on average, will he make you wait? The more sophisticated the instrument is, the more these questions matter. Placing an order with a instrument dealer may, sometimes, save time, but there is actually nothing that can replace communication with the manufacturer. As a matter of fact, there are few dealers who have a good knowledge of the instrument they sell, or who attend to the training of the users. It is very often difficult to go beyond the stage of purely commercial advertising. 1.4.4.6. Adaptation of the instrument It is advisable to get instruments which have been conceived with a “metrological” outlook; i.e. instruments adapted in their principle and realization to the needs of metrologists. For example, all metrologists who work in the timefrequencies scope have “major oscillators”, which are excellent sources of 5 or 10MHz. It is therefore eminently desirable that any synthesizer or frequency meter should be able to work either on its internal oscillator or on an external signal. The best-equipped “frequency” laboratories possess a caesium clock, or at least a
Analysis of the Metrological Requirements Needed to Ensure Quality
29
rubidium clock, from which a 10MHz signal is drawn and distributed in the firm in order to synchronize frequency meters and synthesizers. 1.4.4.7. Possibility of traceability When you buy a measuring instrument, you have to raise the question of traceability to national or international standards before you eventually make up your mind to proceed with the purchase. Is it or is not possible to relate your measurements validly to the accepted standards at the national or international level? The question of traceability is developed in 1.5 below. 1.4.4.8. Computerization and the speed of taking measurements There is a technical parameter which has a direct consequence on the cost of quality to the firm: how fast it will be to take a measurement? The question, and its answer, is as much about how quickly the instrument can provide the necessary information as about the transcription of the measurement in a simple form. A digital display offers ease of reading and can, in the case of the vernier calliper for example, reduce by a factor of five the time it takes to take measurements. In addition, it may be important to computerize the measurement. Computerization makes it possible: – to increase the speed at which measurements are obtained by decreasing the input time; – to increase the Quality Assurance by reducing the risk of making mistakes while, for example, writing the results out by hand; – to incorporate the measuring instrument into a computerized “Statistic Process Control” (SPC). Of course, computerization is possible on adaptable instruments, for instance digital display instruments which have an outlet to connect to an RS 232 plug. These remarks refer, in particular, to those instruments which are used on sites or in production, and also in metrology laboratories. 1.4.4.9. Ergonomics Several types of instruments can be selected for a specific measurement. However, some will turn out to be less “handy” to implement. The ergonomic aspect of the utilization must not be forgotten: ease of handling, utilization by a left-handed person, integration into the work surface, bulk and weight, etc. 1.4.4.10. Capability of measuring instruments This is a very important parameter that people in charge of metrology and people who use measuring instruments must keep in mind. The “capability” of the
30
Metrology in Industry
measuring instrument is an indication which is the extent to which the instrument makes it possible to assess whether, and to what extent, the measuring system fits with the tolerance that is being checked. The measuring system includes the measuring instruments (the material), applied measurement processes (the methods) and the personnel who do the measuring, that is to say the users (the person). To put it another way, it is about whether the prescribed interval of tolerance properly fits with the overall uncertainty of measurement. Choosing too effective a means would result in a superquality which would lead to too high a price. On the other hand, a lack of effectiveness would bring about an unacceptable percentage of defective parts being manufactured. Who amongst us has not had to struggle with too strict intervals of tolerance, which are hard to comply with in manufacture, and also in measurement? What is the good of striving to get a result to the hundredth of a unit (0.01 volt for example) when the dispersion of a series of measurements is already equal to one tenth of this unit? You need to take into consideration the limits (and the cost) of the measuring instruments to be used to check the technical specifications (intervals of tolerance) when you choose the instruments. Consequently, the choice of the instrument depends on the tolerance to be verified. You have to clearly delimit the uncertainties of measurement that will appear when you use the material. The French standard NF E02204 (which concerns mechanical engineering, but which can serve as a basis for other purposes) provides very useful supplementary information and definitively repeals the widespread “10%” rule. In production, the capability index (whole or by centering) is given by the following formula: Cp = [upper tolerance - lower tolerance]/6 s with s = standard deviation of the series produced Cpk = MIN [ (upper tolerance - mean)/3 and (lower tolerance - mean)/3 s] In metrology, the capability index of the means of measurement (Cmm) is often: Cmm = IT/6 Ig with IT = interval of tolerance (from specifications) Ig = overall uncertainty of the measurement
1.4.5. Economic criteria For reasons that are the very bases of the metrological function, it is necessary to practice metrology with well-known measuring equipment. It is possible to reckon
Analysis of the Metrological Requirements Needed to Ensure Quality
31
how much a measurement costs, but this does not mean anything unless all the parameters of the cost are taken into account: – the purchase price of the material and its resale price after it has been used a certain number of years; – the costs of operation (expenses for operating the material, usually the lowest cost), electric power, recording paper, accessories; – the cost of maintenance (including calibration, and preventive and corrective maintenance); – the cost of lack of availability: will a replacement material be needed while it is being maintained? Will there be any financial consequence? These different parameters are interdependent; automation increases the purchase price, but it reduces the operating cost. High reliability also increases the purchase price, but it cuts down the cost of maintenance. 1.4.6. Grid of the analysis of the choice There are two stages when you select a measuring instrument. 1.4.6.1. Stage 1: primary technical requirements (unavoidably necessary) The point is to determine the quantities, the ranges of measurement and the uncertainties which should be found in the instrument so that you can get the expected quality of instrument. The outcome of this stage will be a list of the instruments available on the market which meeting the technical requirements. 1.4.6.2. Stage 2: secondary technical requirements (desirable) It will be possible at this stage to make a decision based on the results of outside evaluations, and taking commercial and economic conditions into account. Here is a tool to help thinking with the decision-making: a good mind of “Management of Quality” will always try to use practical tools. We suggest that you make a list of the criteria to consider when choosing an instrument, then to attribute to each criterion a coefficient depending on how important each criterion appears to be, and then a mark. The items on this grid should come from the analysis of the criteria undertaken by the manager of the metrological activities (the person in charge of the metrological function in the firm), the user, the buyer and the personnel responsible for the maintenance. Each person’s opinion will thus be taken into account. The important thing is to make a careful list of questions and provide an answer to each one. It is true that experience is not easy to weigh, but the object of this method is just to provide a starting point to work out a decision (Table 1.2).
32
Metrology in Industry
The various people who are concerned with the instrument should meet to determine the values of the weightings. The role of these weightings is to give more weight to one or several items of the grid which, according to the group, have a certain importance. The final mark for each item is obtained by multiplying the mark of the item by the associated weighting (n*c). The weightings c (Σc) are added, then the products c*n (Σc*n) are added. The evaluation of the measuring instrument is obtained by the division: Σc*n -----Σc ∑c = ________
∑c*n = ________
Identification = Type = Manufacturer = Technical – homogeneity of the supply of instruments needs – risk of rapid obsolescence – documents from the supplier – technical assistance – adaptation of the instrument to technological requirements – etc. Outside – evaluation from a centre accredited by the evaluations COFRAC or the DKD – evaluation by users (EXERA, etc.) – experience gained on similar material of the same make – press-cuttings from the specialized press – etc. Economic – cost/price of the competitor’s range and – possibilities of purchase or loan commercial – required time for delivery conditions – time allowed for repair – etc.
Coef. c
Table 1.2. Grid of the evaluation of a measuring instrument
Note n
c*n
Analysis of the Metrological Requirements Needed to Ensure Quality
33
1.4.7. Technical assistance for users of measuring instruments In some countries organizations have established themselves to provide users of measuring instruments with technical assistance. Two examples of such organizations are given below. 1.4.7.1. The EXERA (France) This chapter, which deals with the analysis of metrological requirements, would be left unfinished if no mention was made of the EXERA, one of the few associations which work to support industrial metrology. The EXERA is a non-profit-making association, an amalgamation of companies and organizations that are major users of instruments and systems of measurement, regulation and automation. Since its foundation, in 1970, its purpose has been to produce and circulate original information and to provide its members with assistance when they need to express their requirements, to choose, to install and to operate materials and systems. The EXERA is first and foremost a club; it is a privileged meeting place for users, where specialists (over 500) can freely exchange what information about what experience has taught them, as well as information about instruments and systems. This club acts, in essence, through its members by organizing the technical evaluation of materials. It also initiates the writing of guides about the choice of material in the different technical areas and, at the same time, does its best to develop a constructive dialogue with manufacturers. In a spirit of partnership, groups of users are constituted so that they can take responsibility for their needs and they can better express and defend them in front of manufacturers. This enables the users and the manufacturers to obtain more elements of explanation on investments and technological trends. There are technical commissions about automation, instruments, analyzers, measurements and systems for the tests, etc. In 1982, the EXERA signed an agreement of international cooperation with two other organizations of users: – the SIREP (Britain); and – the WIB (the Netherlands). These two other associations have members in other industrialized countries, for example, the USA, Canada, Japan, Finland, Sweden, Belgium, Switzerland, etc.
34
Metrology in Industry
The main features of the agreement, which ultimately concerns more than 100 large companies, are: – the full-scale and well-balanced exchange of assessments of instruments and surveys, which are all written in English; – the acceptance of common principles regulating the procedures of evaluation and the presentation of the documents; – the harmonization of the work programs; – the gradual adjustment of the formalities regulating the testing of materials. Altogether, there are about 90 members in the three associations; 40 of the members belong to the EXERA, among them are: CEA, CGE, EDF, GIAT, IFP, L'OREAL, PECHINEY, RENAULT, TOTAL, etc. At present, approximately 80 reports are distributed annually by the three associations. In December 1991, the SIREP, the WIB and the EXERA were officially recognized by the European Organisation for Conformity Assessment (EOTC) as “Agreement group”. For more information, see www.eotc.bc or www.exera.com. 1.4.7.2. VDI/VDE-GMA (Germany) In Germany an organization similar to EXERA is the Society for Measurement and Automatic Control GMA (Gesellschaft Mess- und Automatisierungstechnik). This organization is a joint organization of the Association of German Engineers VDI (Verein Deutscher Ingenieure) and the Association for Electrical, Electronic and Information Technologies VDE (Verband der Elektrotechnik, Elektronik und Informationstechnik). The GMA is a network of technical competence in metrology and other fields of activity. It combines expertise of institutions such as the German National Metrology Institute (PTB), the German Calibration Service (DKD), the German Institute for Standardisation (DIN), the International Organization for Standardization (ISO) and several industry associations and societies. GMA activities include: – the promotion of the exchange of information between industry, public authorities and scientific institutions; – the organization of congresses, conferences, symposiums, etc. to promote the flow of information concerning new processes and developments; – the preparation of publications, recommendations, guidelines, etc. to improve understanding; – the scientific preparation for standardization;
Analysis of the Metrological Requirements Needed to Ensure Quality
35
– the national and international representation in the field of measurement and automation controls; – the publication and promotion of technical and scientific literature; – the support of education and post-graduate training. In common technical committees, honorary experts of industry, research and science cooperate in different fields of metrology. Each committee is focused on specific branches of metrology. These committees produce newly-developed or updated technical documents. These documents are first presented as drafts. Views and comments of potential users are evaluated and the documents are modified before they are definitively published. The guidelines published by VDI/VDE-GMA describe standards, e.g. in metrology. These metrology documents define procedures for users of measurement instruments (see the following table). Metrological level
Guidelines, documents and standards
National Metrology Institute (PTB)
National DIN standards or DKD guidelines
DKD accredited calibration laboratory
International EN or ISO standards EA documents
Optional: In-house calibration laboratory
VDI/VDE guidelines
Measurement and testing equipment
DKD guidelines
Product
EA documents
Table 1.3. Metrology literature used in Germany
In the VDI/VDE guidelines there are three series dealing with the treatment of measuring equipment: The series VDI/VDE/DGQ 2618, “Inspection of measuring and test equipment – instructions to inspect measuring and test equipment for geometrical quantities”, contains general considerations and determinations, as well as information on the expression of uncertainty in measurement. In separate documents there are procedures for calibration and surveillance of specific-measurement instruments. An example of such papers is the paper about the procedures for “Callipers for external, internal and depth dimensions”.
36
Metrology in Industry
Another series, the VDI/VDE/DGQ/DKD 2622 guidelines, deals with “Calibration of measuring equipment for electrical quantities”. Again, along with a general introduction including information on the measurement uncertainty, there are separate documents defining the calibration of specific, electrical measurement instruments. An example is the calibration procedure for electrical oscilloscopes. The third series, VDI/VDE 2617, is entitled “Accuracy of coordinate measuring machines – parameters and their reverification”. The calibration, the acceptance and the surveillance of coordinate measuring equipment is defined in separate documents. This series is used as a base for the development of a new ISO standard on coordinate measuring machines. More detailed information is available at the GMA secretariat in Düsseldorf, Germany (e-mail:
[email protected]).
1.5. The traceability of the measurements It has to be said repeatedly: the calibration requirements and the traceability define the quality of the measurements. The metrological function is responsible for the management of the quality of the measurements. This has to be taken into consideration from the beginning of the process that leads to the selection of the method of measurement, and then the means of measurement.
1.5.1. The necessity of traceability of the measurements Traceability is the very basis of metrology. What good is it to take measurements if the measurements do not mean the same thing to everybody? For example, let us look at the measurement of the value of the “foot” in the past. Until about the 18th century (and even later), the “foot” was used as a unit to measure distances. Everyone used the same word. A worthy sample of this quantity was available to avoid arguments, such as “is it a child’s foot, or a woman’s, or a man’s?” The problem was that when the value was translated into the metric system, it gave the following results: – foot of the King of France 32.48 cm – Roman foot 29.63 cm – foot from Bordeaux (South of France) 35.70 cm – foot from Lorraine (East of France) 28.60 cm – foot from Vienna (Austria) 31.50 cm
Analysis of the Metrological Requirements Needed to Ensure Quality
37
These discrepancies resulted from the lack of a national reference (let us not even talk of a European one), and of local comparisons to each reference. Just imagine the Airbus today manufactured from all parts of the world. It is therefore indispensable to have metrological references in one’s firm and to have them compared to national reference quantities by calibration. Comparisons between accredited laboratories are made by national accreditation bodies (the COFRAC in France, the Deutscher Kalibrierdienst (DKD) in Germany) and there are programs of comparison that make it possible to ensure that the standards of different countries are related. It is to be regretted that all the industrialized countries are not at the same level of progress in metrology. However, such European countries as Britain, France, Germany, Italy, Spain, etc. are the leaders. It has been said above that it is important to have reference standards in one’s firm and to have them calibrated in accredited calibration centers or laboratories. However, a choice must be made between having the metrology integrated in the firm and having it subcontracted. As some providers of calibration services propose to calibrate the measuring instruments with standards of their own, you need to be careful. You must absolutely make sure that: – their standards are periodically calibrated in a competent laboratory (whose organization complies with the ISO 17025) accredited by a national organization (COFRAC, DKD, UKAS, etc.); – the provider of the service can guarantee the quality of the measurements provided. An audit of the provider’s system of management of the quality will probably be necessary. You have to be able to demonstrate full traceability of the measurement that has been made, the relationship between the measurement and the instrument used, and also the traceability of the firm’s instrument, in order to show that the chain of calibration has not been broken. In addition, do not forget to verify that at every stage the uncertainties of measurement are not too large. Bringing in a provider of services who has their own accredited laboratory is not a must. However, if the provider has one, it is further evidence of his seriousness and commitment to his job. There is every reason to think, that a provider with an accredited laboratory knows better what the word “metrology” means than a competitor who does not have any accredited laboratory. The provider with the
38
Metrology in Industry
accredited laboratory can grasp the primary technical needs of the client: quantities, scope of measurement and uncertainties. Furthermore, the investment required for launching an accredited laboratory excludes the “transitory” type of company that starts up in “commercial niches” and then vanishes as quickly as it has appeared. Stability is a key word in metrology. It is important not to change your provider too regularly when you decide to subcontract the calibrations; for example, do not consider only the price and have a yearly competition. Nevertheless, let us point out that what has been said so far applies to movable measurement, control, test or analysis instruments. In the case of equipment such as heavy machinery (traction, compression, hardness, etc.), scales, air conditioning chambers, etc., the verification can only be done on-site. It is not necessary for the provider to have their own laboratory since the whole intervention is carried out onsite. However, the provider must use working standards which are related to the calibration chains.
1.5.2. Calibration requirements Several problems come to mind when thinking of calibration. First of all, how can a particular measuring instrument be calibrated? If it is a calliper, you will think about using gauge blocks. Has anyone even considered measuring rods for a micrometer? What is to be done with dynamometric spanners, balances, etc.? If you go into physical chemistry, etc. it gets even more complex! Some methods of measurement demand equivalent methods of calibration. Fortunately, some manufacturers of materials provide tips. When you look deeper into the matter, you realize that quite often you talk about calibration, but what you actually need is a verification, perhaps even a metrological confirmation (see ISO 10012 standard). Therefore, it might be necessary to proceed to an internal checking between two interventions, which is just a simplified examination of good working order. Calibration must be done intelligently, which means doing just what is necessary; it is not only a means to avoid auditor’s critical views. How many firms, which work in mechanical engineering and have their sets of gauge blocks calibrated in an accredited calibration laboratory simply open their calibration certificate?
Analysis of the Metrological Requirements Needed to Ensure Quality
39
Never must it be forgotten that the major purpose of calibration is to verify the measuring instrument and calculate the uncertainties that go with the results of the measurements taken with that instrument. The question of the interval of the calibrations inevitably arises quickly. The answer, which should make everybody happy, is that it depends. Some methods of measurement meet a few demands, particularly in the field of physical chemistry. In any event, measuring instruments should be calibrated reasonably frequently, so as to detect and prevent any possible drift, but not too often because of the overall cost involved. On the question of follow-up interval, the reader’s attention is drawn to Chapter 6, as well as to the handbook of documentation published by AFNOR on the subject of the surveillance intervals. The reader should wary of any person who claims that they can tell which intervals are the right ones. As a matter of fact, you always start quite randomly and then, with experience, you define the necessary intervals more accurately. There is the question of subcontracting the calibration; it is not cheap regardeless of whether you do it yourself or subcontract it. It is our opinion that a compromise can be considered. In fact, even though the metrology is not the firm’s chief activity, it is a part of the “Management of Quality”. If you retain part of it in the firm, it makes it possible to maintain the user’s awareness of the importance of the measuring instruments, of the notion of connected uncertainty, etc. However, a firm cannot excel in everything and it must avoid spreading its resources too thinly. It is always possible to ascertain whether there are any local providers of services in metrology and, if so, their charges.
1.5.3. The selection of standards The content of this technical paragraph does not concern all firms; the small- or medium-sized firms that do not use many standards (merely a set of gauges or masses for example) need not worry. What is presented here is a practically complete line of thought which can reveal useful for the firms with a metrology service. However, let us first recall the definition of the word “standard” in the “International Vocabulary of basic and general terms in Metrology” (ISO document, 1993):
40
Metrology in Industry Standard: “material measure, measuring instrument or measuring system intended to define, realize, conserve or reproduce a unit or several known values of a quantity to transmit them by comparison to other measuring instruments.”
Examples: 1 kg mass standard 100 ohm standard resistor standard ammeter gauge block For a given metrological quantity, the standard will be the “reference” of the firm. The standards may, or may not, differ from the usual measuring instruments. The standards of the lowest orders often have the same shape as the standards of usual instruments. They are selected according to their type and their individual characteristics. Thus, they will have to be differentiated from the other usual measuring instruments because they will not have the same assignment, calibration or verification. Consequently, the mode of management concerning them, choice, identification and conservation of the references, will have to be clearly defined. The management of the standards will have to take into account: – the metrological level of the standard; – the technical level and the complexity of the standard; – the abilities of the users; – the assignment of the standard (reference, work standard, etc.); – the importance of the standard for the firm; – special cases of utilization. All this information must be described in simple and accessible documents, because it concerns the references of the measurements made by the firm. An error made on a standard can have more serious consequences than one made on a measuring instrument. When you select a standard, you have to take metrological, technical and economic aspects into account. The metrological aspects are about the following: – the methods that can be used to compare the measuring instrument submitted to calibration to the standard, as well as to calibrate or verify the standard itself; – the assessment of the results of the measurements made with the standard; – the basic metrological characteristics of the standard, that is, the accuracy, the stability and the metrological reliability.
Analysis of the Metrological Requirements Needed to Ensure Quality
41
Note 1: the accuracy of a standard is established either: – by comparing it to a reference standard of a superior order and of the same quantity; or – by assessing its errors using methods and means of measurement (for example, standards of other quantities, calibration devices, etc.) that make it possible to preserve the compatibility of the standard with the national standards. Note 2: you realize a standard is stable and metrologically reliable: – by studying the working principle, the conception and the structure of the standard and coming to an opinion about them; – by scrutinizing the materials that make up its structure, the method of manufacture and assembly; – by studying the registers (monitoring cards, etc.) containing the detailed information about the standard. Note 3: the metrological reliability is the ability of a standard to fulfill its expected function while maintaining the required freedom from bias and repeatability during a predetermined period of time and in set conditions. Besides these basic characteristics, other metrological characteristics can be important in certain cases, for example: – measuring range or nominal value, total or partial, in the case of a material measure; – reference conditions; – reading security; – sensitivity; – linearity or maximum permissible error of reversibility (hysteresis); – dynamic metrological characteristics, etc. The technical aspects are about: – ease of use, simplicity and reliability of the standard; – ease of transport, of taking to pieces and putting together again, of installation, of connection and of setting up in the calibration or verification device; – protection against deterioration, pollution, interferences, etc. either when the standard is being used or when it is just being preserved; – special accessories necessary for the utilization or the preservation of the standard (installation, reading, recording, electric power, etc.).
42
Metrology in Industry
The economic aspects are about: – the price of the standard and its accessories; – the cost and the interval of the calibrations (including the costs resulting from non-availability) during the calibrations; – the cost of its utilization, its maintenance and its preservation; – the possibilities of repairing, and the lifespan; – the qualification of the personnel needed. If you consider the restrictions imposed by the prescribed metrological characteristics, selecting a standard can be regarded as the pursuit of an optimum solution. Yet, in practice, there are even other restraining factors, for example: – the order of standardization of the equipment; – the absence of national or international instructions for some models of standards; – the trend towards the automation of measurements and calculations; – the influence of traditions, etc. Most of these factors have overall effects (metrological, technical, but also economic) and are liable to considerably restrain the choice.
1.6. Conclusion Today, more than ever, the firm focuses its attention on its particular activity. In relation to its metrological function, it must focus its energy on its primary responsibility: the analysis of the requirement, the selection of the materials and the authentication of their metrological capability. Doing that requires data that no one else possesses. The periodical follow-up and the administrative management are somebody else’s affair; a quick economic survey will, most of the time, show that having these activities carried out externally is less expensive – just add up the investments (initial and periodical), the training, the drafting of the procedures and the periodical calibration of the reference standards. The reality is that doing metrology, that is, trying to give meaning to the results of a measuring instrument, is a full-time job which requires you to be independent. However, one must not forget the necessity to compare the specifications (tolerances) on the measured parameters to the uncertainties of measurements of these parameters.
Chapter 2
Organization of Metrology: Industrial, Scientific, Legal
2.1. A metrological organization: why? The authors have purposefully devoted the first chapter to the analysis of metrological needs. The reason for this choice is simple. People’s needs for measurements of all kinds and the necessity to be sure of their reliability and their universality have given rise to metrology, the science of measurement. It is only through satisfying the needs of industry that metrology finds its raison d’être, whether at the international or national levels, or at the very core of each firm. Therefore, the metrological organization could only comply with the rules that make it possible to meet these needs, and in the modern day to anticipate these needs; of course, this task falls to the metrologists who intervene at the scientific, technical and industrial levels. Thus, an intra-firm, national and international coherence of measurements is achieved. In short, metrology is, and it must remain, a universal language. It is easily understood that a universal language involves a certain amount of dialogue between people from different ethnicity, hence its elaboration may seem
Chapter written by Luc ERARD – Laboratoire National de Métrologie et d’Essais (LNE), Jean-François MAGANA – Organisation Internationale de Métrologie Légale (OIML), Roberto PERISSI – ENIQ/Italy, Patrick REPOSEUR – Comité Français d’Accréditation (COFRAC), Jean-Michel VIRIEUX – METAS/Switzerland.
44
Metrology in Industry
laborious. For a long time, talking of quantities or units sounded more like a babel of languages than a modern means of communication. From very early days, trade required measuring instruments and thus standards. The need for universal and unified measurements made it necessary to establish an independent organization which would guarantee the fairness of exchanges that were affected by deep-rooted economic, political and social realities: localization of the exchanges, economic and political interests, not to mention the various national habits and customs which are the hereditary enemies of metrology. It needed powerful triggers to change these customs. The scientific developments of the 17th and 18th centuries prepared the ground for the French Revolution to create the metric system. In spite of political vicissitudes, the industrial developments of the 19th century, which increased needs tenfold, accelerated the process of establishing a metrological organization. It emerged from concepts which will be studied later on in the chapter. It is interesting to quote Lavoisier, who said that: “never has anything greater, simpler, more coherent in all its parts come out of man’s hand.” Although the essential notions of coherence and simplicity influenced the creation of the metric system, it was not adopted in France, as the only and compulsory system, until 1st January 1840. A statement made at that time by the Minister for Commerce is still relevant today: “if man’s needs are something permanent that cannot be modified by a law, his habits are not, they are mere accidents that can be defeated and dominated after more or less time, more or fewer efforts ...” Expressing the real needs, and fighting poor practices, is one of the missions of a metrology organization. The metric system medal, stamped in 1840, commemorating the law of 4th July 1837, has on one side “To all times – To all peoples”, and “Unity of the Measurements” on the other. This states the need that was felt very early in the world of industrial measurement, the need to collaborate regardless of political differences and, in addition, to establish and use a coherent and universal system. If the word “need” is a dominant recurring theme, it is because metrology is not reserved for isolated, initiated people in their ivory towers. Each day, it is constantly resorted to, often unknowingly, for tasks that are regarded as commonplace. Measuring is closely related to any human, scientific, industrial or commercial activity. Its role is constantly increasing and it concerns such vital sectors as energy,
Organization of Metrology: Industrial, Scientific, Legal
45
health, communications, food, armament, the security of goods and people, environment, transport, public works, etc. In order that measuring should have some meaning, and its results should be unquestionable and might be compared to those obtained at other times and in other places, each measurement must be related to a standard by an unbroken chain. It is the role of metrology to forge the different links of the chain and to make sure it does a good job.
2.2. Metrology: how? The organization of metrology cannot, and must not, be arbitrary. It can, and it must, evolve. It ever tends towards being more universal, which explains the success of the metric system that has become the International System of units (SI). Coherence, on which legal metrology in particular depends, has to be ensured at the international and national levels. International coherence means an SI resting on sound scientific bases and comparisons of the national standards of the different countries. It is essentially the sphere of the Conférence Générale des Poids et Mesures (CGPM) and its laboratory, the Bureau International des Poids et Mesures (BIPM). For about 30 years, regionalization of the world has been witnessed. Regional organizations that bring together national organizations have been created, and this allows a keener harmonization which makes the user’s task easier. Europe set an example by creating EUROMET in which the European National Metrology Institutes collaborate. It has also created European cooperation for Accreditation (EA) which brings together the accredited calibration laboratories, and its aim is to harmonize the operation of the national calibration chains. National coherence mirrors international coherence, although each country has its own national standards. A national organization studies, creates, maintains and upholds its standards. It sets up a system that connects the industrials’ standards, the measuring instruments and the results of the measurements themselves. As any system drifts, there is a need for a periodic follow-up in the field: it is the accreditation of the calibration laboratories that assumes the checking function, providing a link between the needs of industry and the National Metrology Institute. In France, this system was installed in 1969 by the Bureau National de Métrologie which consisted of five primary metrology laboratories and was in charge of the system of traceability chains and of the accreditation of the calibration
46
Metrology in Industry
laboratories. The latter activity was taken up by the COFRAC in 1994. In 2005, the monitoring of French metrology was transferred to the Laboratoire National de Métrologie et d’Essais (LNE). In Italy, the metrology system is based on three primary institutes (IMGC-CNR, IENGF and INMRI-ENEA) which have established and supervised the national standards since 1950, and on the Italian calibration service (SIT) which has been accrediting the calibration laboratories in Italy since 1979. It provides a chain for the dissemination of the standards and guarantees the traceability of all measurement results to the International System of units (SI). The Italian system has been acknowledged since 1991 by a law, no. 273, which establishes a national system of calibration (SNT) which in turn integrates all the structures (primary laboratories and accrediting institutes) (see Figure 2.4). The list of the accredited calibration laboratories, which are called SIT centers, is published in the Official Journal of the Italian Republic, Gazzetta Ufficiale; these are the only laboratories that guarantee traceability to the standards. In Switzerland, the federal government is responsible for the legislation related to metrology and for the diffusion of units. Legal metrology, which would be called regulated metrology today, is the business of the cantons. The confederation has created a federal office of metrology and the cantons have set up verification offices to carry out the tasks. All the official activities of metrology are to be found gathered in one institution and one place, the federal office of metrology and accreditation, METAS, which also manages the Swiss Accreditation Service, the SAS. This centralized organization was adopted at the beginning of the confederation’s activities related to metrology, after the Convention of the Meter was signed in 1875. The Swiss accreditation service (SAS) sets the examinations and delivers the accreditations in all the fields covered by the European or international standards in relation with accreditation and, in particular, in all the domains of metrology (Swiss Calibration Service – SCS). In order that correct values of units be disseminated with the required accuracy, Swiss metrology has set up traceability chains that guarantee the traceability of physical quantities, and of some chemical quantities such as gas mixtures. These chains originate from the METAS’s primary laboratories which materialize the units in accordance with their definition and transmit them to the METAS’s calibration laboratories through material standards. These calibration laboratories calibrate the standards of the clients.
Organization of Metrology: Industrial, Scientific, Legal
47
The concept of legal metrology arose as soon as man expressed a need to guarantee the integrity of commercial exchanges. It is this fundamental aspect of metrology that appears in the industrially-developing countries where “weights and measures” are still such as they have been traditionally known. Legal metrology, with its organization at the national and international levels, is not redundant; it relies on scientific and technical metrology to develop its specific mission; see section 2.4. Only the essential elements of the general nature and the history of metrology have been retained as they make it possible to better understand the current structures; but the history of metrology is fascinating; it is closely tied to the evolution of science and techniques, and to the evolution of mankind.
2.3. Scientific and technical metrology Organization at the international level (the BIPM) With the volume of commercial transactions expanding and with science and techniques developing in the 18th century, the necessity of making sure of the unity of measurements was powerfully felt by the middle of the 19th century. Difficulties were caused by the use of many of standards in commercial and cultural exchanges (such problems were especially conspicuous at the World Fairs), and the Convention of the Metre (20th May 1875) advocated a commitment to found and maintain, on a common foundation, an establishment whose initial aims would be: – to make sure that the metric system was used worldwide, while undertaking the realization and the upkeep of the (international) materialized standards of the meter and the kilogram, – and to ensure the coherence of national standards. The Convention adopted French as its official language. The international level In addition, the BIPM was intended to improve the processes of comparison and transfer between standards. Once the aims of the BIPM were established, all that was needed was a venue. On 22nd April 1876, the French government set the former Breteuil pavilion at the disposal of the Comité International des Poids et Mesures. The pavilion was situated at the heart of the Saint Cloud park, far away from any sources of vibration, and was a 4 hectare international enclave in French territory.
48
Metrology in Industry
CIPM 18 members
BIPM Laboratories
10 consultative committees
USA
UK
Germany
France
NIST
NPL
PTB
LNE
CH METAS
Italy SNT
Figure 2.1. The BIPM and the national laboratories of metrology
CGPM CIPM BIPM NIST PTB NPL LNE METAS SNT
Conférence Générale des Poids et Mesures (France) Comité International des Poids et Mesures (France) Bureau International des Poids et Mesures (France) National Institute for Sciences and Technology (USA) Physicalisch Technische Bundesanstalt (Germany) National Physical Laboratory (UK) Laboratoire National de Métrologie et d’Essais (France) Office Fédéral de Métrologie (Switzerland) Sistema Nazionale de Taratura (Italy)
2.3.1. The BIPM Today, at the beginning of the 21st century, the BIPM continues to attend to the standardization of physical measurements in the world. Its scientific activity aside, the BIPM is certainly the oldest establishment that “standardizes”; it is indeed possible to consider the SI as the oldest published document of international harmonization.
Organization of Metrology: Industrial, Scientific, Legal
49
The BIPM, together with the national metrology institutes, are responsible for the SI, which is the key to the uniformity of measurements internationally and one of the unquestionable bases of the industrialized world. In order to fulfill this mission of standardization, the BIPM has to establish the basic standards, as well as the scales of the physical quantities, and keep the international prototypes. To this day, only the unit of mass is kept under the form of a “materialized measure”. The other basic quantities of the SI are defined today from physical constants, such as the distance traveled by light in 3.34 nanoseconds (the physical constant is the speed of light in vacuum): – to compare the national standards to the international standards; – to organize international comparisons at the level of national standards; – to ensure the coordination of the corresponding techniques of measurement; – to bring into existence the determinations relative to the basic physical constants and coordinate them. The scientific activity of the laboratories of the BIPM is divided in relation to the units of the SI into: – length; – mass; – time; – electricity; – ionizing radiations; – chemistry. The CIPM supervises and guides the BIPM’s work and it is itself under the authority of the Conférence Générale des Poids et Mesures (CGPM). The CGPM is composed of delegates (51 in 2004) from all the states, which have signed the Treaty of the Metre Convention. The CGPM meets every four years and its mission is, in particular: – to debate and prompt the necessary steps to bring about the propagation and the improvement of the SI; – to approve the results of the new basic metrological determinations and adopt the various scientific resolutions of international significance.
50
Metrology in Industry
At the conference 18 members of the CIPM are elected, half of which are reelected every four years. In October 1999, the directors of the national metrology institutes (NMI) of the states belonging to the Metre Convention signed an arrangement (MRA) to mutually recognize the national measurement standards and the calibration and measurement certificates issued by their laboratories. In order that the criteria of mutual recognition be unbiased, the agreement is based on, first, the results of a set of key comparisons carried out according to specified methods that lead to a quantitative assessment of the degree of equivalence of the national measurement standards; secondly, the setting up by each NMI of appropriate means so as ensure the quality of the measurements; and thirdly, the actual participation of each NMI in suitable additional comparisons. This agreement is in two parts: in the first part, the signatories recognize the degree of equivalence of the national measurement standards of the participating national laboratories; in the second part, the signatories recognize the validity of the calibration and measurement certificates delivered by the participating laboratories. Thanks to the work of the CIPM and to the coordination by the BIPM, it is possible to compare measurements made in Europe, in North America, in South Asia, or in a nation which has joined the Metre Convention.
2.3.2. Results of the international activities As a consequence of these scientific activities, it has been possible to sign such international recognition agreements as the BNM/NBS agreement of 1989 (which has become the NIST). The agreement concluded there was not any significant gap between the American and French standards; this point turned out to be very important for the approval of the French manufacturers of fastening systems, within the scope of the American law (see the Fastener Quality Act (FQA) 1990). The result of these scientific works and agreements of equivalence is that it is now possible for European exporters to prove that they meet the requirements of many American contracts which still stipulate that the supplier has to be “traceable to NIST”. Hence metrology has lowered a technical obstacle to the export of our products to the North American continent. More generally, one of the goals of metrology is to make sure that a measurement made at Ulan-Bator (Mongolia) is comparable to the same measurement made later at La Paz (Bolivia), after possible corrections due to the environment among other conditions, have been applied. It is then possible to determine the exactitude of the comparison, and thus to reach the same conclusions,
Organization of Metrology: Industrial, Scientific, Legal
51
regardless of the geographical location, and excepting the measurement uncertainties of the two laboratories. Therefore, metrology is unquestionably useful in bringing people closer to one another by avoiding contentions and malfunctions directly related to measurements. More precisely, by taking into account the human factor, the major objective of the world organization of metrology is to determine the causes of the deviations and to define the uncertainty of the measurements (reproducibility, repeatability). The quality of the measurements that ensues will be synonymous with quality in essential spheres at the world level. These spheres include multinational industries which involve the development of subcontracting, the international trade of products, the networks of communication and navigation as well as a multitude of theoretical or applied technical and scientific activities. Metrology, as a universal language, contributes to the harmonization of scientific, technical and commercial relationships between peoples.
2.3.3. Regional organizations 2.3.3.1. EUROMET EUROMET is an organization which was officially founded in Madrid, in September 1987, following the signing of a “Memorandum of Understanding” (it was amended in August 1990 and July 1998), and it is made up of the NMI of the countries from the European Union, of the NMI of the European Free Trade Association (EFTA) and of the Commission of the European Communities. It is now also open to all the European countries, including new members, e.g. Turkey, Bulgaria and Romania. It was set up to develop cooperation between the national laboratories of metrology of Western Europe and provide an efficacious utilization of the means which are available. 2.3.3.1.1. Objectives and structures EUROMET’s aims are: – to develop a closer collaboration between the members, in the work on the standards, within the current decentralized metrological structure; – to optimize the use of the resources and services the members have at their disposal and emphasize the trend of members to satisfy detected metrological needs; – to improve the existing metrological services and make them accessible to all members; – to make sure that the new calibration benches created within EUROMET are accessible to all members.
52
Metrology in Industry
EUROMET’s specific aims are: – the coordination of studies about standards; – the coordination of the major investments in metrological means; – the transfer of competence between members in the domains of primary or national standards; – the facilitation of collaboration between members interested in a particular project; – the supply of information on resources and services; – the cooperation with the European accreditation bodies; – the cooperation with the European legal metrology services. EUROMET does not have any funds of its own; it operates on the basis of a voluntary participation. The expenses for cooperation and research are borne by the participating laboratories. Total autonomy is retained by the members. However, external financing is not excluded: the European Community in particular financially participates in the research programs. Structure Each member (the national metrology organizations) appoints a delegate; all the delegates constitute the General Assembly of EUROMET which meets at least once a year to debate its aims and objectives. EUROMET’s president is elected for two years and he provides a secretariat staff. 2.3.3.1.2. Technical activities There are 11 spheres of activities: – mass (force and pressure included); – length (dimensional measurements included); – electricity and magnetism (direct current and quantum metrology, low frequency, high frequencies); – time and frequency; – thermometry (thermal properties and humidity included); – ionizing radiations (dosimetry, radioactivity, metrology of neutrons); – photometry and radiometry (fibronics included); – flowmetry (properties of fluids included); – acoustics, ultrasonics and vibrations (accelerometry included);
Organization of Metrology: Industrial, Scientific, Legal
53
– chemical metrology (gas, organic and inorganic, electrochemistry); – interdisciplinary metrology. A “technical chairman” is elected by the committee in each subject field for a two-year mandate, which is renewable once. His main task is to coordinate the projects which are presented by the “contact person”, a specialist in the sphere of activities in question who has been appointed by the national organizations of metrology. Each collaborative project in a given activity is classified in one of the following categories: – cooperation in research; – intercomparison of measurement standards; – traceability; – consultation on facilities. On 1st May 2004, 256 projects were in progress; 368 have been previously carried through and have been concluded with a report. Four to five participants on average have collaborated in each project. It can easily be imagined that an important role in the European metrology is played by the countries with a larger GNP or possessing a larger size of metrology institute; they play the largest part in the projects. The number of projects is a proof of the success of EUROMET in terms of European cooperation, and some countries have taken advantage of their participation in EUROMET to develop their own metrological infrastructure. The spheres that give rise to the greatest number of projects are electricity, mass and length; time/frequency, acoustics and flowmetry give rise to the fewest. The spheres which have the highest number of projects are those that arouse a high interest, or are developing. There is often a collaboration outside EUROMET for those whose number of projects may seem low. Likewise, the number of projects are not the same within the categories of cooperation. The realization of common surveys is the type of collaboration that has the greatest attraction, which shows that metrologists are determined to pool their work. Interlaboratory comparisons come second because they are used to demonstrate the equivalence of standard realizations; they also make possible the gathering of information about traceability in Europe for the use of accreditation organizations.
54
Metrology in Industry
EUROMET remains closely linked to many European and international organizations. Among them, EA (European Cooperation for Accreditation), whose technical support is EUROMET, must be mentioned; so must WELMEC, EUROMET’s twin for legal metrology. EUROMET also collaborates with such organizations as EURACHEM; they have developed a common technical domain or sphere of activity, called the “amount of substance”, now called chemical metrology, and it is related to physicochemical analyses and measurements. As a regional organization, EUROMET has links with international and similar regional organizations. The BIPM, the OIML and COOMET are regularly invited to the plenary meetings of EUROMET’s committee to contribute towards its work and extend the cooperation between the different organizations. This cooperation is now extending to such regional organizations as the APMP (Asia Pacific Metrology Program) for South-East Asia and the West Pacific, the SADCMET (Southern African Development Community Cooperation in Measurement Traceability) for Southern Africa and the SIM (Sistema interamericano de metrologia) for the Americas. The most significant works to be carried out within EUROMET in the coming years will be the interlaboratory comparisons and the accreditation of the national laboratories of metrology which are the two major components of the planned elaboration of the mutual recognition agreements. 2.3.3.2. European Cooperation for Accreditation (EA) The Western European Calibration Cooperation (WECC)’s object was to testify to the collaboration of the official services of calibration-laboratory accreditation that operated in Western Europe. Originally in 1975, the WECC was a working section of the WEMC (Western European Metrologic Club) and it was called the Working Group on Calibration Services. The objective of the WECC was to establish and maintain a mutual and reciprocal confidence between the different accreditation services of Western Europe, so as to obtain the signing of recognition agreements and thus eliminate the technical obstacles to free trade resulting from calibrations, traceability or measurements. Another goal of the WECC was to secure and maintain the free movement of the know-how between the different organizations, in order to bring the capacities of calibration in Europe to the same level and to give the clients of the service the required guarantees.
Organization of Metrology: Industrial, Scientific, Legal
55
In June 1994, the WECC merged with its counterpart that dealt with testing and analysis laboratories, the Western European Laboratory Accreditation (WELAC) to form a new structure, EAL (European Cooperation for Accreditation of Laboratories), which in 1997 became the EA when it merged with EAC (European Cooperation for Certification), a counterpart which carried on the coordination between the organizations of accreditation and certification organizations. There are other regional or international organizations, such as APLAC (AsiaPacific laboratory Accreditation Cooperation) for the Asia-Pacific zone, IAAC (Inter America Accreditation Cooperation) for all the countries of the two Americas. Together with EA, these organizations work in the ILAC (International Laboratory Accreditation Cooperation) for a recognition of the calibration results. That recognition is validated through the agreements concerning the equivalence of the calibration certificates. 2.3.3.2.1. The EA Recognition Agreements These agreements have emerged from a long and rigorous process which begun at a time when the standards of the EN 45000 series did not exist, not even as projects. After an evaluation by a group of experts from the member countries of the EA, each organization has at its disposal a document that reports the deviations from the EA criteria. This process makes it possible to limit the number of crossed evaluations and especially the time spent on these evaluations, at the same time as it makes sure that the arrangements that appear in the EA’s report are still being applied in the assessed organization. The international agreements depend on the same principle: ILAC’s assessors make sure the regional agreement works well with regard to the requirements of the guide ISO/CEI 58 (EN 45003) by evaluating the work of the committee responsible for handling the agreement, as well as observing evaluations made in several countries that have signed the regional agreement. The organizations which have been invited to sign the multilateral recognition agreement declare that: There is no significant difference which might induce a user not to grant the same confidence to the calibration certificates issued by someone accredited: they are equivalent and can then be considered as such by those the certificates are addressed to.
This declaration of equivalence concerns all the calibration certificates stamped by one of the mentioned organizations (see Chapter 5). In no way do these agreements alter the operation of the organizations which, individually, retain their independence, their mode of functioning and their characteristics.
56
Metrology in Industry
The purpose of these agreements between national organizations of accreditation of calibration laboratories is to facilitate the recognition of the soundness of the measurements recorded in the calibration documents. They enable calibration certificates to circulate freely, in the spirit of the directives of the Commission of the European Union (new approach, global approach, modular approach, etc.). The EA makes bilateral recognitions easier between the different economic regions of the world through technical and organizational audits. 2.3.3.2.2. Definition of accreditation Accrediting a calibration laboratory is to recognize that the laboratory is apt to perform calibrations in a specified sphere, for clearly defined methods, in an identified measurement range and with associated uncertainties, while integrating the characteristics of the equipment which is to be connected to the standards. Each physical quantity is the object of a similar analysis; this leads to the drawing up of an accreditation certificate which defines the calibration which can be accredited for a given domain (dimensional metrology, electricity, mass, forces, ionizing radiations, temperature-hygrometry, etc.). The main objective of the national traceability chains is to make possible the connection of industrial measurements to national standards and to understand the needs of industry in the field of metrology, while ensuring there is a dialogue between laboratories and industrialists. The firms are then in a position to show that their products meet all of requirements, by means of tests carried out in their own laboratories, in accordance with the standards or rules the product is subjected to. The accreditation bodies take into consideration the competence and the experience of the personnel, the equipment, the calibration methods used and the connection to the national standards. Those elements are ensuring the coherence of the technical activity of the accredited laboratories and their calibration capabilities and associated uncertainties. Traceability to the national standards is a priori ensured only by the calibration certificates which bear the logotype of the national accreditation organization and are delivered by accredited laboratories. 2.3.3.2.3. Guarantees provided by accreditation Accreditation is the recognition of a certain competence and the assurance of the durability of this competence by an organization which is accepted as an authority on the subject. (EN 45020)
Organization of Metrology: Industrial, Scientific, Legal
57
That is why, in order to eventually guarantee a traceability to the national standards or to the SI, it is necessary to make sure the provider of the service is accredited for the domain in question. Having obtained this guarantee, you have to verify that: – the accreditation has been attributed to the firm or the agency that is likely to carry out the calibration (VIM section 6.13) and that it is valid at the relevant time; – the technical annexes to the convention do cover all the needs of the firm. The latter can, if necessary, call on several accredited laboratories to cover all the physical quantities and fields of measurement to which it wants its equipment connected. The annexes also specify whether the provider of the service is accredited to operate on site. This is important in the case of the connection of scales, power machines and other equipment that cannot be moved, either because the movement would ruin the calibration operation, or because it is not reasonable to move the equipment. In order to inform industrialists, accreditation organizations regularly publish facsimiles which reproduce in full the technical annexes of the accredited laboratories as soon as the annexes appear. These publications concern different physical quantities: dimensional metrology, pressure, mass, force, accelerometry, acoustics, flowmetry, electricity, time-frequency, ionizing radiations, temperature, hygrometry, radio-photometry, and reference materials. 2.3.3.2.4. Criteria of accreditation The following items are examined before a calibration laboratory is accredited for a field in which the measured physical quantities and the calibration uncertainties are defined by: – the qualification of the personnel and the presence of a technical supervisor, answerable for the validity of the calibration documents and responsible for the accredited laboratory; – the equipment and reference standards which suit the domain of measurement and the uncertainties stated by the laboratory; – the environment of the laboratory (temperature, hygrometry, vibrations); – the calibration methods; – the exhaustive assessment of the causes of uncertainty for each domain; – the means of traceability to national standards (reference standards, recalibration program and periodicity); – the internal calibration procedures (follow-up and checking of reference standards, periodicity and program for traceability of working standards).
58
Metrology in Industry
The experience of the laboratory is also examined; in some traceability chains it can be confirmed by means of proficiency testing. It is important, however, that there may be a certain repetition, so that the metrology service may, by itself, be able to detect any fault or abnormal drift. For example, the laboratory can use a high-stability generator to verify a high-resolution measuring instrument; two smooth rings, connected to an approved center, can be compared on a measuring machine; the laboratory has two systems at its disposal which make it possible to compare two sets of standard gauge blocks, or two standard gauges, etc. In addition, the accreditors have established a number of procedures that are meant to ensure the quality of the calibrations performed in accredited laboratories is permanent: – technical audit of the laboratories; – quality audit of the general requirements; – periodical re-examination of the accreditations; – yearly survey of the connections achieved; – numerous comparisons are organized, within the EA, between the different accredited laboratories, to ensure the calibrations are coherent and the clients, whatever their nationality, receive equivalent services. 2.3.3.3. Accreditation procedure The object of accreditation is to ensure that: – the minimum requirements which are indispensable to guarantee the traceability of the references to the national standards are set up; – the potential calibration of the implemented measuring instruments and the measurement and uncertainty ranges claimed are coherent; – the demands for quality assurance of standard ISO/CEI 17025 and of the EA’s specific documents are met. One should be careful to differentiate between calibration and handling a bank of measuring equipment (see Chapter 11).
Organization of Metrology: Industrial, Scientific, Legal
59
B.I.P.M (Sèvres) Laboratories SIM
EUROMET
USA
UK
BDR
France
CH
Italy
NIST
NPL
PTB
LNE
METAS
SNT
National Laboratories National Standards Reference standards of the accredited laboratory
LNE LNE-INM LNE-LNHB LNE-SYRTE
Accredited laboratories Cofrac-calibration
METAS
IENGF IMGC ENEA
Accredited laboratories SCS
Accredited laboratories SIT
Reference standards of the firms National and European firms Control of the process of measurement
Figure 2.2. Example of traceability scheme in Europe
2.3.4. Organization at the national level 2.3.4.1. The Laboratoire National de Métrologie et d’Essais (LNE) 2.3.4.1.1. Role and missions Metrology became organized in 1969 in France, when the Bureau National de Métrologie (BNM) was created; its mission was to animate and coordinate the actions initiated by the different ministry departments in the sphere of metrology. A structural reform was undertaken in 1994 to consolidate its action and diversify its activity.
60
Metrology in Industry
By a ministerial order on 22nd December 1994, the BNM became a public interest group whose mission was to prepare and implement the national policy for metrology. The group was renewed by notice on 22nd May 2001 for a term of four years, with a structure slightly different from that established in 1994. However, the organization as a public interest group (GIP) had two principal disadvantages: first, its temporary nature whereas metrology is a perennial task, and secondly, its position as an intermediary that did not have a high visibility at the international level, as most countries have only one national metrology institute (NMI) linked to designated bodies, if necessary. In January 2005, the Ministry of Industry and the Ministry of Research decided to dissolve the BNM and transfer the central task of metrology to the LNE which was renamed the Laboratoire National de Métrologie et d’Essais. One of the objectives of metrology is to ensure the national and international coherence of the measurements made in the firms. The qualitative and quantitative checks, the development of subcontracting, and the technological evolution strengthen the role of metrology in industrial processes, scientific research, and in trading as well. The metrological needs that the LNE is charged to satisfy arise from very various spheres of activity: car manufacture, aircraft, space and nuclear industries, armament, public works, health and security, communications, transport, environment, chemistry and analysis, etc. To meet these needs, the LNE, together with three other national metrology laboratories and six designated laboratories, form a coherent and coordinated body of four national metrology laboratories and six designated laboratories associated to the LNE (they have signed a contract with the LNE). They carry out: – research in physics and chemistry, which leads to new definitions and realization of units, some of them based on fundamental phenomena; – work to improve and maintain current national references; – linking the references of firms and technical organizations to national standards, with the best uncertainties.
Organization of Metrology: Industrial, Scientific, Legal
61
2.3.4.1.2. Organization of French metrology (monitored by the LNE) General organization The scientific and technical activities related to metrology are divided between the different partners as follows. The laboratories of the LNE: metrology in chemistry, dimensional metrology (material standards), mass and related quantities (accelerometry, mass, force, couple, pressure, viscosity), temperature as a complement of the standards of the LNE-INM and thermophysical properties of materials, radiometry and photometry as a complement of the standards of the LNE-INM, electricity (quantum metrology), development of references in the ranges, direct current and low frequency, high frequency and electromagnetic radiations, and guided optics. The laboratory LNE-INM (National Institute of Metrology), at the CNAM (Conservatoire National des Arts et Métiers): wavelength and refractometry, mass, temperature (unit, scale), radiometry-photometry and acoustic pressure in cavity. The laboratory LNE-LNHB (National Laboratory Henri Becquerel), at the CEA (Commissariat à l’énergie atomique): ionizing radiations (activity, flux, exposure, kerma in the air, absorbed dose, dose equivalent, dosimetry of X-rays. The laboratory LNE-SYRTE (Time-space Reference Systems), at the Paris Observatory: time and frequency, with the basic unit (second) and the derived unit (hertz), unit and scales of time (dissemination, legal time, references of frequencies (from the radioelectric domain to the optical domain) and chains of measurement of frequencies (from the radioelectric domain to the optical domain)). The associated laboratories are: – LADG: gas flow; – IRSN: neutron dosimetry; – CETIAT: hygrometry, liquid flow, anemometry; – ENSAM-PARIS: dynamic pressure; – Observatory of Besançon: time (time interval, stability of time and frequency); – FEMTO-ST: frequencies (oscillators, spectral density of phase). Traceability of the industry’s standards and references The LNE, together with the national metrology laboratories and designated laboratories, develop high-level means of transfer and calibration. These means are used to make possible the traceability of any working standard, measuring instrument or reference material to national standards set up in the calibration services of the national metrology laboratories and designated laboratories.
62
Metrology in Industry
The calibration services of the national metrology laboratories and the laboratories accredited by COFRAC for calibration are responsible for performing calibrations that ensure traceability to the national standards. This traceability is guaranteed by the logos of the organizations that have accredited the calibration laboratories, COFRAC in France, SIT in Italy, and SCS in Switzerland (see Chapter 5). In addition, the LNE is responsible, together with COFRAC, for encouraging and coordinating the actions undertaken within the system of the calibration chains. The LNE, together with the national metrology laboratories (NML), defines the structure of the calibration chains and provides COFRAC with its scientific and technical competence. The principle structure of these chains is shown in Figure 2.3.
Standard
LNE NML
Laboratories Accredited for calibration Transfer to users
NMI
National Standards
Laboratories Associated to LNE
Laboratories Accredited for calibration Not Accredited Calibration laboratories Industrial measurements
Figure 2.3. Traceability scheme in France
International cooperation The LNE is France’s representative to international metrological organizations (Conférence Générale des Poids et Mesures, Comité International des Poids et Mesures). It is on all the consultative committees and chairs several working groups. This presence enables it, together with its counterparts, to ensure the coherence of the implementation of the SI, and of the new Mutual Recognition of the CIPM. Information and training Another mission of the LNE is to “gather, exploit and circulate the information and documents touching the developments of metrology”. To that end, it publishes a scientific and technical journal La revue française de métrologie, the aim of which is to inform scientific and industrial circles about the achievements, programs and prospects of French metrology. The LNE also publishes sector-based monographs,
Organization of Metrology: Industrial, Scientific, Legal
63
organizes theme days about specific metrological sectors, and takes part in exhibitions about metrology. Every year, the LNE, the NML and the designated laboratories organize training courses in the various fields of metrology, as part of continuing education. 2.3.4.2. The Italian national calibration system (SNT) In Italy, the traceability of measurements is guaranteed by law when the Sistema Nazionale di Taratura (SNT) is used; it is comprised of the metrology institutes and the accredited laboratories, as provided in the law no. 273 which established the national calibration system, SNT, in August 1991. The SNT has a three-level structure, as shown in Figure 2.4. The first level is the primary metrology institutes which establish the SI units and maintain them; they also ensure their dissemination at the highest level and guarantee the traceability of the measurement results. For historical reasons, three institutes are responsible for the different units of the SI: – the IMGC-CNR, units of mechanics and science of heat; – the IEN-GF, units of electric quantities, time and frequencies, photometry, optics and acoustic quantities; – the INMRI-ENEA, unit of ionizing radiations. The IMGC-CNR and the IEN-GF recently merged to create the INRIM, a single national institute covering all the metrological activities. The IMGC-CNR is situated in Torino where it has, since 1968, been carrying out research in the field of metrology. The national standards established by this institute, in compliance with the SI, cover the following basic quantities: length, mass and temperature. The IMGC also uses the units derived from the basic units: angles, force, pressure, volumic mass and flowmetry, scale of hardness, hygrometry and accelerometry. The IEN-GF is also in Torino, not far from the IMGC, in an area which is known as the Italian metrological pole. The national standards developed by this institute are: power intensity, luminous intensity, acoustic pressures and electric quantities (farad, volt, ohm, watt, joule, henry, magnetic flux, luminous flux). The IEN-GF’s activity is not limited to metrology; it also involved in the sector of materials and technological innovation.
64
Metrology in Industry
The INMRI-ENEA is situated at Roma Casaccia; it is responsible, in the domain of ionizing radiations, for the units of absorbed doses, the activity of a radionuclide, the flux of neutrons and exposure. The IMGC, the IEN-GF and the INMRI, which set up the metrological standards for Italy, enjoy an environment that is very conducive to innovation in the different fields of measurement sciences. These three institutes cooperate in the activities carried out as part of the Convention of the Metre, at the level of the International Committee of Weights and CIPM, as well as at the level of the consultative committees for the definition of the meter, mass, thermometric quantities, time and frequency, electricity and magnetism, photometry and radiometry. In addition, they contribute to the activities of EUROMET.
NATIONALCALIBRATION CALIBRATIONSYSTEM SYSTEM SNT – NATIONAL
MINISTRY MINISTRY of INDUSTRY and INDUSTRYan COMMERCE
METRIC METRIC CENTRAL COMMITTEE COMMITTEE
MINISTRY MINISTRY of UNIVERSITY UNIVERSITYand an SCIENCE RESEARCH
NATIONAL CALIBRATION SYSTEM PRIMARY PRIMARYMETROLOGY METROLOGY INSTITUTES INSTITUTES IMGC ENEA IMGC – IEN – ENE SIT SIT Accreditation Structure Structure Accreditatio SIT SIT Users
EA Calibration Calibration services service in Europe
SIT SIT COMMITTEE COMMITTEE Secretariat Secretaria Technical Committees Technica Committee Working Groups WorkingGroup
centers Researc center • Research laboratories • Test Tes laboratorie • Industrial sectors sector • Services Service
SIT SIT Calibration CalibrationCenters Centres
Figure 2.4. Accreditation of the calibration laboratories in Italy (SIT)
The SIT (Servizio Italiano di Taratura) is found at the second level; it is the national accreditation organization with full authority to deliver accreditation to calibration laboratories.
Organization of Metrology: Industrial, Scientific, Legal
65
The laboratories which are accredited by the SIT (SIT centers) perform calibrations and deliver calibration certificates which are technically as valid as those of the primary institutes, but with higher uncertainties. 2.3.4.3. The Swiss national calibration system The federal office of metrology and accreditation has gathered all the official activities of metrology into one institution and on one site. It also manages the Swiss Accreditation Service (SAS). This centralized organization was established as soon as the Swiss Confederation began to deal with metrology, after the Convention of the Metre was signed in 1875. At the beginning of 2001, the office adopted Metrology and Accreditation Switzerland (METAS) as its name; thus, scientific and legal metrology, as well as the SAS, came together under one name. METAS’s tasks METAS’s tasks are defined in Article 17 of the federal law on metrology; this law sets out the scope of official metrology in Switzerland. Article 17 states that the office has, in particular, the following tasks: – it prepares the legislation related to metrology and ensures that it is enforced; – it determines and circulates sufficiently precise standard values of the units used in metrology and does the necessary research and the scientific and technical work of development; – it elaborates the requirements needed for the determination, the transmission and the accurate estimation of physical quantities; – it examines measuring instruments and metrological testing methods and makes decisions about their conformity, their acceptance or approval and, if applicable, their verification; – it advises and trains the personnel of the cantonal offices of verification, draft directives for these offices and checks their measuring instruments; – it oversees the enforcement of the law in the cantons; – it gives consultations and performs evaluations; – it performs the activities that third parties request it to do (and is paid for those activities) within the limits of its capabilities. the agreement of the relevant department is needed for important activities. METAS has adopted a matrix organization and a matrix distribution of the work and responsibilities to carry out these different tasks. Teams of experts are formed for the particular objectives to be reached, which ensures cooperation between all the specialists and a rational and efficient utilization of the experience and knowledge of each specialist.
66
Metrology in Industry
METAS’s general organization In METAS there are: – two scientific and technical divisions which oversee seven sections altogether; – two technical, logistic and administrative support services; – the SAS; – one research and development staff member; – one management staff member. At the international level, METAS cooperates in research, the object of which is to establish new definitions of basic units, or to improve their implementation. In particular, we would mention the quantified Hall and Josephson effects in electricity, Watt’s scales for the kilogram and, in relation to length, the new definition of the meter and the length metrology. To ensure the availability and the transmission of the correct values of units with the required accuracy, Swiss metrology has set up traceability chains which guarantee the traceability of physical quantities and of some chemical quantities such as gas mixtures. These chains originate in METAS’s primary laboratories which materialize the units from their definitions and pass them to METAS’s calibration laboratories in the form of material standards. These calibration laboratories calibrate the standards of the clients, most of them being accredited. In their turn, the accredited laboratories calibrate the standards of industry, commerce and research. In legal metrology, METAS itself calibrates the standards of the verification organizations which are usually dependent on the cantonal authorities. This well-documented system contributes towards the international recognition of the certificates of conformity issued in Switzerland. In order to meet the needs of its clients as satisfactorily as possible, METAS collects and distributes as much information as it can about metrology, conformity, accreditation and the recognition of certificates. It publishes a scientific and technical journal, the “METAS Info”, which informs all those that are interested in scientific realizations, technical problems, international cooperation and the decisions of Swiss metrology. It regularly organizes seminars on topics of general interest, such as uncertainties, and also some training courses for those who verify weights and measures, which enable them to obtain a federal certificate of capability. METAS takes an active part in the works of the following organizations and it collaborates with many of their subcommittees.
Organization of Metrology: Industrial, Scientific, Legal
67
In metrology CGPM General Conference of Weights and Measures (CGPM); OIML International Organization of Legal Metrology (OIML); ISO International Organization of Standardization (ISO); EUROMET International Electronic Electrotechnical Commission; WELMEC European Cooperation for Legal Metrology (WELMEC); and other more specialized organizations. 2.4. Legal metrology1 2.4.1. Scope of legal metrology The term “legal metrology” applies to any application of metrology that is subject to national laws or regulations. This definition means that the scope of legal metrology may vary considerably from one country to another. With the exception of research, any application of metrology may fall under the scope of legal metrology if regulations are applicable to all measuring methods and instruments, and in particular if quality control is supervised by the state. This is the case in some countries, whereas in most countries the regulated area generally concerns measurements for trade. However, many countries also regulate Health and Safety policy and evidential measurements. Legal metrology covers measurements and measuring instruments that the state considers to be to much a sensitive subject for society. The Technical Barriers to Trade Agreement (World Trade Organization) sets up a framework under which technical regulations may be developed, and this framework applies to the scope of legal metrology. Article 2.2 defines what is and is not be covered by legal metrology: Article 2.2 (...) technical regulations shall not be more trade-restrictive than necessary to fulfill a legitimate objective. (...) legitimate objectives are inter alia: national security requirements, the prevention of deceptive practices, protection of human health or safety, animal or plant life or health, or the environment.
The first aim of legal metrology is to define which units of measurement are acceptable in the relevant country and for what purposes. In most countries, legal 1 This section has been written with the help of Gérard Lagauterie, Sous-Directeur de la Métrologie, France.
68
Metrology in Industry
units are the SI units, plus special units for specific applications and, in some countries, customary units. In relation to measurements, legal metrology regulations may require that certain measurements be carried out, that transactions be based on these measurement results and it may require minimal performance levels for these measurements. It is however quite unusual for regulations to prescribe the maximum uncertainty of such regulated measurements as defined by the Guide to the Expression of Uncertainty in Measurement (GUM) Regulations on measurement results, which generally consist of: – setting acceptable limits to the content of prepackages compared with their nominal value, – prescribing that measurements shall be performed with instruments of a given accuracy class subject to legal control. The third part of legal metrology consists of submitting certain categories of measuring instruments to legal control. Depending on the country, this regulatory scope may be limited to a few categories used in domestic trade (weighing scales, petrol pumps, etc.), or may cover categories of instruments used for transactions between companies. Most often the instruments used for levying taxes are the object of special attention from the regulatory authorities. Usually, the instruments used for the implementation of technical regulations are submitted to legal control, but the list may be diverse according to the countries. This is to give confidence to the public that regulatory controls are carried out with appropriate and reliable instruments (brake efficiency of vehicles, exhaust gas analysis, sound level of equipment for industry or public works, lighting in the workplace, etc.), and that prosecution of offenses is based on reliable measurements (radar speed meters for vehicles, breath analyzers, etc.). Instruments used for healthcare, for public safety or environmental protection and monitoring are more and more frequently submitted to legal metrology control (medical instruments, measurement of pollutant emissions, etc.). Although its organization differs from one country to another, legal metrology is present in nearly all countries – hence an international organization, the OIML, was set up 50 years ago to deal with this aspect of metrology. 2.4.2. The International Organization of Legal Metrology (OIML) The OIML is an intergovernmental organization established by a treaty in 1955, and whose general objective is to organize mutual information and cooperation among its members in the field of legal metrology, to harmonize legal metrology
Organization of Metrology: Industrial, Scientific, Legal
69
regulations and to foster mutual confidence. The OIML has about 60 member states (who are signatories to the treaty, committed to implementing common decisions), and about 50 corresponding members. One of the main activities of the OIML is to harmonize legal metrology regulations by developing international recommendations, which are model regulations proposed to its members when they intend to regulate a category of measuring instruments. These international recommendations have three parts: requirements, test procedures and the test report format. In addition to recommendations, the OIML develops international documents, which are of a more informative nature. Due to this harmonizing role, the OIML is an international standard-setting body and has been accepted as an observer in the Technical Barriers to Trade Committee in the World Trade Organization. To complement to its harmonizing activity, the OIML has developed – and is continuously developing – systems to facilitate mutual recognition and mutual acceptance of legal metrology controls. The OIML Certificate System for Measuring Instruments was established in 1990 and allows member states, under stated conditions, to appoint the authority which issues certificates of conformity for types of measuring instruments that comply with the requirements of the OIML recommendations. The OIML System is now completed by a Mutual Acceptance Arrangement which came into force in 2005 and which will result in Declarations of Mutual Confidence in the type testing results. The OIML then intends to establish systems for certifying the conformity of prepackages, and for certifying the conformity of individual instruments against the OIML requirements. The purpose of these activities is to set up a global legal metrology system. Harmonization of regulations and elimination of technical barriers to trade form two important elements of the global system under development, for that will reduce the costs of selling instruments on the market and the costs of international trade. However, this harmonization and cooperation will also present important benefits for all countries and for society. Cooperation within the OIML allows the level of protection of consumers, trading partners and the public worldwide to be raised, and allows states to develop an efficient legal metrology system at an acceptable cost, by networking and avoiding costly duplication of resources. The executive headquarters of the OIML are the Bureau International de Métrologie Légale (BIML), located in Paris. The BIML coordinates and supports the work carried out by the OIML technical committees and subcommittees, supports the work of all OIML structures, and edits and publishes OIML publications.
70
Metrology in Industry
The structure of the OIML is as follows: – The International Conference of Legal Metrology, which is the highest level. The Conference meets every four years and is composed of delegations from all member states. It takes all fundamental decisions concerning the OIML, and in particular its budget, its policy, the formal adoption of OIML recommendations and any decision for common action by member states. – The International Committee of Legal Metrology (CIML) is composed of one delegate from each member state, in principle the persons responsible for legal metrology in their respective countries. The CIML follows the technical work of the technical committees and subcommittees, engages discussion and undertakes studies for further decisions at the Conference, and approves the OIML recommendations and other publications. The CIML elects a president and two vice-presidents. – The technical committees and subcommittees are the bodies in charge of developing the OIML recommendations and documents. These committees are composed of experts appointed by the CIML members and observers from corresponding members and organizations in liaison. In addition to this structure, two advisory groups must be noted: – The CIML Presidential Council, composed of the CIML president and vicepresident, plus seven CIML members appointed by the CIML president. The CIML Presidential Council advises the CIML president and vice-president on strategic issues. – The Permanent Working Group for Developing Countries is an advisory group in charge of studying any action necessary to support developing countries in the OIML and of carrying out these initiatives. The OIML has close liaisons with a number of international organizations, and in particular with the Metre Convention, ISO, the International Electrotechnical Commission (IEC), the International Laboratory Accreditation Cooperation (ILAC), the International Accreditation Forum (IAF), the World Trade Organization (WTO), etc. Regional legal metrology organizations (the Asia-Pacific Legal Metrology Forum (APLMF), the Euro Mediterranean Legal Metrology Forum (EMLMF), the Southern African Development Community Cooperation in Legal Metrology (SADCMEL), and European Cooperation in Legal Metrology (WELMEC)) and regional metrology organizations (the Euro Asian Cooperation of National Metrology Institutes (COOMET), the Systema Interamericano de Metrologia (SIM), the European Collaboration in Measurement Standards (EUROMET), etc.) are also key liaisons for the OIML. The OIML languages are French (official language) and English (working language). The BIML publishes a specialized bulletin four times a year. Bureau International de Métrologie Légale 11, rue Turgot 75009 PARIS - France Tél.: +33(0) 1 48 78 12 82 - Fax +33(0) 1 42 82 17 27
Organization of Metrology: Industrial, Scientific, Legal
71
The BIML maintains a website (www.oiml.org) which presents information on the OIML, its members, structures, work and publications. There is also a restricted access members-only area, where circulars, drafts of recommendations and news of interest to members are regularly posted. 2.4.3. The European level 2.4.3.1. European Union harmonization The European Commission (DG Enterprise) has among its missions to harmonize the national regulations that could create technical barriers to trade. The national legal metrology regulations have been harmonized by four series of European Directives: – Directive 80/181/EEC on 20th December 1979 (modified) on legal units; – Directive 71/316/EEC (“Old Approach” Directive) and the Directives adopted in its application; – Directive 90/384/EC modified (“New Approach” Directive), related to NonAutomatic Weighing Instruments; – Directive 2004/22/EC on 31st March 2004 (“New Approach” Directive, usually called the “Measuring Instruments Directive” or MID), which covers 10 categories of measuring instruments. These Directives are applicable through their adoption into the national legislative and regulatory texts. In addition, under DG Transport, the European Regulation 3820/85 EEC on 20th December 1985 (directly applicable without being adopted into national legislation), completed by European Regulation 3821/85 EEC on 20th December 1985, requirements were set up for legal control of the instruments (that is, tachographs) installed on trucks and collective transport vehicles to measure and record speed, driving time, etc. A new generation of tachographs has been defined and regulated by adapting the European Regulation to technical progress (European Regulation 2135/98 on 24th September 1998). 2.4.3.2. WELMEC WELMEC was created in 1989: it is an organization which coordinates the national authorities of legal metrology of the Western European countries within the European Union and common European economic frameworks. WELMEC grew after its creation by accepting as associated members the countries of Central Europe which were committed to entry into the European Union. Today WELMEC has 28 members and two associated members, as most of these original associated members have since joined the European Union.
72
Metrology in Industry
The purpose of WELMEC is to facilitate the exchange of information and favor the mutual acquaintance of the member countries, to harmonize the regulations and checking methods, and to promote mutual recognitions. The objective of WELMEC’s first works was to harmonize the enforcement of the European Directive 90/384 about non-automatic weighing instruments. WELMEC published enforcement guides about this new-approach Directive so that the notified organizations might enforce it in as homogeneous a way as possible. Afterwards, different works were undertaken to harmonize the approach of the member states on different subjects of legal metrology, in particular: – checking prepacking; – requirements applying to notified organizations; – methods for the examination of the software of prescribed instruments; – surveillance of the market for the enforcement of European directives; – various technical fields: weighing instruments, measuring sets for liquids other than water, household meters (used by public utility services). WELMEC has also acted as a group of experts supporting the European Community in the finalization of the future European Directive on measuring instruments, and assisting in the tasks of the working party of the European Council related to the Directive. Since the Directive was published (30th April 2004), the European Commission has reasserted its interest in the work of WELMEC, which has organized itself (new working groups have been created) and launched many initiatives intended to ensure a harmonized implementation of this Directive. In addition, WELMEC has published a repertory of the organization of legal metrology in the member states and corresponding members. WELMEC has concluded a multilateral agreement to recognize model approvals; it states that when, in a member country, an instrument is simultaneously granted an OIML certificate of conformity and a national model approval, this instrument is automatically granted a model approval in the other signatory countries, barring any pressing, and justified, reason. WELMEC’s organization is comprised of: – the Committee of WELMEC, which meets every eight months; –the “Chairman’s group” of WELMEC; – the working parties of WELMEC.
Organization of Metrology: Industrial, Scientific, Legal
73
The federations of manufacturers of prescribed instruments are associated with the activities of the working parties of WELMEC. WELMEC’s member for France is also a member of the Chairman’s Group. 2.4.3.3. Other regional bodies Most regions in the world have set up bodies for cooperation in legal metrology. These regional bodies, like WELMEC, are in line with the OIML and their actions complements that of the OIML. The activities differ from one region to the other, but typically aim to: – develop mutual knowledge at regional level; – develop exchange of experience on legal metrology; – develop mutual confidence; – study and address the needs for training and drawing up training programs. The following are examples of regional bodies: – WELMEC (see above); – APLMF; – SADCMEL; – SIM; – COOMET; – EMLMF; – South Pacific Legal Metrology Forum (SPLMF).
2.4.4. At national level 2.4.4.1. Legal metrology in Italy In Italy, the legal metrology is included in the “Harmonization and Market Surveillance” Department of the Ministry of Industry. The Ministry takes on the main responsibilities of legal metrology. In the Ministry, the metric central office is responsible for the following activities: – drafting regulations; – organizing the metric services and the analysis of precious metals; – protecting consumers; – looking into the activity concerning prepacked products; – looking into the activity concerning the market surveillance.
74
Metrology in Industry
All the inspections and controls have been recently delegated to the local offices of the provincial Chamber of Commerce. These activities are related to: – the approval of the model; – the initial verification; – the periodic verification and inspection assessment (control at the user’s place, where the equipment used for trade is inspected every other year). The measurements related to the above activities are performed in governmental laboratories for calibration and testing, which are located in the different Italian regions and controlled by the local the chamber of commerce. Measurements made at the producer’s laboratory are accepted, if they have been made in accordance with the official procedures and in the presence of official inspectors from the legal metrology offices. The metric central office of the Ministry keeps close contact with the primary metrology institutes which are described in section 2.3.4.2. Representatives of the primary metrology institutes are members of the bureau of the metric central office; they link legal metrology and scientific metrology and ensure the traceability to the SI units. 2.4.4.2. Legal metrology in Switzerland In Switzerland, the federal government is responsible for the legislation in relation to metrology, and for the diffusion of the units. The enforcement of legal metrology – it is called regulated metrology – is the concern of the cantons. The confederation has created a federal office of metrology where the cantons discharge their tasks and the cantons have set up verification offices. In METAS, the federal office of metrology and accreditation, all the official activities of metrology are brought together in one institution and on one site. It is also in charge of the SAS. This centralized organization was established as soon as the confederation became involved in metrology after the signing of the Convention of the Metre in 1875. Since 1st January 1999, METAS is managed according to the principles of the new public administration. They require a budget and acceptance of the cost of the services by the public authority. In this context, METAS has defined four groups that characterize the provided services. These four groups are as follows. National basis of measurement This group deals with the services provided by the primary laboratories which, each in its own sphere, are responsible for the first link of the traceability chains, or
Organization of Metrology: Industrial, Scientific, Legal
75
for the whole traceability chain. The units are established with as high an accuracy as possible, according their definition. Legal metrology The group deals with the preparation of requirements, the supervision of their execution, the training of the operators, and the strict surveillance of the market. Legal metrology deals with the domains of trading, public health and security and also with the official measurements of data related to physical quantities. Industrial metrology In this group are all the calibration services which provide interested parties with sufficiently accurate values of the units. Model pattern approvals, which will become certificates of conformity, are dealt with by this group, some of the measuring instruments used in legal metrology. Accreditation The SAS does the tests, performs the evaluations and delivers the accreditations in all the domains coming under the European or international standards related to accreditation. The support activities needed to run the METAS office, but which do not provide services to third parties, have to be added to the above four groups. The federal government made provision of delegations of competence and then passed a series of ordinances to deal with the following areas: – the selling of goods in bulk or prepacked; – the principles relating to approvals and verifications; – the tasks and the competence of the verification offices and verification laboratories; – the remuneration paid for metrological work. The technical directions regulations concerning the different types of measuring instruments are in the domain competence of the federal councilor (minister) in charge of METAS. Seven sections make up METAS; six of them are concerned with a specific domain of physics or chemistry, and the seventh is responsible for legal metrology. The Swiss constitution states that the legislation on metrology is the domain of the confederation. From this constitutional foundation, the parliament has brought into effect a federal law on metrology which stands as the framework for all metrology in Switzerland.
76
Metrology in Industry
2.4.4.3. Legal metrology in France Legal metrology in France is dependent on the ministry of industry, more precisely the metrology department part of the DARPMI (Direction de l’action régionale et de la petite et moyenne industrie); and at the territorial level legal metrology is dependent on the regional departments of industry, research and environment (DRIRE). Legal metrology is the modern form of the very old control of weights and measures; it includes all the statutory measures as well as the administrative and technical procedures that have been introduced by the authorities to guarantee the quality of the measuring instruments used in trading (scales used for retail sales, petrol pumps, etc.), in official controls when safety is involved (cinemometers (“radars”), chronotachygraphs (“black boxes”), manometers, etc.). Consequently, some categories of measuring instruments are subjected to regulations and controlled by the state. This control is exercised at several levels: – at the conception (approval of model, or EC-type test for the “new approach” directives); – at the manufacturing level (initial verification, or corresponding European procedures when the instruments come under a “new approach” directive); – at the level of the daily use of the instrument (periodic verification and control of the instruments in service). The control is presently in full (r)evolution, since it is possible, thanks to the techniques of quality assurance, to entrust third parties, such as approved repairers or the manufacturers of the measuring instruments, with some checking operations, in certain conditions. Taking this possibility into account, the new Decree of 3rd May 2001 (no. 2001-387), about the control of measuring instruments, clearly states that the processes of metrological control would be delegated to some organizations, with the assent of the regional préfet, or of the minister in charge of metrology, according to circumstances. It nevertheless specifies that the operations are to be performed by state agents, if there are no suitable organizations. Concerning the assessment of the design of instruments and the approval of the quality systems of manufacturers, repairers and fitters, the delegation process has been completed. It is well on its way for primitive verification and almost over for in-service checking. Once the delegation process is over, the role of the state will chiefly consist of approving or appointing the verifying organizations and ensuring that the system as a whole is soundly implemented. That role will include:
Organization of Metrology: Industrial, Scientific, Legal
77
– supervision of the organizations and other operators (manufacturers, repairers, fitters); – supervision of the market, which means ensuring that the new instruments that are marketed and put into service meet the requirements; – supervision of the stock, which means ensuring that the instruments in service conform with the regulations and are used correctly and, in particular, that they have been duly verified. The appointed, or approved, organizations must prove their competence, quality and absence of impartiality. Such proof usually comes from the systems of reference applicable to laboratories (standards of the EN 45000 series). The assessment of these organizations is done in line with the accreditation methods, even when accreditation is not required by specific regulations. The new policy is to require accreditation for most of these organizations. The new European Measuring Instruments Directive (MID) encompasses most of the regulated measuring instruments (that is, the number of instruments, if not the number of categories of instruments). This Directive comes into force on 30th October 2006 and, once it has been adopted into national law, will take over from the national regulation on new instruments. The French organization is already compatible with the European Directive. However, the state will continue to deal with: – the development of the regulations; – the involvement of the French legal metrology in international works; – the harmonization of texts at the European and international levels; – the approval of models; – the coordination of metrological controls. The metrology department has bilateral cooperative relations with a number of national legal metrology authorities. Some cooperative relations give rise to agreements of bilateral recognition of approval testing; the agreements concluded with the PTB in Germany and the NMI in Netherlands are examples. Informal bilateral agreements of recognition with all the countries of the European Economic Area, even if they have not been formally officialized by bilateral agreements of recognition. The relations which are built up in the OIML enable the exchange of information with numerous countries about the statutory requirements, the testing techniques and procedures, etc. More formalized cooperation with some other countries is being developed. The topics of collaboration are: activity of model approval, exchange of experts and technical information. This is true of Poland, Romania, Morocco and Tunisia, in particular.
This page intentionally left blank
Chapter 3
Mastering Measurement Processes Approach to the Setting up of a Metrology Function
3.1. What to do at the beginning? Metrology is neither a fashion nor a fad of auditors. Metrology is necessary to make pertinent decisions, for example: – to control the manufacturing processes; – to verify and certify the products are true to the specifications; – to guarantee the safety of goods and people; – to protect the environment. Firms that are setting up a metrological function find the following difficult: – obtaining a good understanding of the aims of, and reasons for, metrology in a firm; – obtaining a good understanding of the basic concepts of metrology, such as traceability, calibration, reference standard, uncertainty of calibration, uncertainty of measurement, etc; – understanding the metrological requirements of the ISO 900: 2000 and 9004: 2000 standards and adapting them to the specific needs of the firm.
Chapter written by Marc PRIEL – Laboratoire National de Métrologie et d’Essais (LNE), Patrick REPOSEUR – Comité Français d’Accréditation (COFRAC).
80
Metrology in Industry
There are many pitfalls which have to be avoided when setting up a metrology function: – overdoing the function; – confining oneself to formal aspects without technically exploiting the results of a well-controlled metrology; – accepting too much guidance, for example from sometimes not very competent representatives or from an auditor, instead of bringing in one’s own views. What should be done then? We are inclined to answer: – try to get a good understanding of the basic concepts of metrology; – become informed about the EN ISO 10012 standard: measurement management systems requirements for measurement processes and measuring equipment; – make it your business, first and foremost, to define the real needs of the firm; this is the most difficult step, but the most momentous because it will give the company a choice of solutions and consequently lead to a budget. You have to adapt yourself to today’s needs, but remain aware of what tomorrow will be. Thinking ahead is certainly not reprehensible. There is a real need to define the physical or chemical quantities, as well as the characteristics of the products that the firm is to measure, to set the measuring ranges, and to define the measurement uncertainties with regard to the requirements of a standard of products, of the method of testing or of any other criteria which have to be complied with.
3.2. Goals and role of the measurement management system – metrological function The EN ISO 10012 standard introduces the concept of a “measurement management system” and defines it as a set of interrelated or interacting elements necessary to achieve a metrological confirmation and a continual control of measurement processes. Therefore: – the metrological confirmation of the measuring equipment must be seen; – a control of the measurement processes must be organized.
Mastering Measurement Processes Approach
81
The first point (the metrological confirmation of the measuring equipment) represents the traditional activity of the metrology function of firms as it was conceived a few years ago. The continuous control of the measurement processes has been added to this typical activity of management of a set of instruments. This developments has led to a new definition of the metrological function. It is to be found at paragraph 3.6 of the EN ISO 10012 norm: “Function which is administrative and technical responsibility for defining and implementing the measurement management system.” Consequently, the metrological function will be responsible for the metrological confirmation of the measuring equipment. This operation is defined as follows: – metrological confirmation (EN ISO 10012 section 3.5); – set of operations required to ensure that measuring equipment conforms to the requirements for its intended use. Note 1: metrological confirmation generally includes calibration and verification, any necessary adjustment or repair, subsequent recalibration, comparison with the metrological requirement for the intended use of the equipment, as well as any required sealing and labeling. Note 2: metrological confirmation is not achieved unless and until the fitness of the measuring equipment for the intended use has been demonstrated and documented. Note 3: the requirements for intended use include such considerations as range, resolution and maximum permissible errors. Note 4: metrological requirements are usually distinct from, and are not specified in, product requirements. The EN ISO 10012 standard introduces the notion of measurement process and defines it as: – measurement process (ISO 10012 section 3.2); – set of operations to determine the value of a quantity.
82
Metrology in Industry
Signatories of Multilateral Agreement “calibration”
Austria – BMwA Abteilung I/12 Dampfschiffstrasse 4 AT - 1030 Vienna Tel: 43 1 71 100 8248/Fax: 43 1 71 43582 Belgium – BKO-OBE Federal Public Service Economy/Division Accreditation WTC III - 5th floor, 30 Boulevard Simon Bolivar BE - 1000 Brussels Tel: 32 2 206 46 80/Fax: 32 2 206 57 42 Czech Republic – CAI Opletanova 41 CZ - 110 000 Praha Novemesto Tel: 420 2 2100 4501/Fax: 4202 2100 4111 Denmark – DANAK Dyregaardsvej 5 B DK - 2740 Skovlunde Tel: 45 77 33 95 36/Fax: 45 77 33 95 01 Estonia – EAK Estonian Accreditation Centre Aru 10, Tallinn 10317 Estonia Tel: 372 602 18 01/Fax: 372 602 18 06 Finland – FINAS c/o Centre for Metrology and Accreditation P.O. Box 239 FI - 00181 Helsinki Tel: 358 9 616 7553/Fax: 358 9 616 7341 France – COFRAC Secteur Laboratoires 37 rue de Lyon FR - 75012 Paris Tel: 33 1 44 68 82 28/Fax: 33 1 44 68 82 23
Mastering Measurement Processes Approach
Germany – DKD member of DAR Postfach 3345 DE - 38023 Braunschweig Tel: 49 531 592 83 20/Fax: 49 531 592 83 06 Greece –ESYD Hellenic Accreditation System 8 Sissini street 115 28 Athens Tel: 30 210 7204514/Fax: 30 210 7204500 Holland – RvA Radboudkwartier 223 P.O. Box 2768 NL - 3500 GT Utrecht Tel: 31 30 239 4500/Fax: 31 30 239 4539 Ireland – NAB Wilton Park House - Wilton Place IE - 2 Dublin Tel : 353 1 607 30 03 / Fax: 353 1 607 31 09 Italy – SIT Strada delle Cacce 91 1 - 10135 Torino Tel: 39 011 397 73 35/Fax: 39 011 397 73 72 Latvia – LATAK 157, Kr. Valdemara St LV - 1013 Riga Tel: 371 7 37 3051/Fax: 371 7 36 2990 Lithuania – LA Algirdo 31 LT - 2006 Vilnius Tel: 370 5213 6138/Fax: 370 5213 6153 Norway – NA Justervesenet Fetveien 99 NO - 2007 Kjeller Tel: 47 648 48 484/Fax: 47 648 48 485
83
84
Metrology in Industry
Poland – PCA - POLSKIE CENTRUM AKREDYTACJI ul. Klobucka 23 A PL - 02 - 699 Warsaw Tel: 48 22 548 80 00/Fax: 48 22 647 13 01 Slovakia – SNAS Slovak National Accreditation Service PO Box 74, Karloveská 63 SK - 840 00 Bratislava Tel: 421 7 654 12 963/Fax: 421 7 654 21 365 Slovenia – SA Slovenian Accreditation Smartinska 140 (BTC City, 4.nadstropje) SI - 1000 Ljubljana Tel: 386 (0)1 478 3080/Fax: 386 (0)1 478 3085 Spain – ENAC Serrano, 240 - 7° psio E - 28016 Madrid Tel: 34 91 457 32 89/Fax: 34 91 458 62 80 Sweden – SWEDAC P.O. Box 878, Osterlanggatan 5 SE - 50115 Boras Tel: 46 33 17 7730/Fax: 46 33 10 1392 Switzerland – SAS c/o OFMET Lindenweg 50 CH - 3003 Bern Wabern Tel: 41 31 323 3520/Fax : 41 31 323 3510 United Kingdom – UKAS 21 - 47 High Street Feltham Middlesex TW13 4UN Tel: 44 20 8917 8400/Fax: 44 20 8917 8500
Mastering Measurement Processes Approach
Signatories of Bilateral Agreements
Australia – NATA 7 Leeds Street NSW 2138 Rhodes Tel: 61 29 736 8222/Fax: 61 29 743 5311 Brazil – INMETRO Rua Santa Alexandrina 416 - 90 andar - Rio Comprido CEP 20261-232 Rio de Janeiro Tel: 55 21 502 6531/Fax: 55 21 502 6542 Hong-Kong – HKAS 36/F, Immigration Tower 7 Gloucester Road Wanchai Tel: 852 28 29 4830/Fax: 852 28 24 1302 Israel – ISRAC 2 Habonim Street Ramat Gan 52522 Beit Habonim Tel: 972 3575 1690/Fax: 972 3575 1695 New Zealand – IANZ P.O. Box 914 2142 1136 Auckland Tel: 64 9 525 6655/Fax: 64 9 525 2266 Singapore – SAC-SINGLAS The Enterprise #02-02 No.2 Science Centre Road 609077 Singapore Tel: 65 826 3000/Fax: 65 822 8326 South Africa – SANAS P.O. Box 914-2142 Wingate Park 0153 Pretoria Tel: 27 12 349 1267/Fax: 27 12 349 1249 United States – A2LA 5301 Buckeystown Pike Suite 350 MD 21704-8307 Frederick Tel: 1 301 644 3212/Fax: 1 301 662 2974
85
86
Metrology in Industry
3.3. The measurement processes One of the principles laid down in the ISO 9000 standard lies in the so-called “process oriented” approach. The measurement processes have to be considered as particular processes meant to introduce a support to obtain quality for the products manufactured by the firm. Figure 3.1 illustrates the model of system of management of measurement and provides the references to the different paragraphs of the ISO 10012 norm. 8.4 Improvement
Clause 5 Management responsibility
Customer measurement requirements
Clause 6 Resource management
Clause 8 Measurement Management system analysis and improvement
Customer satisfaction
Clause 7 Metrological confirmation and realization of measurement processes Input
Output
7.1 Metrological confirmation
7.2 Measurement process
Measurement results
Figure 3.1. Model of measurement management system (ISO 10012)
3.3.1. Conception and development of a new measurement process 3.3.1.1. Analysis of the requirements It is vital to accurately and unambiguously define the expectations of the client for the product or service. The marketing, development and research units are consulted to ascertain the expected characteristics of the product or service. These characteristics are then translated into specifications and tolerances that ensure that the product or service is functional and/or interchangeable, and/or that the process can manufacture the product or perform the service required. The specifications are subjected to measurements at the conception, manufacturing and final stages.
Mastering Measurement Processes Approach
87
Paragraph 4.7 of the ISO/CEI 17025 states that: “The laboratory must cooperate with its clients or their representatives to clear up the client’s request and supervise the laboratory’s performance with regards to the work done ...” 3.3.1.2. Transcription of the characteristics of the product in “measurand” form or “characteristics to be measured” form It is important that the characteristics of the product, service or process be transformed into quantities to measure on the product, or into characteristics to test. It is in the firm’s interest to develop the synergies between the “conception”, “quality” and “metrology” functions to translate the specifications into characteristics. The next step will be to verify that the metrological requirements, such as they have been defined, are compatible with the state of the technique and with the firm’s strategy; in particular, the economic aspects will be examined. The purpose of the synergies is that the developer will take the performances, the costs of the measurement and test processes into account. The person in charge of the metrology function will have to be made aware of the critical nature of the characteristics to be measured and he will see that processes are developed that are suitable for the controls of the specifications. 3.3.1.3. The development of a measurement process can be managed as a project It is advisable to manage the development of a measurement process as a project. Quite obviously, there is a link between the importance of the development and the structure of the project, but a few essential conditions have to be met: – Someone has to be in charge of the project. – A specification of the process, stating the goals to be reached, has to be defined: - technical: repeatability, reproducibility, rapidity, etc., - economic: cost of implementation, operation, demolition, life cycle cost, etc., - clear definitions of the input data (quantities to measure, expected uncertainty, rapidity of the process, ergonomics, safety, investment cost, operating cost, etc.). – A clear definition of what the project is supposed to deliver is required (the notion of the existence of a process is not clear enough): it can be a measurement procedure, an instruction, an assessment of the “prototype” process for a given period, or training. – Planning of the development (steps, go/no go stages, etc.), assignment of the tasks and resources.
88
Metrology in Industry
– At the end of the development process, a report (with regards to the defined deliveries and the specifications) corresponding to the characterization of the performances of the measurement process (repeatability, reproducibility, uncertainty, rapidity, ergonomics, costs, etc.) is made, and a decision about whether to put the project into service is taken. It is essential for the firm that the development of the measurement process should accumulate knowledge. The results obtained, and the recording of those results, are important, but the process is important as a way to pass on learning and knowledge. This accumulation of knowledge is a vital factor in the continuous improvement of measurement processes. In no way does this continuous improvement concern the improvement of the result uncertainty, for instance, if the latter meets the expectations; the object of continuous improvement is to reach efficiency, that is to say to do well at a lower cost.
3.3.2. Exploitation of a valid process It seems important for critical measurement processes that a “pilot” be appointed in order to ensure a continuous supervision of the process. Chapter 6 describes the main methods of supervision of measurement processes. Figure 3.2 shows the “pilot” of the process being provided with the available information to enable him to act on the process.
3.3.3. Continuous improvement of measurement processes It would be wrong to think that the aim of continuous improvement is to ameliorate, for instance, the uncertainty of the measurement or test results obtained through the process. The purpose is to improve the control of the process and thus reduce the costs; in short it is to do as well as possible, but more cheaply. Figure 3.2 is an illustration of the information the “pilot” of the process has at his disposal to optimize the process.
Mastering Measurement Processes Approach
Characterization Characteristion Data data method mathod
Data Characteristics characteristics instrument
Characteristics environment
method Method
Instrumentation
Environment
output data result of measurement measuremente
Input data specification
Information
89
Elements o Element of processus processes
manpower Manpower
Measure object
Qualification Continuing continuing education
Batches Batches, manufacture
Numerical value + uncertainty
Input/output Input / output data
Figure 3.2. Information available for the control and the optimization of measurement processes
Measurement process piloting indicators Every process must have its own indicators. They are useful to assess the improvements achieved and the regressions. Some examples of indicators are: – uncertainty of the measurement and test results; – how many times nonconformity has been the result of a fault of the measurement process; – rate of availability of the measurement process; – operating costs of the measurement process.
3.4. Management of the measuring equipment (metrological confirmation) One of the roles of the metrological function is to ensure that all the measuring equipment used in the firm, and likely to have an influence on the quality of the product or the service, are suitable for the task. This is so as to be able to guarantee, with minimum risk, that the measuring equipment as a whole is within the limits of permissible errors. For that purpose the firm must implement a system of management of all its measuring equipment. This system will establish traceability
90
Metrology in Industry
to the International System of Units (SI) and carry out the verification of all the measurement equipment in use. For internal services within a firm, the supplier’s metrological function is to have at its disposal all the equipment necessary for carrying out the calibrations and verifications needed to guarantee the quality of the product or the service. The metrological characteristics of this equipment (measuring range, resolution, freedom of bias, repeatability, etc.) must correspond with the needs of the firm (which can be expressed as a measurement uncertainty). Some of the activities of the metrological function can be subcontracted inside or outside the firm (calibration, maintenance, etc.). However, all circumstances bringing, or bringing back, the measuring equipment into service is the sole responsibility of the person in charge of the metrology function, who confirms that the equipment is suitable for the expected use. To secure the traceability of its reference standards to the SI, the metrological function must resort to subcontracting. In all circumstances it is the responsibility of the metrological function: – to ensure that the subcontractor satisfies the requested demands, for example, through audits or any other method of evaluation; – to limit the choice of subcontractors to only those calibration laboratories that are accredited by the national body in charge of the accreditation of calibration laboratories. In Europe, within the framework of EA (European Cooperation for Accreditation), a multilateral agreement has been signed for the recognition of calibration certificates which have been issued by the laboratories accredited by the organizations that have signed the agreement. The list of accredited laboratories is updated monthly on the Comité Français d’Accréditation (COFRAC) website (www.cofrac.fr). The metrological function must be able to demonstrate at each level of the traceability chain that the traceability to the SI is ensured through an unbroken chain of comparisons. An uncertainty must be associated to each one of the comparisons (see Chapter 2). A firm may resort to subcontracting for the management of its measuring equipment (see Chapter 4).
Mastering Measurement Processes Approach
91
Whichever solution is implemented by the firm, the metrological function remains responsible for the decision to confirm the measuring equipment entering into the quality of the product or the service. The purpose of the management of the measuring equipment is to establish and maintain the measuring equipment necessary to satisfy the requirements of the firm. This management must take into consideration: – the analysis of the requirement, and the selection of the measuring equipment; – the reception, the implementation and the follow-up of requirements; – the traceability to the SI; – the calibration, the verification and the supervision; – the statement of compliance with the requirements (the confirmation). 3.4.1. Analysis of the requirement and selection of the measuring equipments The selection of measuring equipment is made after taking the following factors into consideration: technical requirements, economic and commercial conditions, and evaluation of this measuring equipment. It is advisable to ensure that the measuring equipment meets the requirements of the application in the firm: Conception ------> Tests on the materials or the components Development ------> Tests on the prototype or prototypes Manufacture ------> Setting and supervision of the production tool Control and acceptance of samples Quality control ------> Entrance/exit Marketing ------> Tests of compliance to norms or passed orders At all these levels, the requirements concerning the instruments will be modulated. In some cases, a large resolution will be required, in others it will be a capacity of measurement in dynamic conditions, in others an excellent freedom of bias and repeatability, etc. In fact, the specification of the measuring instrument depends on the needs of the firm. 3.4.1.1. Technical requirements An understanding of the technical needs can be understood from the following points:
92
Metrology in Industry
– The main thing is ensure that the performances and the accuracy class, or the freedom of bias and the repeatability of the measuring equipments, meet the technological requirements of the firm; the restraints of implementation and use (influence quantities, handling, maintenance, etc.) of these means must be taken into account. – A firm’s measuring equipment is often used when assessing whether a product complies with its specification. Therefore, it is up to the user of the equipment to decide whether the measuring equipment will be submitted to a calibration test and then used, the corrections notified in the calibration certificate being applied, or whether a verification will be made, which would set the limits of permissible error as well as some acceptance criteria making it possible to qualify the equipment. – At the time when the decision is made, the homogeneousness of the measuring equipment of the firm can be a deciding criterion if use or maintenance are, for example, taken into consideration. – It is judicious to make a prospective and retrospective analysis of the use of the measuring equipment and its possibilities of evolution so as to limit the risks of obsolescence and, mostly, to keep the firm advised of anticipated developments. – Measuring equipment must be delivered with the information necessary to bring it into service, use it, adapt it or repair it. – When the measuring equipment is new to a firm, or outside its usual scope, it may be important to discuss with the supplier the conditions in which the equipment will be used and the content of the assistance required. – For specific or complex measuring equipment, it is recommended that a file of the specifications be opened with, in particular, definitions for: – the requested characteristics of the measuring equipment; – the conditions of use, environment and maintenance; – the particular requirements concerning the calibration and the verification; – the conditions of acceptance. The following elements show that the firm has the technical information that will enable it to have the measuring equipment adapted for use: measuring range, resolution, freedom of bias and repeatability, parameters ruling the acquisition of data, conditions of traceability to national standards (interval/uncertainty), standards needed to verify that the test or control equipment is fit for use, drafting of the acceptance criteria, which makes it possible to say that the measuring equipment is suitable.
Mastering Measurement Processes Approach
93
Moreover, the user alone is aware of the future environment in which the measuring equipment will be used and of the measurement method into which it will be used (see Chapter 8). 3.4.1.2. Economic and commercial conditions These conditions must be determined jointly by the purchase function and the metrology function of the firm with the following factors in mind: – should the measuring equipment be bought, rented or borrowed, depending in particular on the conditions of depreciation and the risks of obsolescence; – delivery time; – maintenance contract and/or technical assistance; – demands for availability (what time of unavailability allowed, what time for repairs, etc.). It will be the role of the metrological function to provide the “purchase” service, using the technical information about the measuring equipment and its projected use; this will make it possible to justify the cost of one solution or another at the expense of a less expensive option, but which would be unsuitable for the projected use. These ideas are embodied in the following standards: ISO/CEI 17025, ISO 9001: 2000, ISO TS 16949: 2002 and ISO 15189: 2003. 3.4.1.3. Assessment of the measuring equipment The selection of the measuring equipment can also be made from evaluations based the experience of other firms, or that of metrology laboratories. So it might be advisable to obtain all the information or documentation possible to help the firm in its choice. In Europe, three associations of measuring-equipment users have laboratories of metrology and tests to evaluate equipment (France: EXERA (Association des Exploitants d’Equipements de Mesure, de Régulation et d’Autoisme); the UK: EI (Evaluation International); the Netherlands: WIB (Werkgroep voor Instrument Beoordeling) (see Chapter 4).
3.4.2. Receiving the measuring equipment and putting it into service As soon as measuring equipment arrives, the metrological function carries out the following operations.
94
Metrology in Industry
3.4.2.1. Compliance with the order Conformity to the order and to the specifications of the manufacturer or of special instructions is verified; supplied technical documents are checked. 3.4.2.2. Identification of the measuring equipment An identification number is attributed to each piece of equipment. The choice of the codification system may use a classification which makes it possible to group the equipment together in categories, or in relation to type of use. The number will be affixed to the measuring equipment in such a way as to ensure its indelibility. When justified by technical reasons, the number will be affixed on the case of the equipment. The manufacturer’s identification number, if the manufacturer uses one, can also be used. 3.4.2.3. Inventory (description) The identification number makes it possible to develop a permanent and quantitative inventory of all measuring equipment. This inventory is useful when following the technical evolution of measuring instruments, and is also useful in relation to calibration operations, verification or repair, or any other event related to any particular instrument. Depending on the requirements of each firm, the inventory can be in the form of a set of cards, which are called life cards. Some suppliers are marketing software for the management of measuring equipment (see Chapter 4). 3.4.2.4. Technical dossier of the equipment It may in some cases turn out to be useful, upon receiving new equipment, to open a dossier in which all the documents concerning the equipment can be filed (specifications, order, report of receipt, instructions, calibration certificates, etc.). 3.4.2.5. Technical documentation Make sure, when new equipment is brought into service, that all the operators have the information needed for a correct use: copy of the instructions, drafting of the procedures, etc. 3.4.2.6. Basic definitions At this stage, it is a good thing to be able to communicate either with a subcontractor, or with potential auditors and, for that purpose, to master the basic vocabulary of metrology, which is as follows.
Mastering Measurement Processes Approach
95
Traceability (VIM section 6.12) and (ISO 8 402 section 3.15 note b) Property of the result of a measurement whereby it can be related to generally stated national or international standards through an unbroken chain of comparisons. Metrological confirmation (EN ISO 10012 section 3.5 without the notes) Set of operations required to ensure that measuring equipment conforms to the requirements for its intended use. Verification (ISO 9000: 2000 section 3.8.4) Confirmation by clear evidence that the stated requirements have been met. Calibration (VIM section 6.11) Set of operations which establish, under specified conditions, the relationship between the value of the quantity indicated by a specific measuring instrument or measuring system, or the values represented by a material measure or a reference material, and the corresponding known values realized by standards. Note 1: the results of a calibration make possible either the assignment of the corresponding values of the measurand to the indications, or the determination of corrections with respect to indications. Note 2: a calibration may also determine other metrological properties, such as the effect of influence quantities. Note 3: the result of a calibration may be recorded in a document, sometimes called a calibration certificate or a calibration report. Measuring instrument (VIM section 4.1) Device intended to be used to take measurements, alone or in conjunction with a supplementary device (or devices). Material measure (VIM section 4.2) Device intended to reproduce or supply, in a permanent manner during its use, one or more known values of a given quantity. For example: – a weight; – a measure of volume (of one or several values, with or without a scale); – a standard electrical resistor; – a gauge block; – a standard signal generator; – a reference material.
96
Metrology in Industry
Measuring equipment (EN ISO 10012 section 3.3) Measuring instrument, software, measurement standard, reference material or auxiliary apparatus, or a combination of thereof, necessary to realize a measurement process. Uncertainty of measurement (VIM section 3.9 without the notes) A parameter, associated with the result of a measurement, which characterizes the dispersion of the values that could reasonably be attributed to the measurand. Metrological characteristic (EN ISO 10012 section 3.4) Distinguishing feature which can influence the results of measurement. Maximum permissible errors (VIM section 5.23) Limits of permissible errors Extreme values of an error permitted by specifications, regulations, etc. for a given measuring instrument. Accuracy of a measuring instrument (VIM section 5.18) Ability of a measuring instrument to give responses close to a true value. Note: “accuracy” is a qualitative concept. Accuracy class (VIM section 5.19 without the note) Class of measuring instruments which meet certain metrological requirements that are intended to keep errors within specified limits. Bias (VIM section 5.25) Systematic error of the indication of a measuring instrument. Freedom from bias (VIM section 5.26) Ability of a measuring instrument to give indications free from systematic error. Repeatability (VIM section 5.31 without the note) Ability of a measuring instrument to provide similar indications for repeated applications of the same measurand under the same conditions of measurement. Correction (VIM section 3.15 without the notes) Value added algebraically to the uncorrected result of a measurement to compensate for systematic error.
Mastering Measurement Processes Approach
97
Adjustment (VIM section 4.30) Operation of bringing a measuring instrument into a state of performance suitable for its use. User adjustment (VIM section 4.31) Adjustment employing only the means at the disposal of the user. 3.4.3. Calibration and verification operations Both the calibration and the verification operations are based on a comparison to a standard and, except for the preliminary operations, do not include any intervention on the measuring equipment. They are indispensable operations, which make the indications provided by the measuring equipment meaningful. The result of a calibration comprises all the values which have got out of the comparison between the measurement results of the equipment and the standard. The calibration, in the strict sense of the VIM, will generally result in a calibration certificate with a view to applying corrections to the measurement results afterwards; exploiting them will make it possible to decrease the uncertainty of the measurements taken with the equipment. These uncertainties about the values of the corrections will also be used when assessing the causes of the uncertainties so as to determine the compound uncertainty that will be connected to the measurement results (see Chapter 7). The result of a verification makes it possible to assert that the measuring equipment meets, or does not meet, requirements that had been set beforehand (generally as tolerated error limits which allow the measuring equipment to be brought, or brought back, into service). A verification can then be made either by: – comparing the results of a calibration operation with the tolerated error limits; – materializing the tolerated limit indications of the measuring equipment that it is compared to directly by means of a standard. This method does not require figures. The result of a verification can be either: – a record of verification, which means for the user that the equipment can be brought back to service; or – a decision to adjust, repair, scrap or downgrade the instrument, materialized by a appropriate mark indicating the state of the measurement equipment.
98
Metrology in Industry
We wish to draw the reader’s attention to the need to adapt the verification program (measurement points, distribution in the measuring range, etc.) to the use intended of the instrument (see section 3.4.3.2, note 4) rather than verifying the compliance with the manufacturer’s specifications because what matters is that the instrument should be fit for use.
Customer2 1 2
C a lib ra tio n (te c h n ic a l c o m p a ris o n o f m e a s u rin g e q u ip m e n t w ith a m e a s u re m e n t s ta n d a rd )
Calibration
N e e d Id e n tifie d S ta rt
C a lib ra tio n C e rtific a te /R e p o rt
M e tro lo g ic a l R e q u ire m e n ts E x is t?
Yes
Recalibration Loop
Metrological Verification
C a lib ra tio n S ta tu s Id e n tific a tio n 1
E q u ip m e n t C o m p lie s W ith R e q u ire m e n ts ?
Y es
Is A d ju s tm e n t O r R e p a ir P o s s ib le ?
No
V e rific a tio n / C o n firm a tio n D ocum ent
No
V e rific a tio n Is N o t P o s s ib le
No
Decisions And Actions
METROLOGICAL CONFIRMATION PROCESS
The different operations for calibration and verification are shown in Figure 3.3.
T e s t R e p o rt: V e rific a tio n F a ile d
C o n firm a tio n S ta tu s Id e n tific a tio n
Yes
A d ju s t O r R e p a ir
S ta tu s Id e n tific a tio n
R e v ie w C o n firm a tio n In te rv a l
R e tu rn T o C u s to m e r End
C a lib ra tio n id e n tific a tio n /la b e llin g m a y b e re p la c e d b y m e tro lo g ic a l c o n firm a tio n id e n tific a tio n . O rg a n iza tio n o r p e rs o n th a t re c e ive s a p ro d u c t. E xa m p le : C o n s u m e r, c lie n t, e n d -u s e r, re ta ile r, b e n e fic ia ry a n d p u rc h a s e r. N o te : A c u s to m e r c a n b e in te rn a l o r e xte rn a l to th e o rg a n iza tio n (re f. IS O 9 0 0 0 :2 0 0 0 § 3 .3 .5 ).
Figure 3.3. Diagram of metrological confirmation
Mastering Measurement Processes Approach
99
3.4.3.1. Calibration or verification program The technical comparisons program is a document that makes it possible to take into accounts all the operations to be carried out on the measuring equipment. It is arranged depending on the calibration or verification program set for each measuring instrument, on when each measuring instrument is easily available and on the work schedule corresponding to the tasks to be done. 3.4.3.2. Calibration or verification intervals Whichever measuring equipment is considered, a systematic process of comparison done at set intervals ought to make it possible to prevent any weakening of the quality of the measurements taken and to ensure the equipment’s credibility over time. To determine the interval of the comparisons (calibration or verification), it is necessary to take into account such factors as the rate and type of utilization, the expected drifts in view of the acquired experience, the nature and wear of the equipment, possibly the economic, normative and statutory restraints, etc. The interval initially determined for a given measuring equipment must be reconsidered and, if necessary, readapted according to the experience that has been acquired. Note 1: proceeding to limited controls within the set period is not to be ruled out; they make it possible to detect any glitch at the measurement points that are normally used. In no way can these controls replace1 the planned calibration and verification operations (see Chapter 6). Note 2: any intervention likely to alter the metrological characteristics makes it necessary to examine the initially determined interval. Note 3: some measuring equipment is used only now and then; the strict periodicity rules are not to be applied to them. In those circumstances there should be written instructions that the instruments be submitted to comparison operations before they are used if the validity period of the previous comparison has expired. Note 4: some equipment is only used for one or a few of its functions; it can be agreed to calibrate (or verify) the equipment only for the function or functions used. In this case, the equipment has to be identified so as to avoid any risk of error if they occasionally were used for a non-calibrated (or non-verified) function; clear mention of use restrictions must be stated on the equipment. 1 In some measuring processes it can be considered that if the measuring process remains “under control”, the calibration at set intervals is not required.
100
Metrology in Industry
3.4.3.3. Supervision of the measuring equipment Measuring equipment is the essential element of measurement processes. Different methods are proposed in Chapter 6 for the supervision of measurement processes and equipment. The idea of supervision has been developed in order to prevent malfunctions, drifts between two calibrations or verifications. The reader should read the EN ISO 10012 standard “Measurement management systems – requirements for measurement processes and measuring equipment”.
3.4.4. Fitness for use of measuring equipment Just as one has to periodically make sure that employees still have the qualifications required to perform the task(s) required – one cannot rely on the initial training and the diploma possibly obtained – likewise, it is important to ensure that the measuring equipment, which contributes to the quality of the product or the service, still possesses the performances and characteristics required to what it is meant to do, or what it is planned that it will do. 3.4.4.1. Freedom from bias, repeatability, stability Three metrological characteristics are essential for measuring equipment: – freedom of bias (VIM section 5.26); – repeatability (VIM section 5.27); – stability (VIM section 5.14).
••• •••• •••
•
Repeatability
• • • • • • • • • • • • • • • •
Freedom of bias
Figure 3.4. Repeatability and freedom of bias
Mastering Measurement Processes Approach
101
The traceability to standards will make it possible to know the value of the corrections to make to indications of the instrument to compensate for its biases. The repeatability of the instrument will be assessed by repeated observations of the same measurand; when assessing the repeatability of the instrument one must take care not to introduce fluctuations coming from the measured quantity. The evaluation of the repeatability can be made by, for example, measuring a standard. You have to be aware that the repeatability you will find that way will generally be better than the measuring process, because other factors of variability come into the measuring process. The stability will be noted of by watching the calibration results obtained at given intervals. These three characteristics have to be supervised by the firm’s internal metrology function. 3.4.4.2. Maximum permissible errors The above data materialize the limits that can be set to start the operations of user adjustment, but it is sometimes preferable to set more restrictive limits if you do not want to have to proceed to corrective actions when a verification reveals that a piece measuring equipment does not meet the specifications (see Chapter 6 for the methods of supervision of measuring equipment and measurement processes). 3.4.4.3. Demands for an assurance of the quality The demands for quality assurance clearly indicate that it is necessary to regularly keep track of the measuring equipment. These demands are made clear in paragraph 7.6 of the ISO 9001 standard – “Control of the measurement and supervision devices” – and in paragraph 5.5 of the NF EN ISO/CEI 17025 standard – “Equipment”. The ISO standard of the 14000 series concerning the system of environmental management states that the firm should measure, supervise and assess its environmental performances (ISO 14004 section 4). Why? As it has been seen, measuring equipment requires that its drift should be supervised so that its indications can be brought back within the tolerated limit of errors, by means of a user adjustment or an adjustment. This can also be achieved by applying corrections. Supervising the drift is equally essential because if the errors become considerable the indications of the instrument might lie outside the tolerated limits of errors.
102
Metrology in Industry
3.5. Setting up a metrological structure within the firm 3.5.1. Analysis of the metrological requirements and setting up standards A look at the inventory of the measuring equipment will make it possible to group the equipment according to the three following criteria: – measured physical quantity; – measurement field; – freedom of bias and repeatability. The analysis of these groups reveal three typical cases. Case of one instrument only Generally, buying reference standards to calibrate only one measuring instrument will not be contemplated. The easiest and most efficient solution will be to request a calibration laboratory to calibrate the equipment; this will ensure its traceability to the SI. Either a national laboratory of metrology or a calibration laboratory accredited for the quantity and for the measuring range expected would be acceptable. Case of equipment of widespread use in the laboratory It will be possible, with the help of the inventory of the measuring equipment, and taking the measuring ranges and uncertainties into account, to define the standards needed to calibrate and verify such measuring equipment. Let us take a particular case to illustrate this point: the calibration of voltmeters. When there are a large number of voltmeters in a laboratory, it is better to use a tension generator whose calibration will be entrusted to a laboratory – it makes it possible to ensure traceability – the competence of which is guaranteed by accreditation, rather than send away each one of the voltmeters for calibration. Several benefits are derived from this type of organization: less expense, shorter immobilization periods and the possibility of using a local reference if there is a doubt about a measurement (metrological redundancy). Case of measuring and testing equipment where the connection to physical quantities raise technical problems It is the case when those measurements result from the application of conventional methods. Two types of approach are possible: utilization of reference
Mastering Measurement Processes Approach
103
materials or interlaboratory comparisons; in some cases, the two approaches can complement each other. In paragraph 5.6, “Traceability of measurement”, of the NF EN ISO/CEI 17025 norm stresses that: (...) there are calibrations which cannot at the present time be strictly performed in an SI unit Calibration, in such cases, must introduce confidence into the measurements by establishing traceability to appropriate measurement standards such as: – the use of reference materials – it must be certified they are from a competent supplier – to characterize physically or chemically a material in a reliable way, – the use of specified methods and/or standards chosen by consensus, clearly described and accepted by all the parties concerned. Taking part in an appropriate program of interlaboratory comparisons is required whenever it is possible.
In the case of physical methods of chemical analyses (chromatography, spectrometry, etc.), the pre-analysis operations compulsorily include an operation known as calibration or gauging which implements solutions obtained by the laboratory or by reference materials (see ISO 32 guide, “Calibration in analytical chemistry and utilization of certified reference materials”). A procedure has to be established when the firm uses reference materials; this makes possible the control, the implementation of a new sample of reference materials and the answer from the measuring equipment when two samples of reference materials are used. The criteria that have led to the decision to renew the reference material must be in writing. Case of the measuring equipment that cannot be connected to an accredited calibration laboratory Credibility of the measurements will be sought by means of comparisons and cross-checking between laboratories. Contact can be made with the national institutes of metrology and even foreign laboratories may be used to do the calibration, within the scope of EUROMET; the national metrology institutes collaborate and are in a position to direct the requests toward laboratories that can satisfy them. A Mutual Recognition Agreement (MRA) of the standards and the calibration and measurement certificates issued by national laboratories was signed
104
Metrology in Industry
in 1999. See the Bureau International des Poids et Mesures (BIPM) website (www.bipm.org).
3.5.2. Traceability of the measuring instrument(s) to the firm’s reference standards The traceability to the firm’s reference standards determines whether a measurement result can be connected to appropriate standards through an unbroken chain (see traceability in section 3.4.2.6). The traceability of the measuring equipment to the firm’s reference standard can be achieved through a working standard. There does not have to be a working standard; it will depend on the technology of the instruments and the conditions of their use. The number of intermediary firm’s reference standards must be chosen in such a way as the degradation of the uncertainties caused by the use of successive standards is compatible with the uncertainty which is obtained by the measuring equipment: a judicious choice should make it possible to obtain a chain of standards well adapted to the intended application as regards their uncertainties, their stableness and their domains of use. Note: if there is no chain of standards, the traceability can be done through fundamental constants, by the methods of reference measurement (chemical analysis, for example) or by using reference materials. Reference materials make it unnecessary to move an instrument: the reference material is the metrological information medium. For example, a viscosimeter can be calibrated if it is sent to a calibration laboratory, but the user can calibrate it himself by using standard oil (reference material) which will, beforehand, have been calibrated by an accredited laboratory.
3.5.3. Traceability of the firm’s reference standards to the SI The purpose of the connection to the SI is to make sure that a measurement result obtained at one point on the globe is unquestionably comparable to another measurement result obtained in the same conditions at another geographic location. The organization of metrology at national and international levels is intended to guarantee consistency between the standards of the different nations and to ensure that the deviations which occur are not significant at the level of the measurements made in the firm.
Mastering Measurement Processes Approach
105
The traceability to the SI of all the equipment which can influence on the quality of the product has to be guaranteed. The connection of the firm’s references to the SI is comprised of the following operations: – external calibrations of the firm’s reference standards, which guarantee their connection to the SI; – internal calibrations of the working standards. For either operation, a connection program sets the list of equipment involved, the interval between calibrations, the points to be calibrated and the possible requirements. This program can be drawn up with the help of a national laboratory of metrology or an accredited laboratory. Note: the optimization of connection programs is one of the major tasks of the metrological function. This optimization must be: – technical: uncertainties are to be optimized; – economic: the costs of the calibrations (traceability program and periodicity) are to be optimized. When the traceability of the measurements to national standards or to the SI units is not feasible, the firm’s metrological function must be in a position to demonstrate that the measurement results are correlated; it can be done, for example, by taking part in national or international interlaboratory campaigns. It would be wise, in any case, to look into the ratio between the uncertainty of the calibration of the equipment and the measurement uncertainty requested by the firm
3.6. Suggested approach for setting up a metrology function It is important not to set up a metrology function at random; the order of the operations can be of some importance. A suggested approach is as follows: – to nominate someone to deal with this operation. However, the person must know the firm and its techniques very well; it would not be a good idea to entrust a trainee or a new employee with this task; – to analyze your real needs for information from a measurement or test result; – to make a list of your measurement processes and choose those you regard as critical;
106
Metrology in Industry
– to make an inventory of the measuring means already found in the firm (identification, localization, etc.); – to open a technical file in which to store the information related to these instrument (instructions, certificate, order copy, etc.), for comparatively important instruments; – to analyze your manufacturing processes and testing methods, then pick out the instruments which play an essential part in controlling the processes, or in demonstrating the quality of the products. These instruments are the ones you must deal with first; – to analyze your measuring processes and determine the uncertainty of your measurement results; – to analyze your products, testing methods and manufacturing processes, then verify if your measuring processes are appropriate to your intended objectives (ratio tolerance/uncertainty); – to supplement your equipment when necessary; – to think of the different possible traceability patterns for each type or each instrument; try to optimize them economically and technically (ease of use), the uncertainty being adapted to the needs; – to send your reference standards to accredited laboratories for calibration and optimize your calibration intervals; – to examine and write down your procedures of calibration, of verification of your own instruments; establish supervision methods for your measurement processes; – to put in writing all the measurements you take; – to analyze the malfunctions and your errors; to take steps to ensure that they do not happen again; – do not forget that perfection is out of reach; what is sought is to establish a system that will enable you to make progress.
3.7. Bibliography International vocabulary of basic and general terms in metrology (VIM) ISO-IEC-FICCIUPAC- IUPAP-OIML-BIPM 1993 ISO 10012 (2003) Measurement management systems – Requirements for measurement processes and measuring equipment ISO/CEI 17025: 1999 General requirements for the competence of testing and calibration laboratories
Mastering Measurement Processes Approach
107
EURACHEM/CITAC Guide 2: Quality Assurance for Research and Development and Nonroutine Analysis (1998) EURACHEM Traceability in chemical measurement (2003) ILAC P10:2002 ILAC Policy on Traceability of Measurements Results EA-4/07 (rev 01): Traceability of Measuring and Test Equipment to National Standards (previously EAL-G12)
This page intentionally left blank
Chapter 4
Handling of a Bank of Measuring Instruments
The object of this chapter is to suggest an approach to the implementation of a bank of measuring instruments, with a double purpose: – to give confidence in one’s own measurement results; – to show one’s clients that the measurement processes are controlled. There is no particular chronology to follow, except that one should start with the inventory, define the responsibilities, etc. However, this chapter has been written with certain logic, following the order of successive steps. The time needed to initiate the handling of a bank of measuring instruments is also stressed. It is a long process that you cannot complete in a couple of months, unless there are only a dozen instruments or the handling is fully subcontracted; and even in the latter case, subcontracting specifications will have to be drafted, the subcontractor will have to be found, you will have undertake thorough quality audits to ensure that the subcontractor is competent, and responsibility for the follow-up of the metrology function will need to be given to somebody in the firm. Do not forget that it takes time to analyze the measurement requirements and to select the suitable means, and that is something that is not subcontracted. Initiating the handling of a bank of measuring instruments has to be done with the desire to improve the current organization of the firm while taking the firm’s culture into consideration.
Chapter written by Jean-Yves ARRIAT – Ascent Consulting.
110
Metrology in Industry
Throughout this chapter, the term “measuring” is used in the broad meaning of “measuring, checking, analyzing and testing”, as in each of these actions the result is obtained through measuring equipment which has to be looked after.
4.1. Acquaintance with the bank 4.1.1. Inventory The first step is to draw up a complete list of the measuring equipment, without omitting those which are never used (the question of why some are never used can then be raised) and those no longer in working order, and including the gauges, templates, height gauges, etc., as well. You must take advantage of this step to build up contacts with the users; knowing them with an ability to sense their problems will turn out to be very useful later on. At the same time, you should take note of the assignment (to places and/or persons) of the measuring instruments and of the people who keep them (in the case of statutory-use measuring instruments). The inventory of the material is very useful for several reasons: – the importance and the size of the bank make it possible to define the policy that directs the metrology function; – it is used as a database when a new instrument needs to be chosen; – it may save buying new instruments if some are not used; – furthermore, it is indispensable within the context of contractual relations, for those instruments supply results concerning the quality and conformity of products.
4.1.2. Identification After you have listed all the measuring equipment, you have to identify them in a concrete form. It means you have to define a code system. For example, you could take numbers in numerical order, from 1 onwards. You could also make the allocated number more significant; for example, you could assign the numbers: – from 0001 to 0999 to the metrology laboratory; – from 1000 to 1999 to the testing laboratory; – from 2000 onwards to the workshops.
Handling of a Bank of Measuring Instruments
111
You could also use a combination of letters and figures; for example: DG 1117 which would mean: DG: depth gauge, - - 1 - - -: assigned to the testing laboratory; - - - 1 - -: assigned to the laboratory of ground mechanic-testing; - - - - 17: 17th gauge in the laboratory. You could also use a two-part number: 000 - 0001, where first part (on the left) would be the category: – Series 100: mechanics category; – Series 200: electricity category; – Series 300: weighing category; – do not forget the “others” category. This code system makes the management of the codes easier when the handling of the bank is computerized. The main thing is to establish a clear, simple system, preferably one that can be used for the codification of the documents related to the measuring instruments (see section 4.3.1 below). The individual number of the measuring instrument, provided by the manufacturer, can also be used. Almost all measuring instruments have an identification number provided by the manufacturer. Even if this number is not relevant for the firm’s identification system, it appears on the instrument, which can spare trouble when marking instruments. The identification must be clearly affixed, preferably with a mark or label fixed on the instrument without altering it; if the marking is engraved, you must be careful about which method is used. It may also be helpful to identify the instrument’s container, especially if it also contains documents or data useful for the operation of the instrument. In the same way, in case the data about the periodicity of the followup (e.g. date of the next calibration) cannot appear on the instrument (e.g. because of lack of space), the data could appear on the container, on the condition that the container remains in sight of the instrument and mentions its reference. In most cases, a label is simply affixed to the instrument; depending on its size, it indicates the instrument number, the date of the last calibration and the date of the next one. The periodicity can be seen immediately by using labels of different colors; for example: yellow = 6 months; blue = 1 year; green = 2 years.
112
Metrology in Industry
The date can be mentioned in “week – year”; the individual number and the identification sheet make it possible to easily go back to the verification or calibration report. Here are two models of labels:
Last calibration: Instrument number: Next calibration:
or more simply:
06/02 M/Y
perhaps in green, to indicate the conformity M for month and Y for year
Though it is easy to use, the label may sometimes not be the ideal solution because it may come unstuck. However, much progress has been made in this area, and a little effort may allow you to uncover a good solution. When there are many measuring instruments to handle, some firms use bar codes which are stuck straight on to the instrument by means of a label. It is an attractive solution, but it involves risks. The bar code refers directly to the data-processing unit for the whole of the information concerning the instrument; it also requires a very advanced computerized management and the ownership of bar code scanners (in good working order) by the users; finally, it makes the follow-up of the instrument anonymous (which runs counter to the users being made to feel responsible). Nevertheless, these difficulties can be circumvented by putting the individual number of the instrument near the bar code. When a firm subcontracts the handling of its measuring material to an outside service company, it is important that the contract should specify which of the two parties is responsible for the marking. It does sometimes happen that there is no marking (each party thinking it would be done by the other), or that the service company simply attaches a label with an arbitrary date for the next visit without consulting the firm. The firm can define its policy about the handling of the metrology function before or after proceeding materially to the identification of the instruments. In any event, both marking and identification have to be done right at the beginning, after the inventory.
Handling of a Bank of Measuring Instruments
113
4.2. Metrological policy of the firm 4.2.1. Objective and commitment of the firm’s management The firm must clearly state what objective it wants to reach: for example, to satisfy the demands contained in the ISO 9001 norm, to obtain an aeronautical acknowledgement of the JAR 145 type, or to become qualified for the QS 9000 (American motor referential) or the ISO TS 16949. The firm then defines the objective of the metrology function. It has to decide, from the information gathered during the inventory, whether it wants to do everything internally, or subcontract part of it, or all of it. According to what it has chosen, it draws up a plan of what has to be done and defines the responsibilities of the various people who are to intervene as well as their “sphere of influence” and the functional connections.
4.2.2. Plan of actions to launch Once the objectives have been defined and the commitment has been clearly stated, the actions to be launched must be specified. The list of the missions to be carried out will be established, and a degree of priority for each one will be indicated. As far as it is possible, it is worth trying to estimate the time which is needed to perform each operation. This makes it possible: – to draw up a schedule and a work program; – to assess a part of the cost for launching the firm’s quality system. Then someone has to be made responsible for each action; one person can be in charge of several actions: heading the metrology function, identifying the material, drafting the documents, training the users, assessing the capability, verifying the instruments, etc.; he or she must, however, make sure that the documents are verified and approved by another person.
4.2.3. Awareness, training and vocabulary You have to make sure, before any action, of the personnel’s adherence. You will then have to start informing people and making them aware of the importance of looking after the measuring instrument; they will be reminded that natural drifts are possible, that uncertainties are related to measurement results, that it is important not to believe spontaneously in a “top level electronic” instrument, etc.
114
Metrology in Industry
When you analyze what the firm needs for the handling of metrology, you must not forget to train the person in charge; among other things, he or she should have technical knowledge in metrology, in quality handling and about the notion of traceability. He or she should also ensure that the users of the measuring instruments have the necessary ability to use the material. A training program, adapted to the needs of users, will have to be set up, both for the person in charge and for the users. It is usually at this stage that a number of difficulties arise, often linked to vocabulary problems. People talk about the same subjects but with different words: they do not understand one another, or worse, they agree on terms they understand differently. Therefore, it is of paramount importance to rapidly define the meaning of the words to be used, and especially the words: standard, calibration, verification, gauging (“calibration” in French, so confusion can arise with “calibration”; the English translation of the French “étalonnage”), mean of the measurement results, uncertainty, etc.
4.2.4. Selection of the material to be followed periodically Faced with all the demands one is supposed to comply with, it is easy to panic and consider that it is too heavy, constraining and expensive a job; as a rule, in this situation, one may prefer to do nothing. On top of this, one wonders, quite rightly, whether the same strict handling applies to all the measuring instruments, apparatuses, gauges, sensors, etc. The answer is obviously no, all the measuring instruments are not handled in the same way, although this position is far from being unanimously accepted among metrologists and quality managers. This is for a very simple reason: the cost of the operations. However, all the means have to be seen to, but not necessarily all in the same way. Some are merely listed in an inventory, others are followed with normative strictness, complying with ISO 10012. What are the criteria which can be selected in order to perform the sorting out? The main question to ask is: “how important is the measurement which is to be carried out as far as the contractual requirements of quality assurance, security and safety are concerned?” All the instruments that fulfill these requirements should be followed very strictly. Let us underline the importance of carefully defining the contractual requirements (and reading attentively the documents attached to the contract; for example, the military American military norm 45662-A does not leave much room for instruments that are not followed), of not guaranteeing a lot of parameters on a product if you do not look after the instruments which are used and if you do not record the results obtained.
Handling of a Bank of Measuring Instruments
115
Concerning the other measuring instruments, which are not subjected to constraints from outside the firm, you should ask yourself: “what will be the consequence of an undetected drift of my instrument?”, then assess how likely it is this “risk” will occur and compare the risk to the total cost of a follow-up. This leaves us with two categories of measuring instruments: – for those that are strictly handled over time, all the requirements are applied to them; – for those whose handling is not subjected to a plan, they may simply be listed. You must not, for this second category, let metrology indulge in the free and easy attitude of former times: a minimum of work should always be undertaken. The instruments must all be listed in an inventory, even if they have to be put together in series (this is especially the case in the chemical industries which can have 10,000 temperature gauges) and whenever it is possible, the importance of the measurement in the process will be determined. This material will never be followed over time; it will be either on account of doubts the users might have or only when it is first verified before it is put into service. This strictness makes it possible to eliminate all the useless measuring equipment from the firm: they are sources of errors, conflicts between the users, unnecessary immobilization (an important element these days when uncontrolled costs are hunted down), etc. However, you must not forget to clearly identify the instruments of this second family, so that they are not thought to be periodicallyhandled equipment that have lost their labels. 4.3. Drafting of the documents 4.3.1. Codification of the documents The efficiency of the handling of a bank of measuring instruments cannot last if the handling is not formalized, especially within the context of a “quality management” process; but before the documents are drafted, it is important to list those you need (it is advisable to refer to Chapter 8, section 8.3) and think of structuring their relationship with the firm’s documentary system. It is vital to attempt to present and codify them along the same principle as the documents of the firm’s quality assurance system. Given the vast amount of “paper”, only useful documents should be created and they have to be clearly identified. Let us go back to the example at the beginning of the chapter; if you use “DG” for the depth gauge, you will use “DG” in the codification of all the documents concerning the measuring instrument: ISDG1117: identification sheet of the 17th depth gauge of the laboratory of ground mechanic.
116
Metrology in Industry
It is not essential to use an abbreviation of the name of the instrument, but it is useful to create differences between the classes of documents: – CBI- - - -: calibration instruction no. CR-0201: 1st calibration report in 2002 – TSI- - - -: test instruction no. TR-0269: 69th test report in 2002 – VFI - - - -: verification instruction no. VR-0275: 75th verification report in 2002 The identification sheet mentions the references of the instrument in question, as well as the number of the report that contains the results which have been taken into consideration to authorize the instrument being put into service again. The report mentions the references of the verification instruction which was used to proceed to the said verification. The codification is important because it enables you to find your way through all the documents. Thinking things out a bit when finalizing the document can make the work much easier. After you have defined the codification for the documents, you have to draft them. These make up two different categories: the work instructions and the documents which will show the results.
4.3.2. Work instructions It is important to emphasize here that this approach is only one way to proceed. The first work document is the general procedure for dealing with the measurement processes. It provides the outlines of what is to be done and refers to the work instructions for further details. You have to set out how the material is identified, the meaning of the labels when necessary. That is the identification instruction. It enables all the services and shops to identify the material similarly. You also have to document: – the instruction that sets the intervals for the periodical follow-up of the material over time; – the instructions about the verification of the measuring means to define the way each category is verified; – the calibration instructions for the metrological references which have to be calibrated; – the instructions about upkeep and maintenance, when the materials would be put at risk if these operations were not done correctly.
Handling of a Bank of Measuring Instruments
117
It is advisable to have one instruction per category of measuring instruments: it makes it possible for the documents to evolve more easily as a function of the various demands (normalization, clients, etc.). If possible, get ideas from existing norms and from suppliers’ advice. Regarding the verification of the measuring instruments, it is often difficult to thoroughly apply all that the norms prescribe. Do not worry too much about it. To begin with, the main thing is to define what you want to do and stick to it. If what you have decided turns out not to be enough, work on it to further it. You should give an instrument only the time it requires; this is dependent on how useful and important it is. The technical content of the instructions must take the users’ standard into consideration. As a rule, the users are technicians whose basic standard is reasonably good, so the documents can be simple. However, the instructions have to give plenty of details if the personnel are not well-trained.
4.3.3. Result-recording documents At the same time as you define the work instructions, you have to define the documents in which the results are recorded. The most important one among these is the identification sheet. To draft it, you can get your inspiration mostly from the national norm, when there is one. There is one identification sheet per measuring instrument and it holds all the information about the life of the instrument in question, especially: – the name of the instrument (or standard) and its individual identification; – the name of the manufacturer; – the date of its receipt and setting up; – its usual location, if the question arises; – the account of the interventions it has been subjected to, by referring to the documents containing the details of the operations and the figures of the results; – the maximum length of time between two successive calibrations (periodicity); – the references of the work instructions (verification, maintenance, etc.) to be used. The identification sheet does not contain the detail of the operations which have been performed; it only indicates the result. If it is possible, only one type of identification sheet should be used in order to facilitate the use of the documents.
118
Metrology in Industry
It is advisable to file the sheets by spheres of activities or categories; depending on the size of the bank, the place of use can also be taken into account. Hence, the classification can be by spheres of activities: mechanics, electricity, optics, chemistry, temperature, time, etc. By category, there are the following: – dimension (length, surface, angle, etc.); – mass, force, pressure, hardness, resilience, roughness; – time (hour, frequency, duration); – flow (liquid, gas); – volume (gauging); – acoustics; – non-destructive testing; – electricity (potential difference, current intensity, power, resistance); – thermodynamic temperature; – light intensity; – quantity of matter; – molecular composition (spectrophotometry); – chemical analysis (acidity, etc.); – and many others. The other documents which have to be formalized are, according to the firm’s particular needs: – calibration certificates; – verification reports; – test reports, maintenance reports, etc. The reports are the documents that contain the details of the results obtained, step by step, and whose conclusion generally appears in the identification sheet. A template report should be established at the same time as the work instruction it refers to.
4.3.4. Other documents At the stage when a system for handling a bank of measuring instruments is set up, what has been achieved is both the easiest and the hardest parts: the hardest part because it is never easy to lay the first stones of a construction as they are the ones upon which the stability of the work rests; but the easiest part, too, because the first steps are simple, understandable by everybody and, more and more often, demanded by the clients, who have contributed to their development.
Handling of a Bank of Measuring Instruments
119
Beyond that, only what is strictly necessary should be documented. A craze for documents often arises, especially if the first ones have been launched easily. Some exception rules, if they are peculiar to the firm, can be documented in writing; they are usually called “internal norms” and might be a formalization of the processes of physical handling of the measuring instruments, such as they are described in the next paragraph. 4.4. Physical handling of the measuring instruments 4.4.1. Receipt The process of acquisition, receipt and implementation of new equipment should be defined. As soon as a new means of measurement is delivered, you should make sure the following operations are adhered to: – to verify that the equipment conforms to the order, the manufacturer’s specifications or to particular prescriptions; also, do not forget to check the technical documents that are provided; – to identify the means of measurement (with a registration number, for instance); – to introduce the means into the inventory; – to calibrate, or initially verify before the implementation (possibly done by the manufacturer), which makes it possible to determine the class of the instrument; in cases of mass and (static or dynamic) volume measurements, it is generally the approval of the model which defines the class, for reasons related to the demands of legal metrology; – to inscribe a mark concerning the calibration or verification and thus start the periodicity. Until these operations have been completed, the equipment must not be implemented except, exceptionally, in an emergency; even then, it must be handed back as soon as possible, in order to be put through the correct steps. Concerning the initial verification, you should do this yourself, or have it done by a laboratory which has the required competence; ideally, however, this should not be done by the supplier, unless he or she can give guarantees of his or her impartiality. It is important to say who in the firm is in charge of acknowledging receipt of the measuring instruments (whether new or being returned); and who replaces him or her if he or she is absent. Actually, a badly-controlled handling of receipts rapidly leads to disorder in the handling of the bank, because it is very likely these materials will be put into service without first being identified or verified.
120
Metrology in Industry
4.4.2. Transfer On top of these processes of receipt and follow-up of the material, it is important to perfect control of all the operations concerning the transfer of the metrological equipment: their entry in/out of laboratories or shops, dispatching, changing assignments, occasional moving, etc. These operations should be subjected to particular procedures that state what the possibilities of transfer are and, as the case might be, their limits, as well as which precautions should be taken. 4.4.2.1. Traceability In order to know at any moment the state of the bank of the measurement means, it is vital to ensure a traceability, which should be both satisfactory and adapted to the firm’s requirements, of all the components of the bank. It should be possible to locate all the instruments, name the person responsible for them, and know their latest places of assignment or use according to the contractual importance of the measurements made or the cost of deviation in the case of wrong measurements. Traceability of measuring instruments and measurements also means being able to determine which instrument has been used to make a particular measurement. This should make it possible to obtain: – a good progress of the program of calibration and/or verification; – the detection of the measurements which need to be checked or done again in case a deviation in the operation of the instrument is revealed during a calibration or verification. Let us take the example of measurements made on testing benches. There are three benches and one of them turns out to be faulty. If measurements have to be redone, it is indispensable to identify the benches in order to repeat only those measurements which were made on the faulty bench. Depending on the importance of the bank of instruments and the size of the equipment, it might be a good working technique to establish a computerized procedure of the “outgoing equipment ticket” type, which would make it possible to know all the transfers. 4.4.2.2. Transfer Any transfer must be performed under someone’s responsibility. Transfers have to be controlled so that the equipment scheduled for maintenance may be called in due course, without disrupting the program of the measurements to be made. Several systems are possible.
Handling of a Bank of Measuring Instruments
121
For example, there could be: – a computerized automatic call in the case of a computerized file; – a call per type of instrument, with the purpose of planning calibration campaigns on metrological themes; – a user sector could be made responsible – its mission would be to keep the verification of its means of measurement up to date. In all cases, the follow-up of the procedure has to be ensured; it might be necessary to send reminders. However, you have to be sure that taking the means of measurement away from the user services can be done without its absence causing a disruption in the operation of these services; if not, appropriate arrangements (e.g. official derogation on the verification date, or supply of a replacement means) have to be made in cooperation with the users of the means. 4.4.2.3. Precautions Every time the means of measurement have to be transferred from the place of use to the place of calibration, or vice versa, or between the place of storage and the place of use, appropriate precautions have to be taken. The conditioning of the measuring instruments is well defined and the transfer is subject to instructions which are pre-established and which concern handling, packing, transportation, and maybe, intermediate storage. Some elements of the equipment may have to be secured before the transfer, for example, the arm of a measuring column, the arm of one-pan scales, etc. The accesses to the adjustment devices which may affect the performances should be protected so that untimely or accidental handling is prevented. This does not concern the devices which are meant to be accessible to the user without any outside help; that is the case, in particular, for zero adjustments. Instruments which are subjected to regulations are protected by lead seals whose location is indicated in the model approval. You must ensure that the seals are unbroken. No uncontrolled intervention by the state should be performed on these instruments.
4.4.3. Storing and environment To successfully carry out the processes of storing and control of the environment, a number of operations have to be undertaken and followed up: – to provide suitably fitted-out safe storing areas or premises to prevent the equipment from damage or premature deterioration;
122
Metrology in Industry
– to define suitable methods to allow receipt into and dispatch from these areas; – to have perfect environmental conditions (temperature, hygrometry, dust, vibrations, etc.) and knowledge of the consequences of the variations of any of them; – to equip the premises with the necessary energy sources; – to have a device to watch over the surrounding parameters (if it is felt to be necessary). It is advisable to keep the measuring instruments in their original cases and keep them flat (when possible) on an appropriate piece of furniture. It is advisable to store separately the common measuring instrument and the standards of the firm. See Chapter 8 for more details.
4.4.4. Maintenance It is important to assess the life span of each instrument, even though it is very difficult, if not impossible, in some cases to do so. Replacement should be prepared beforehand so that the services that use the instrument may be as little inconvenienced as possible. To assist you in this task, take heed of the manufacturer’s advice, of the calibration results, and of the identification sheets. It is advisable to use the method of the control charts (see Chapter 6, section 6.2) in order to keep track of the variations of the equipment over time. There are some elements of the measuring equipment which you know will wear out: in particular, the batteries, springs, belts, etc. Spares should be kept handy to make the immobilization time as brief as possible while any of these elements is being replaced. As measuring instruments can be downgraded, it is necessary to define that accuracy limit which can be tolerated, as well as the location where lower-class equipment can be sent and used of for less accurate tasks. Regarding those instruments for which several accuracy classes were provided by the norms when the new instruments are received, the downgrading is done along the classes as they have been defined. Regarding those instruments for which only one accuracy class was defined when they were new instruments, four classes at most will be defined for their use; for example: – class 0 (wear out limit = tolerance as new); – class 1 (wear out limit = 150% of tolerances as new);
Handling of a Bank of Measuring Instruments
123
– class 2 (wear out limit = 150% of tolerances of class 1); – class 3 (wear out limit = 150% of tolerances of class 2).
4.5. Follow-up of the measuring instruments over time Keep in mind that a measuring instrument cannot go off limits, except for two reasons: – “natural” drift (whether it is used or not); – accident. It is therefore of paramount importance to ensure that the personnel are fully aware of the precautions to be taken and the necessity to report any accident (fall, overload, etc.). Following the drift in time will make it possible for the users to avoid facing the very embarrassing situation of the measuring instrument being declared “off limits” at the end of its periodical verification, and avoid the question: “what am I doing with the measurements taken with this equipment since its previous verification?”
4.5.1. Periodicity of the follow-up The systematic and periodic comparison of measuring instruments to metrological references is meant to prevent, for as long as possible, the risk a measuring instrument yielding wrong results. It is impossible to say that a lapse of time would be sufficiently brief to eliminate the risk of a measuring instrument becoming faulty before the end of the period. In addition, too high a calibration frequency is costly for the following reasons: – the process is never free of charge; – there is a drop in the production of measurements when the instrument is immobilized; – there may not be a substitute instrument. However, too long intervals may make it impossible to detect a drift of the metrological qualities of the measuring instrument early enough. Therefore, a compromise is necessary, though it is difficult to draw up a list of universally applicable validation intervals. You should bear in mind two fundamental and opposing criteria which have to be balanced when you set the follow-up intervals.
124
Metrology in Industry
They are: – to make the risk of the measuring instrument straying out of the tolerances while being used as small as possible; – to make the costs of verification or calibration as cheap as possible. The calibration frequency does not have to be constant. The time intervals between two verifications or calibrations can be adjusted. They will be reduced when the results of the previous comparisons do not allow you to permanently guarantee the accuracy of the means of measurement. They will be increased if the previous comparisons show that longer intervals do not impair the reliability or the accuracy of the means of measurement. Sporadic checks to detect any malfunction should be ruled out within these periods. For some instruments, you can use surveillance standards to check the condition of the measuring instrument, as a quick verification, before each use. If this operation is done in a strict and well-documented way, it can replace the scheduled verifications, or make it possible to adjust the periodicity of the verifications. If some means of measurement are used only now and then, or exclusively used for one or only a few functions, a specific method of verification could be used for these means. These means must then be identified so as to avoid any risk of error. The periodicities may be granted a tolerance to give the quality system some flexibility; for example, the follow-up periodicity can vary from 1 year to a month, or from 6 months to 15 days. See Chapter 5 about this issue. 4.5.2. Campaign of recall It is of paramount importance to tack the instruments down over time. Once the periodicities have been settled, all that is left to do is to proceed to the recalling of the means of measurement. It is important to make clear who is responsible for the follow-up, whether the material is to be collected, whether it is to be brought in by the users, whether it is to be checked on site, etc. Knowing a instrument is going to be out of use for a time, you must avoid hampering the production line that might need it. It is strongly recommended that: – the recalls be planned; – the users be forewarned; – some replacement material be provided.
Handling of a Bank of Measuring Instruments
125
By this way, the inconvenience caused to the users is reduced, and the application of the scheduled plans is facilitated. A planning-board with “T-shaped” cards could be used to follow the shifts of the measuring instruments, if there are not too many. For example, you only display on a board the work to be done over the next three months; you can move the cards back and forth and remove them when the work is done; the instruments which have been sent back can be easily identified by using different colors of cards. Hence, at one glance, you can see the progress of the follow-up. Also, the users of the instruments are easily informed if there is easy access to the board.
4.5.3. Follow-up of the results It is important to periodically analyze the results of the follow-up of the measuring instruments. You must not keep an eye on them only to meet the requirements of the ISO 9000 norm or those of the client. The aim is to detect a possible drift of the measuring instruments and to make use of the results to reduce the uncertainties related to the measurements. The results of the calibration are used. Also, you can use graphs of the results, which make it possible to detect a drift before it occurs and to react before it is too late. This method is also useful when you decide on monitoring intervals. See Chapter 6 for more information. You must not forget to follow the handling of the bank. It must periodically be subjected to audits to ensure that the procedures are followed, that the system is developed, and that research is being undertaken to improve it.
4.6. Software for the handling of the means of measurements As all companies are becoming increasingly computerized, and as the market offers various software for the handling of measuring instruments, it is quite tempting to obtain such software. However, you have to be very careful before deciding to purchase software because it may not necessarily meet the needs of metrologists and it is not easy to offset its cost. No particular software will be mentioned here; all we wish to do is introduce a few points of reference. In electric metrology there is software for the handling of multimeters which is almost unanimously approved in the profession (but it is not suitable for other technical fields); you can obtain its details from nearly all the big electricity laboratories. Other software has a “users’ club”.
126
Metrology in Industry
Regarding software, you first have to be sure that it is economically profitable. Of course, pages in a binder do not urge to work with excitement during an audit: but a computer does not do everything, there is a danger of depending a bit too much on computers, and there are limits on the software. The main goals of computerized management are to have easy access to all the data in the files, to make the updating of the documents easier, to prevent the contents of the data being tampered with by anybody, to have handy all the information about the measuring instrument and, finally, not to leave any measuring instruments out of the periodical follow-up. Work has been done by some French associations to help potential buyers (or architects) of software handling measuring material. For example, the French club, Métrologie Centre, has assessed and compared about 30 types of software. A few years ago, the French metrology group FAQ Ouest (Federation of the Quality Associations of the West) established an assessment grid along the following principles: – the main criteria were listed with a rating; – the criteria rated 3 were deemed indispensable when choosing software; – the criteria rated 2 were more specific to the utilization of the software; – the criteria rated 1 were a plus. Handling software is nothing but a tool. If it has been badly designed or if it is badly used, it soon becomes a source of problems and then an unwelcome cost for the firm. However, if it is adapted to the real needs of the firm and is in the hands of the person in charge of the metrology function, it becomes a source of productivity.
Chapter 5
Traceability to National Standards
5.1. Introduction In many fields or activities, the requirements of applicable written standards or contracts require that the measurements performed by the instruments be traceable in relation to the national standards. The same demand applies when you want to be sure of the quality of the measurements performed by a measuring instrument. The purpose of this chapter is to provide the main theories that are necessary to achieve this goal in an organization or company that is faced with this requirement.
5.2. Definitions Traceability, in its technical as well as documentary meaning, should not be dissociated from the technical operations which are related to it: calibration and verification; this is the reason why the definitions of these terms should be known and remembered in order that they might be unambiguously used.
5.2.1. Traceability It is the term that you must base your work on to comply with the demands relative to the traceability to national standards as they appear, mainly in the written standard ISO 9001:
Chapter written by Luc ERARD – Laboratoire National de Métrologie et d’Essais (LNE), and Patrick REPOSEUR – Comité Français d’Accréditation (COFRAC).
128
Metrology in Industry
– 2000 or ISO/TS 16949 for firms, ISO/IEC 17025 for laboratories (or firms when they undertake analyzing, calibrating or testing activities); – EN 45004 for inspecting activities; – EN 45011 (ISO Guide 65) for product certification. A first definition appears in “International vocabulary of basic and general terms in metrology” (VIM), 1993 (section 6.10): “Property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons all having stated uncertainties.” Note 1: the concept is often expressed by the term traceable. Note 2: the unbroken chain of comparisons is called a traceability chain. Note 3: the way the connection to the standards is effected is called connection to the standards. A second definition appears in the written standard ISO 9000: Essential principles and vocabulary, 2000 (section 3.5.4), which defines traceability as the ability to retrieve the history, the implementation or the location of what has been examined. Note: in metrology, the definition in paragraph 6.10 of VIM 1993 is the accepted definition.
5.2.2. Calibration “International vocabulary of basic and general terms in metrology” (VIM), 1993 (section 6.11) defines calibration as: – a set of operations which establish, under specified conditions, the relationship between values of quantities indicated by a measuring instrument or measuring system; or – values represented by a material measure or a reference material, and the corresponding values realized by standards. Note 1: the result of a calibration makes possible either the assignment of values of measurands to the indications or the determination of corrections with respect to indications.
Traceability to National Standards
129
Note 2: a calibration may also determine other metrological properties, such as the effect of influence quantities. Note 3: the result of a calibration may be recorded in a document, sometimes called a calibration certificate or a calibration report.
5.2.3. Verification ANSI/NCSL (1) – standard for calibration – Z540: 1994 section 3.28 defines verification as an evidence, from calibrations, that the specified requirements have been satisfied. Notes 1 and 2: within the context of the handling of a bank of measuring instruments, a verification makes it possible to ensure that the deviations between the values indicated by a measuring instrument and the corresponding known values of a measured quantity are all below the maximum permissible errors, such as they are defined by a norm, some regulation or a requirement specific to the person in charge of the bank of measuring instruments. The result of a verification entails a decision to put the instrument back into service, adjust it, repair it, downgrade it or scrap it. In all cases, a written record of the verification has to be kept in the individual file of the measuring instrument. This written standard should be withdrawn when the ISO 10012 standard about verification comes out.
5.3. Traceability chains At the international level, the decisions concerning the International System of Units (SI) and the recommendations concerning the realization of primary standards are taken into account by the Conférence générale des Poids et Mesures (CGPM). The Bureau International des Poids et Mesures (BIPM) is responsible for coordinating and maintaining the primary standards and for organizing comparisons at the highest level. As the basic principle of traceability consists of linking the measurement “in its most general sense” to relevant standards, most industrialized countries have set up traceability chains which fulfill this function, at least in relation to the most accurate measurements, the instruments which are regarded as reference standards, or those which contribute to the guarantee of the quality of a product or of a test.
130
Metrology in Industry
These traceability chains rely, at the highest level, on one or several national metrology institutes whose principal missions are to realize, improve and maintain the national references. Theoretically, these are directly defined in relation to the SI. The realizations of the national references can, for some quantities, be implemented in associated laboratories which are delegated for this activity by the national organization in charge of metrology (the CETIAT (Centre Technique des Industries Aérauliques et Thermiques) for hygrometry in France, the National Engineering Laboratory (NEL) for flow in the UK, etc.). The national metrology institutes and the associated laboratories are liable, in their calibration services, for the measurement units they provide to users who may be scientists, research laboratories or industries. It is also their duty to make sure that their realizations are coherent at the international level; this coherence is obtained through the participation of the National Metrology Institutes (NMI) in comparisons organized by the consultative committees of the Comité International des Poids et Mesure (CIPM), or by regional metrology organizations such as European collaboration on measurement standards (EUROMET), Asia-Pacific Metrology Program (APMP), Sistema interamericano de metrologia (SIM), etc. More and more often, so as to secure the quality of the calibrations, these laboratories are requested to become accredited, or to set up a quality system for their calibration activities in accordance with the requirements of the written standard ISO/IEC 17025. In their calibration services, the national metrology institutes and the associated laboratories directly provide traceability to the references of the accredited calibration laboratories (frequently identified as SMH (Service de Métrologie Habilité) in France) and provide the organization which accredits the calibration laboratories with their technical competence and their support. They have other activities which include: calibration, both internally and for third parties; training personnel; and technical assistance, especially for setting up calibration laboratories. In Europe, the accredited organizations (national metrology institutes (NMIs) and SMH calibration laboratories) comply with the requirements of the ISO/IEC 17025 written standard and the specific documents of the accreditation organizations. They can issue calibration documents referring to their accreditation body, for example: Comité Français d’Accréditation (COFRAC), Deutscher Kalibrierdienst (DKD), Service d’Accréditation Suisse (SAS). Since 1984, the NMIs and the national accreditation bodies (NAB) have been collaborating in order to allow the free movement of calibration documents. The process of securing the traceability to the SI system is made simpler for industries
Traceability to National Standards
131
by this recognition of equivalence; the document which is issued to an industry only has to bear the seal of either the national NAB or the NMI. Based on interlaboratory comparisons, the arrangements of the Bureau International des Poids et Mesures (MRA-BIPM) and the European organization of cooperation for accreditation (MLA-EA) are accessible on the internet: – MRA-BIPM: www.bipm.org – MLA-EA: www.european-accreditation.org Each seal that has been recognized as equivalent can be consulted and any additional information can be obtained from the national accreditation body (COFRAC in France). The presence of this “symbol” proves the accreditation and the recognition of equivalence.
5.4. Traceability It is clear from the definitions that the traceability of measurements, which is a basic requirement of many written standards dealing with quality assurance, is of a technical nature on the one hand and of a documentary nature on the other. Technical traceability is always secured by a connection through an unbroken chain: – to national or international standards in relation to physical measurements, then, as a last possibility, to the basic quantities of the SI; – to basic constants, perfectly referenced and with documented procedures, or to reference materials which are well known in the field involved. Documentary traceability is generally ensured by complying with the requirements of the quality assurance written standards such as the ISO/IEC 17025 written standard (see section 1.6 of this standard). The traceability has to be secured when the firm cannot technically show that the absence of traceability does not have any influence on the result of the measurements, or on their associated uncertainty (see ISO/IEC 17025, section 5.4.6) The needs of the firm and the causes of uncertainties of measurement will make it possible to determine the consequences of the absence of traceability.
132
Metrology in Industry
5.5. Calibration 5.5.1. Calibration in an accredited laboratory After calibration has taken place in such a laboratory, a calibration certificate is issued by the laboratory’s accreditation body, in its field of accreditation (range of measurement and uncertainty). Calibration accreditation guarantees traceability from a technical as well as from a documentary point of view.
5.5.2. Calibration in a non-accredited laboratory Such a laboratory may issue calibration certificates, but they are not guaranteed by an accreditation body, and they do not mention any certification of a system of quality management in compliance with the requirement of the IAF (International Accreditation Forum, which, at the international level, includes the certification organizations, their accreditation body and the principals) no. G.3.5.7, which prohibits these logos from appearing on anything that can be related to a product or a result. The traceability will not be secured unless the following conditions are met: – the technical traceability must be justified by traceability of the laboratory’s reference standards to the national standards, or equivalent, and by appropriate calibration procedures completed by calculations of the uncertainties; – the documentary traceability must be justified by compliance with the requirements concerning quality assurance. The user, from the laboratory or from outside, should make sure the service of the laboratory is in conformity with the different requirements and relevant, thorough, appropriate assessments, not limited to the audit of the system of quality management based on the service company’s ability to perform the measurements requested by the user. Note: whatever the nature of the laboratory which has delivered a calibration certificate, it is important to point out that it is not sufficient to look only at the flyleaf (or the label stuck on the instrument); it is necessary to ensure that the calibration program is relevant and sufficient for what is expected from the instrument. A calibrator or a multimeter may be calibrated for one function and one range, but the traceability will be secured only for that range if the calibration program includes a sufficient number of measurement points.
Traceability to National Standards
133
5.6. Verification 5.6.1. Verification in an accredited laboratory and in its accreditation scope The verification operations carried out in such a laboratory, and within its accreditation scope, entail the issuing of verification reports in conformity with the requirements in effect, and referring to the national accreditation body. The requirements of the national accreditation body (COFRAC, DKD and UKAS, for example) guarantee the traceability from a technical, as well as documentary, point of view.
5.6.2. Verification in a non-accredited laboratory or out of the accreditation scope Such a laboratory may perfectly deliver verification reports. The reports, quite obviously, cannot refer to any guarantee from an accreditation body. The traceability cannot be secured unless the following conditions are met: – the technical traceability must be justified by the traceability of the laboratory’s reference standards to the national standards, or equivalent, and by appropriate verification procedures which include the calculations of the uncertainties of the measurement that have led to the drafting of the report; – the documentary traceability must be justified by compliance with the requirements concerning quality assurance. The user, from the laboratory or from outside, should ensure that the verification reports are relevant and in conformity with the different requirement, by means of audits. Note: as for calibrations, the user should make sure that, at the technical level, the content of the verification report completely fits the use scheduled for the instrument (verification program, uncertainty of measurement, maximum permissible errors, etc.).
5.7. Use of calibration and verification results The measuring instruments which have been subjected to a calibration or verification may, under certain conditions, be used as references for the calibration or the verification of other measuring instruments. The main questions that may be raised on this subject are addressed in the following sections.
134
Metrology in Industry
5.7.1. Use of the results of a calibration The calibration certificate, as defined in the VIM, theoretically contains all the technical elements that enable the beneficiary instrument to be one of the technical links of the traceability chain: – “relationship between the values of the indicated quantity and the corresponding values of the quantity realized by the standards”; – uncertainty of measurement. As a result, the calibration certificate can be used as the starting point of or the reference for a new calibration or a verification in the field for which was been issued; the uncertainty used as a base for the calculation of the uncertainty is the one which appears in the certificate. This point is particularly important because a “calibration document” in which no indication of uncertainty appears cannot be used for the propagation of uncertainties or for ensuring the “technical” traceability of any instrument. The calibration certificate of the calibrated instrument is one of the links in the traceability chain in the field for which the calibration certificate has been issued.
5.7.2. Use of the results of a verification Theoretically, a verification report only contains a judgment about whether the instrument does or does not meet the requirements of the specification (permissible error limits), that is, the information about whether it is apt to do what it is intended to do. The numerical values of the measurement results and the combined uncertainties do not, as a rule, appear in the verification report, but, of course, they are recorded in a file or on measurement sheets. However, the instrument and its associated verification report may no longer be used as a new starting point for traceability to standards. However, if a verification report contains the numerical values of the measurements, as in a calibration certificate, plus the combined uncertainties, it can be used afterwards to ensure the “technical” traceability of any instrument. In this case, the uncertainties mentioned in the report are those used as a base for the propagation of the uncertainties. In the particular case of a verification report issued for an instrument, or a standard, defined by its nominal value, the numerical value being within the limits of permissible errors, or class, the corresponding instrument can also be used to ensure traceability (for example, the standard gauges used to verify calipers).
Traceability to National Standards
135
Likewise, if a verification report contains only the numerical values of the measurements, as in a calibration certificate, but nothing about the uncertainties of measurement, it can also be used as a new starting point for traceability to standards. In such a case – although the principle may be questioned despite being sound on a strictly technical point of view – the values of the uncertainties used as a base for the propagation of the uncertainties are simply the permissible error limits. The verification report, in its usual form and except in the special cases mentioned above, together with the associated instrument, cannot be regarded as one of the links of a traceability chain, but as the end of it.
5.8. Particular cases It may be necessary, in some domains or for some particular instruments, to be more specific or to give examples of traceability to national standards.
5.8.1. “Self-calibrating” or “self-gauging” measuring instruments The new multimeters, or calibrators of electrical quantities, can be mentioned in this category of instruments. The manufacturers of these instruments recommend that they be calibrated with the help of 2 or 3 reference standards, such as a Zener diode reference and two resistors of 1 and 10 k , for example. Using calibrated reference standards is not sufficient, in theory, to assume that traceability has been reached. As a rule, the internal working of the instrument and the processing of the data provided by the standards are not known (from the point of view of the corrections as well as of the uncertainties associated with the measurement) and so it is impossible to be sure that the different domains and ranges of the instrument have been correctly linked to the SI. Ensuring the traceability of these instruments can only be achieved by implementing the classical methods: calibration or verification with the aid of standards that are themselves directly, or indirectly, traceable to the national standards, or their equivalent.
136
Metrology in Industry
5.8.2. Complex instruments in which components/equipments and software are narrowly combined and large measurement ranges are covered for complex quantities The following instruments are included in this category: – the vectorial network analyzer, which measures the components in modulus and phase of high frequency electrical quantities; it is calibrated from a limited number of reference standards; – the coordinate measuring machine, which measures the dimensional quantities of complexly-shaped parts; it is also calibrated from a limited number of reference standards. These instruments work on identical principles; integrated software makes it possible to compensate some systematic errors, to extend the measuring range, to make complex calculations and to reduce the number of random errors, all at once. The problem of the traceability of these instruments is not completely resolved; however, it is possible to suggest a few ways to solve this problem: – a large enough number of standards measured by these instruments are to be used to determine what errors are related to the measurements made in the whole range of operation of these instruments; – there should be software to assess the measurement uncertainties; this software should be validated.
5.9. Metrology in chemistry and physical methods of chemical analysis Whether the concern is the ISO 9001 written standards, or its specific requirements for a particular industry (automotive, aeronautical, etc.), a particular activity (ISO 14001), or the ISO/IEC 17025 or ISO/IEC 17020 (EN 45004) written standards, there is a requirement for the person in charge of the metrology function and responsible for the bank of measuring instruments. It is: “The measuring equipment must be traceable to national or international standards.” To say that this requirement for traceability can be applied only in the domains of science in which it is possible to materialize a basic quantity of the SI, or a derived quantity, is to summarize things all too briefly. There is nothing in this requirement that makes it possible to differentiate between a “physicist”, a “chemist” or a “biologist”. They all have to be able to prove that the measurements made are coherent, and that the measurement should be independent of the measurement equipment and the method used.
Traceability to National Standards
137
The objective of the traceability to national standards is to ensure that a measurement result obtained somewhere in the world is clearly comparable to another measurement result obtained in similar conditions in another part of the world. In the case of physical methods of chemical analysis, the notion of traceability to national standards is understandable, though less clear-cut than for physical measurements. A pragmatic approach has been taken in a conference by Mr Alain Marschal entitled “Traceability and calibration in analytical chemistry” (National testing laboratory, LNE). In all cases, the metrology function should be able to ensure the coherence of the measurement results, for example by taking part in national or international interlaboratory comparisons, so as to optimize its method of analysis or by verifying this coherence by using another method of measurement. According to the international vocabulary of basic and general terms in metrology (VIM), the traceability must be implemented through an unbroken chain of comparisons. This vocabulary was published by the BIPM, the OIML (Organisation Internationale de Métrologie Légale), the IEC/CEI (International Electro Technical Commission), the ISO (International Organization for Standardization), the IUPAC (International Union of Pure and Applied Chemistry), the IUPAP (International Union of Pure and Applied Physics) and the FICC (Fédération Internationale de Chimie Clinique).
5.9.1. Traceabilty in metrology in chemistry No matter what document of reference a firm has chosen, problems of traceability have always been considered by standardization bodies because, they thought, the traceability was not technically feasible, or perhaps there was no standard at the national level. This obviously concerns all the domains of chemical analysis, but also all the measurements to characterize a physical property of a material (bending by shock, hardness, etc.). In the case of physical methods of chemical analysis (chromatography, spectrometry, etc.), the operations prior to an analysis usually include an operation which is said to be a calibration or a gauging; it makes use of solutions prepared by the laboratory, or of reference materials supplied by producers who may be accredited ((IRMM-JRC) Brussels, (NIST) USA, (LGC) UK, (EMPA) Switzerland). Gauging the measuring equipment of a method of chemical analysis means adjusting the output signal, or using a standard curve in such a way that for each
138
Metrology in Industry
level of concentration, the mean of the results coincides with the conventional true value (CTV) which is given by reference samples, or by a reference method (absolute). Gauging errors, by one method or another, causes systematic deviations which can be constant or proportional to the input signal and thus dependent on the concentrations. The notion of gauging error, with a modification of the position of the experimental line in comparison with an ideal line, and also of the shape of the cloud of experimental points, must not be taken into account unless the purpose of the comparison is to “adjust” an alternate method in relation to a method that is taken as a reference. The problems encountered in what could generically be called chemical analysis come from the fact that the “standard” product: – is not a reference material which is certified (or whose traceability is completely established), so no evidence such as an accredited calibration certificate is recorded; – is not sufficiently resistant to the effects of time for instance, as it is subjected to oxidization or reduction reactions, or polluted by gases such as the CO2 from air; – responds in a way differently from the analytes in the real sample owing to matrix effects; – cannot always have its characteristics verified by the buyer, even if it is purchased from a specialized producer. On the other hand, a case which ought to be considered is that of a piece of equipment that is very difficult to calibrate or verify and it is practically impossible to obtain traceability to the national standards according to the usual protocol; for example, calorimetric methods, a thermo gravimetric (TGA) or differential thermo analysis (TDA). As in most methods in which the object is to physically characterize a material, a drawback of chemical analyses applied to liquids and to solids is that they usually destroy or modify the sample by turning it into a solution, by extraction, etc. This inconvenience makes it very hard to maintain an unbroken real chain of comparison to a national-level standard. Faced with this situation, many “chemist-analysts” consider that they have only one means at their disposal to validate their method and verify their measurement results: to repeat the analyses on samples characterized by known values which have
Traceability to National Standards
139
been established by a process deemed to be reliable, that is to say, to use a certified reference material. These samples are selected or prepared by laboratories which are well-known or recognized to be competent because of their experience or because of the results of a campaign of interlaboratory analyses for which the purpose was, for example, to precisely define the concentrations of some elements or components in a specified matrix. However, in many cases, the samples have to be conceived and prepared in and by the laboratory in accordance with the requested analysis. Some laboratories have become accredited for performing these calibrations since the first edition of this book was published. Their accredited possibilities are accessible on the internet sites of the European Cooperation for Accreditation (EA) (access through www.cofrac.fr).
5.9.2. Influence of the principle of the method The influence of the type of the method is not insignificant. It can be classified in three categories according to the principle of calibration which is used. 5.9.2.1. Absolute methods The principle of the method consists of obtaining the result of the analysis from laws which link physical or chemical phenomena. The measurements consist of, for example: – weighing the quantity of a substance; – determining a volume of titration reagent; – weighing a mass of precipitate; – determining a volume of generated titration. The method for the traceability chain consists of separately identifying the elementary quantities which have been measured in the analysis process and linking them to national standards. Should these properties have a significant effect on the results of an analysis, then the requirements defined in the ISO/IEC 17025 written standard in sections 5.4 and 5.6 must be complied with. In practice, it is rather difficult, or too expensive, to control all the parameters used for a calculation, or it is the value of the uncertainty which is dissuasive. It is vital, in all cases, to make an assessment of the causes of uncertainty in order to be able to identify the most influential ones and, possibly, intervene in them to reduce their effects.
140
Metrology in Industry
This assessment can then be used as a tool of the functional analysis of the measurement process. In their concern to help industrials and laboratories as a whole, EURACHEM (European Cooperation for Chemical Analysis) and the CITAC (International Committee for the Traceability in Chemical Analyses) have published a document which is a guide to the assessment of uncertainties of measurement. 5.9.2.2. Relative method The principle of the method consists of comparing the indications given by the instrument for the measurement of the sample with those given for the calibration performed from a range of “reference” products prepared by dilution of the pure analyte in a solvent, the concentrations of which are known by the user. The method of connection consists of connecting the different systems of measurement used for the preparation of the “standards” (mass, volume, temperature, purity of the basic products, nature of the impurities, interpolation between two points, etc.). However, one should always keep in mind that the objective is to satisfy an industrial need and therefore one should estimate the share contributed by each one of the causes of uncertainty and then compare their total sum to the final uncertainty of the result of the analysis. 5.9.2.3. Comparative method The principle of the method consists of comparing the indications given by the instrument for the measurement of the sample with those obtained from a “calibration” curve drawn from samples which are known to be of the same nature and taken as references. The method of connection consists of using reference materials, preferably certified material if there is any, and the nature of which is very close to that of the sample to be analyzed, and without any additional disruptive effects (influence of the matrix, content of the measured-out element, geometry of the standard, etc.). When no reference materials are available on the market, another method consists of using samples prepared from the pure analyte and some blanks, that is to say some samples of the same type which are supposed to not contain any trace of the analyte. This method does not give the same guarantees as the methods which depend on external reference materials. Another technique consists of comparing the results of the sample with those of a reference method from the first two categories, for example, measuring out fat in milk by infrared spectrometry compared with an ether-hydrochloric extraction.
Traceability to National Standards
141
5.9.3. “Documentary” traceability Strictly speaking, as this term is not defined in any published text of terminology, we use it in opposition to the term traceability chain, which is wellknown in metrology; therefore, it corresponds with the general meaning of traceability as defined in the ISO 8 402 written standard, note (b) excepted. In the case of physical methods of chemical analysis, the question is not so much to find the track of a particular document; it is to be able to prove that the techniques used for “adjusting” the method make it possible to have confidence in the measurement result and the uncertainty which goes with it. On the other hand, the question arises: “what is to be connected and how do you prove the connection?”. For example, in France, is the verification of a stopwatch used to determine a time interval traceable to a national standard if you use a method describing the verification process or if you use a “standard telephone” directly linked to the speaking clock? The answer is “yes, it is”. The traceability to a national standard is valid; the accuracy of the means of reference is in the order of 0.1 milliseconds in France and the timekeeping of this clock is controlled from an atomic clock connected to the national standards: in France, the Laboratoire Système de Références Espace Temps (SYRTE), in Germany, the Physikalish Technische Bundesanstalt (PTB), and in Italy, the Instituto Electro Nazionale Galiléo Ferraris (IENGF). The proof of this connection can only be internal as there is no delivery of a calibration certificate issued by a laboratory, which has an accreditation by COFRAC, or by a member who has signed the equivalent recognition agreement of the EA. The error made at the time of the setting of the “chronometer” is much larger than the uncertainty of the connection, but quite suitable for the use to be made of the measuring equipment. Should all the gauged glassware be verified and how often should this be done? This equipment may represent more than 80% of a “chemistry” laboratory’s bank of measuring equipment. Therefore, you need to be cautious about this demand. It is possible to verify one volume per weighing: 1. However, what uncertainty can you guarantee when you weigh a volume of water from a micropipette of 10 µl? 2. Is this uncertainty comparable to the maximum errors allowed for this order of pipettes? Surveys of new or “precision” material have been undertaken and are currently being continued in some laboratories; they conclude, for the moment, that about
142
Metrology in Industry
80% of the verified glassware was within the error limits allowed by their class or their requirements. It is then possible, depending on the various cases, to trust the values of the permissible maximum errors and use them in the evaluation of the overall uncertainty of measurement. Once again, it is the analysis of the need and the calculation of the uncertainties which tell you whether the method of connection is relevant and whether it is reasonable to invest in these verifications. The decision to calibrate a spectrophotometer will depend on the type of analysis made with this spectrophotometer. Instruments such as spectrometers and chromatographs which have to be calibrated every time they are used should be calibrated with chemicals known to be sufficiently pure, or reference materials whose composition is known, knowing that deviations of internal repeatability can reach from 3 to 8%. The laboratories of chemical analysis frequently use this chromatography equipment, either in a gaseous phase, or in a liquid state. To sum up, the principle of this method consists of an elution of the elements constituting a sample; the detection at a time T is depending on each one of the constituent elements. This technique seeks to make use of reference products so as to be able to identify the constituent elements. The reference materials are obtained by the user through successive dilutions (mass and volume). The user, before turning his attention to the connections of these two quantities, has to determine the contribution to the overall uncertainty of the injection system, the geometry of the column, the analysis temperature, the response of the detector and the response of the integrator. Spectrometry techniques are commonly used in the laboratories which practice the determination of chemical elements of a substance. Spectrometry is not an absolute method, it is a relative or comparative method, which requires drawing a calibration curve of the indication from internally diluted solutions or from reference materials when there are some. The result of a measurement is obtained by transferring the determined value on the calibration curve that has been drawn; it is like using an experimental graph. In order to reduce the influence of the successive dilutions, the same parent solution has to be used for each dilution, as long as the sought for uncertainty permits it.
Traceability to National Standards
143
5.9.4. Control of the reference materials When the laboratory uses reference materials (RM) of its own or from outside, a procedure has to be established which makes it possible to check, to use a new sample and to compare the response of the measuring equipment, with the two samples of RM (the older one and the newer one), in order to determine the systematic component of the uncertainty related to the reference. Moreover, the laboratory must be absolutely sure about the homogeneousness of a lot, the sampling conditions, etc. The criteria that rule the decision to renew the RM must be written down. These RMs must meet the previous requirements and be applied to the standards related to the SI; when this cannot be done, the products used as references must be treated like consumable products used as part of the tests or analyses. In addition, the laboratory should have at its disposal a range of RMs adapted to its sector of analysis, if these RMs are available. These procedures concerning the use of reference materials should be described in detail in the documents which are at the disposal of the operators; the observations made should, in all cases, make the traceability of the operations possible, especially when faults are detected. The file relating to the equipment should always contain its follow-up information, especially the follow-up of the monitoring of the coherence of the product which is used to control the drift over of the response of the measuring equipment. The metrological traceability is achieved through reference standards, which belong to and are created by the laboratory, then through the internal traceability chain which implements the metrology function. The chain is broken when the final link is compared to a link of the same nature, that is to say, at the level of the external connection to a national standard kept by the national laboratory of metrology or by an accredited laboratory. The laboratory must be in a position to prove that every calibration of the internal traceability chain has been done according to the set up procedures. The different stages of the manufacturing to be taken into account are the following: – supply and receipt (definition of the expected requirements, check of the products on arrival, etc.); – manufacture of the products when this operation is within the laboratory’s scope; – storing conditions;
144
Metrology in Industry
– handling (preparation, conditioning, etc.); – management (identification, follow-up, inventory, etc.); – assessment of the uncertainty provided or evaluated by the laboratory, in the case of internal reference materials; – reference materials which have or have not been certified and about which the laboratory has to show they are suitable for the use that is made of them. The reference materials and the chemical standards have to be clearly labeled so that they can be unambiguously identified and referenced in relation to the certificates and other documents that go with them, and so that their documentary traceability is secured. The information must be available and mention the duration of preservation, the storing conditions, the applicability, and the restrictions of use. The made-up standards should be treated as the reagents, in relation to labeling. The reference materials provide the essential traceability of chemical measurements; they are used to prove the correctness of the results, to calibrate the material and the methods, to check the performances of the laboratory and to validate the methods. They also make it possible to compare methods when they are used as transfer standards. One is encouraged to use them as much as possible. When there are matrix interferences, using a sample, with a measured-out addition, of a chemical standard is generally acceptable. It is important that the certified reference material (CRM) is produced and characterized in a technically sound way. The users of CRM should be aware that all materials are not validated from the same standard. Details about the tests of homogeneity and stability, the methods used for certification, uncertainties and variations of the stated values are usually obtainable from the producers and this information must be used to assess the quality of the CRM and whether it is appropriate to use it for a given analysis. The required purity of chemical standards can be defined in relation with the tolerances of the method. For example, a tolerance of 0.1% of the targeted value requires the chemical standard to have a precision of concentration significantly better than 99.9%. It is essential to control the impurities for an analysis of traces. Particular attention should be paid to the manufacturer’s advice about the storing and the duration of preservation.
Traceability to National Standards
145
Reference materials and standards should be handled in such a way as to protect them from possible contamination or alteration. Personnel training procedures should reflect these requirements.
5.9.5. Conclusion In the domains of what we have called chemical analysis, as well as in the field of the measurement of the basic quantities of the SI, it is fundamentally important to remain open-minded and to take the whole process of measurement into consideration. Thus, metrology is neither the science of measurements (as defined in the Concise Robert Dictionary) nor the science of uncertainties (Pierre Giacomo – Honorary Director of the BIPM), but “the conscience of the process of measurement”. The measurement process includes additional parameters such as sampling, preparing the sample and relativizing the influence of basic quantities (mass, temperature, volume, etc.), which, admittedly, remain at the base of some methods, particularly in wet process chemistry, and have to be connected, but they must not eclipse other parameters which have a greater influence when assessing the causes of uncertainties and which are not to be dealt with simply by a connection to the basic quantities. The difficulties of guaranteeing the traceability to national standards make a good case for the implementation of a quality system which would evolve as and when corrective actions are applied. Furthermore, the development of “crossed” analyses has to be supported and helped either by: – resorting to two similar methods which, however, are based on different principles; or – promoting campaigns of interlaboratory analyses to verify the result of a measurement, rather than making one method more worthwhile than another. Such campaigns are, and will remain, the safest means to make sure, a posteriori, of the coherence of the measurements.
5.10. Assessment of traceability This is an especially important point because, as procedures to ensure quality (ISO 9004, for example) are currently developing, many firms have to prove that their measuring equipment is connected to national standards or the like. Theoretically, only the assessment of the traceability makes it possible to verify that the corresponding requirements have been met.
146
Metrology in Industry
When the calibration and verification operations have been performed in a metrology laboratory accredited by an organization (for example, COFRAC in France) in its sphere of accreditation, it is not necessary to have an audit done since the traceability, both technical and documentary, is guaranteed by the accreditation. Nevertheless, there is reason to ensure the content of the documents (functions, gauges, programs, uncertainties, etc.) is adapted to the intended use of the instrument. In cases where there is no accreditation, an assessment is necessary to make sure that the technical and documentary traceabilities are satisfied and relevant (calibration procedures, connections to the standards, calculations of the uncertainties, standards used, etc.). The assessment will have to be gone through by, in particular, those firms which subcontract the calibration or the verification of practically the entirety of their bank of measuring instruments. The subcontractor will need to prove that the operations of calibration and verification that he performs are traceable to the SI, by using the document ILAC P 10, for example. In any case, there will come a time when the connection to national standards can only be proved by showing a calibration certificate delivered by an accredited laboratory. As for the principal, he must demonstrate that the services of calibration and verification he has ordered from the subcontractor are relevant. During an audit, the best way to ensure that the stipulated requirements have been met is to rely on the technical and documentary requirements of the ISO/IEC 17 025 norm about technical and documentary requirements. It should be noted that the auditors use the agreements of international recognition (the MRA CIPM, EA, ILAC, etc.) as evidence of equivalent traceability; similarly, the results of national or international comparisons can be used as bases for traceability.
5.11. Bibliography French norm NF X 07-010, The metrology function in the firm, December 1992 (www.afnor.fr) ANSI/NCSL Z540: 1994, American National Standard for Calibration – Calibration Laboratories and Measuring and Test Equipment – General Requirements (www.ansi.org) Guide ISO 35:1989, Certification of reference materials – General principles of statistics (www.iso.org)
Traceability to National Standards
147
Guide ISO 34:2000, Quality system requirement for the production of reference materials. ISO-CASCO EURACHEM-WELAC, Accreditation of CHEMISTRY laboratories: Guide for the no. 1/WGD2 interpretation of the norms of the series EN 45 000; and the Guide ISO/CEI no. 25RNE, February 1993 (www.eurachem.ul.pt, www.european-accreditation.org) EURACHEM-CITAC, Quantifying Uncertainty in Analytical Measurement, 2nd edition, 2000 (www.eurachem.org) ILAC, Calibration of chemical analyses and use of certified reference materials, May 1993/Draft ISO guide 32 (www.ilac.org) ILAC, Guide for calibration and maintenance of measuring test equipment in laboratories, February 1994 (www.ilac.org) ILAC-G17: 2002, Introducing the Concept of Uncertainty of Measurement in Testing in Association with the Application of the Standard ISO/IEC 17025 (www.ilac.org) ILAC P 10: 2002, EA Policy on Traceability of Measurement Results, December 2001 (www.ilac.org) Metrology Congress, “Approach to the metrology function in laboratories of Lille 1993 (MFQ) chemical analyses” – Christian TRICARD/DGCCRF Talence BCR, Report reference materials – checking the quality of the analyses of agricultural produce, JJ Beliardo – BCR EUROLAB Congress, “Reference material for mechanical testing and uncertainty of measurement”, Malcolm Loveday, Division Material Metrology, NPL, Teddington (www.eurolab.fr, www.lne.fr) EUROLAB Congress, “Traceability and calibration in analytical chemistry and Florence – April 1994 material testing – Principles and applications to real life, in connection with ISO 9 000, EN 45 000 and guide ISO/CEI no. 25”, Alain Marschal, Head of Reference Materials Department, LNE (www.eurolab.fr, www.lne.fr) Engineer’s techniques Calibration in analytical chemistry and testing of Reference materials R 52 Measurements and Controls, Alain Marschal, Head of Reference Materials Department, LNE (www.lne.fr) ISO 9004: 2000 System of Quality Management – Guidelines for the improvement of performances (www.iso.org/iso/fr/iso 9000-14000/tour/magical.html)
This page intentionally left blank
Chapter 6
Calibration Intervals and Methods for Monitoring the Measurement Processes
6.1. Normative requirements Calibrations at fixed (possibly variable) intervals are indispensable processes which are usually expensive for firms. Having these intervals well under control is a major technical and economic objective. The control of the measurement processes resulting from the application of the norm ISO 9001 (2000) is an inducement to ensure that the measurement process “does produce” correct results. This type of demand also applies to testing and calibration laboratories. The ISO 10012 norm, “System of management of the measurement – requirements for the measurement equipments and processes” introduces the following demand (section 8.3.3): If the result of a metrological verification prior to any adjustment or repair indicates that the measuring equipment did not meet the metrological requirement such that the correctness of the measurement results may have been compromised, the equipment user shall determine the potential consequences and take any necessary action. This can involve re-examination of product produced using measurements taken with the nonconforming measuring equipment.
Chapter written by Patrizia TAVELLA – IENGF/Italy, and Marc PRIEL – Laboratoire National de Métrologie et d’Essais (LNE).
150
Metrology in Industry
These measures can have significant technical and financial implications for a firm or a laboratory. For firms there are two immediate consequences of this requirement: – the need to have the intervals of calibration of the instruments under control; – the need to set up methods of monitoring the measuring instruments. The determination of the calibration intervals and their modification, plus the setting up of the methods of monitoring, make it possible to minimize the risk, or at least to control it. A selection of the instruments to be monitored will have to be made when setting up the monitoring methods. Those instruments that are especially critical from an economic point of view or for security reasons should be examined first. The ISO 10012 international written standard requires the organization to specify which measurement processes should comply with the measures stated in this international standard. It is advisable to take into account the risks and consequences of not satisfying the metrological requirements when the limits within which a standard has to be complied with have been defined.
6.2. Methods for monitoring the instruments in use – general criteria These methods should satisfy some criteria in order to work efficiently and be applicable when the instruments in use are monitored: – ease of implementation: in many cases, these methods have to be implemented by the instrument operators; – speed of execution: the time needed to implement these methods must be short, so that the monitoring can be done frequently; – use of the results: the results should be easily understood and provide the operators with information; – motivation of the operators: the operators must be interested in the methods and motivated to use them.
6.2.1. First method: metrological redundancies Principle This method consists of deliberately duplicating some critical elements of the firm’s metrological system so as to easily compare information that should normally be in agreement. Any deviation makes it possible to easily detect a fault.
Calibration Intervals and Methods
151
Hypothesis The method rests on the assumption that the probability of a similar and simultaneous drift in two instruments is low. This assumption may lead to choosing measuring instruments which are technologically different or from different manufacturers. Applications This method is implemented, in particular, in relation reference standards: standard rings, standard masses, etc. The reference standard represents the first link of the calibration chain inside the firm; if it happens to drift, this may entail serious errors of measurement and, most importantly, these are undetectable if no duplicating item is available in the firm. Let us mention, by way of example, reference-standard rings whose 80 mm diameter has altered by 2 µm in a year; this alteration is to be compared to the uncertainty on the known diameter of the ring which was +/- 0.17 µm. The cause of the alteration was probably a defect in the stabilization of the material.
6.2.2. Second method: checking the coherence of the results Principle This method is based on the examination of the measurement results and the calculation of characteristic values such as the standard deviation of repeatability, or on drafting graphs, and comparing them to typical values or to standard graphs. Hypothesis This method assumes that instruments of the same nature, or the same technology, perform more or less similarly. Take, for example, platinum-resistance temperature sensors; the model of the variation of the resistance as a function of the temperature is shown by a second degree polynomial: R = Ro (1 + at + bt 2) It is widely accepted that the value Ro can vary between two thermometers, but the general aspect of the curve remains parabolic, and a discontinuity in the curve will act as a warning. Applications The two following examples illustrate this method: – An electronic comparator made of a table fitted out with two inductive sensors is generally used to calibrate standard gauges. A method can be used to monitor
152
Metrology in Industry
these benches; it consists of testing the repeatability of the measurements and to compare them to a “typical” value. The operator is alerted if the values are not similar. Statistical methods can be used as a basis to perfect this type of test (comparison of a variance to a given value). – Some analysis procedures recommend that the measurements be repeated twice and to compare the deviation between these values with a “critical difference” that has been determined by taking the repeatability of the analysis procedure for its basis. The operator will have to examine if the difference between the two results is greater than the critical difference. Graphic techniques also deserve attention. Studying a chart of numerical values does not generally make it possible to detect the abnormal values; graphs, on the other hand, are very rich in information and something unusual (change of gradient, etc.) in a curve very often reveals a faulty measurement. These techniques can be used to verify the calibration of thermocouples by watching out for the regularity of the calibration curves.
6.2.3. Third method: “monitoring standards” and statistical supervision of the measurement processes 6.2.3.1. Statistical control of the measurement processes The measurement process can be considered as a part of the production process For many years, manufacturing companies have shown interest in the monitoring of the means of production. For the first time, in 1924, WA Shewhart explained the principle of control charts. Chiefly adapted to manufacturing processes, they are perfectly suitable for the monitoring of instruments. The measurement process is then considered as a production tool that does not make objects, but which “manufactures” results of measurement. Measurement process concept The measurement process is a set comprising of: – the measuring instrument (means); – the method of measurement and the measurement procedure (method); – the environmental conditions (medium); – the standards used (means); – the operators (manpower). The measurement process provides the results of the measurements.
Calibration Intervals and Methods
MEANS
153
METHOD
RESULTS OF MEASUREMENT
MATERIAL
MEDIUM
MANPOWER MANPOWER
Figure 6.1. Measurement process concept
Just as any manufacturing process, even one that is perfectly controlled, cannot turn out identical products, so the measurement process comes with errors of measurement that fluctuate from one result to the next. That is why it is necessary to attempt to monitor and control the measurement processes. Principle The objective of this method is to place and then maintain the process under “statistical monitoring”: the dispersions of the results that are observed are only due to the random fluctuations of the instrument or of the environmental conditions, and sometimes to the operator’s initiatives, but not to attributable causes that can be controlled. From a statistical point of view, it can be said that the samples represented by the series of measurements of the same object are extracted from the same population and so have the “same mean”. In order to monitor the measurement process, the different causes of variability will be examined when the system of control charts is set up. These causes come from: – the medium, and the environment; – the method of measurement and the measurement procedure; – the measured quantity (measurand); – the operator; – the means (instruments, standards, etc.).
154
Metrology in Industry
Check standard Monitoring standards have to be used to implement these techniques. A provisional definition of monitoring standards may be measuring instrument, material measure, or product, whose function is to generate or achieve the value of a quantity in a stable way in time. These standards are used at regular intervals to ensure a statistical control of the measurement processes. Several examples illustrate this concept. A frequent measurement in a laboratory is the measurement of direct current. In order to monitor the digital voltmeters, or the automatic measurement sequences, a voltage reference can be used; it can be introduced on different measurement benches to monitor them (by the connection of the reference tension generator to a channel of a channel scanner). These techniques are applied in the field of dimensional metrology. It is possible to associate one “check standard”, or even several, to each measurement bench. These standards are of the same type as those that are usually measured on the bench; special care should be taken when these standards are stabilized. The value of the quantity which is measured and represented by these monitoring standards must also be representative of the measurements customarily made; several monitoring standards (representative of the field of measurement) may sometimes be necessary to supervise the measurement process. 6.2.3.2. Control charts A control chart is a graph on which a point is made to correspond to each value of a statistic calculated from successive samples (mean, range, standard deviation). The abscissa of each point corresponds to the number of samples and its coordinate is the value of the statistic calculated from these samples. Checking and control limits have been drawn beforehand on the graph; they make it possible to follow the evolution of the measurement process. There are numerous types of control charts, but in relation to the monitoring of the measurement processes, three should be retained: the chart of the mean, the chart of the standard deviation, and the cumulative sum chart. First step: know your process well It is necessary before you compile a control chart to estimate the parameters µ and the characteristic of the distribution of the measurements of the monitoring standard with the aid of the process that you want to check.
Calibration Intervals and Methods
155
These two estimators will be called m and s. Two cases are to be considered to calculate the value of m. Either you know the value of m, thanks to a calibration of the monitoring standard by a method of a higher accuracy, and you use the value of m supplied by the calibration, or the monitoring standard is only supposed to be stable and m should be estimated by performing a number of series of measurements. The series of measurements should be sufficiently representative of the different operating conditions so as to ensure a proper characterization of the distribution. If k series of measurement, each made of n1, n2 , … nk determinations, have been made:
n0 =
k
∑n
h
h =1
xh =
1 nh
nh
∑x
ih
i =1
then m will be calculated by the quantity:
m0 =
1 n0
∑n x h
h
The variance of each one of the samples should be estimated with ν h = nh − 1 degrees of freedom by the expression:
s h2 =
2 1 nh x ih − x h ) ( ∑ nh − 1 i = 1
The variance of the population should be estimated by use of s by combining the different variances:
s 02 =
(n
1
− 1) s12 + ( n2 − 1) s 22 +A+( nk − 1) s k2
(n
1
− 1) + ( n 2 − 1) +A+( nk − 1)
=
ν1s12 + ν2 s 22 +A+ν k s k2 ν1 + ν2 +A+ν k
The control charts of the mean and of the standard deviation: LS = m 0 + 2
s0 n
LC = m 0 + 3
s0 n
Regularly, the monitoring standard has to be measured using the measurement process that you want to control. If, for each series of measurements, the number of repetitions is n, you can draw the limits of control and warning, knowing the
156
Metrology in Industry
estimators m and
s0
,. The mean of the series of the n measurements will be noted
n
on the graph. The values of the warning limits (WL) and control limits (CL) will be the coefficients 2 and 3 respectively appearing in front of the estimator of the standard deviation of the mean. They can be modified in accordance with the risk you are willing to take. Nevertheless, using a whole multiple of the standard deviation is certainly sufficient and more meaningful for metrologists. The warning and control limits for the standard deviation are:
s ≤ s 0 F1− α ( n − 1,ν ) in which s2 is the estimator of the variance-estimator obtained with the considered series of measurements, F1 - α (n-1, ν) is the “fractile” of 1-α order of Fisher’s distribution with n-1 degrees of freedom (ν = ∑νh in the numerator and degrees of freedom in the denominator) and has the accepted values of the risk of first kind (α = 5% for the warning limit and α = 2% for the control limit, for example). The initial phase of the drawing up of the chart is bound to involve progress because, practically all the time, it will be noticed that the process is not “under control” and the attempts to find the attributable causes will be a indicator of obvious progress. It seems that the control charts of the mean and of the standard deviation both deserve attention; they provide complementary information on the way the process works. A variation of the mean reveals a drift either of the instrument or of the environmental conditions, whereas an abnormal increase of the standard deviation indicates that the measurement process is not stable. Leaving the checking limits means a compulsory examination of the measurement process. The cumulative sum chart Such a control chart can turn out to be a good thing in metrology because it makes it possible to detect small drifts. Its principle is to calculate the mean of the series x for each one of the series of measurements, and then to work out a series of cumulated sums:
S1 = x1 - m0 S2 = (x1 - m0) + (x2 - m0) = S1 + (x2 - m0) S3 = S2 + (x3 - m0) St = St-1 + (xt - m0)
Calibration Intervals and Methods
157
If the successive values are all obtained in the region of m, the cumulated sum remains close to zero, but if on the contrary a phenomenon of drift occurs, it is quickly detected. As, in order to study the tendency, the cumulated sum smoothes out the paths, it saves supplementary treatments on the chart of the mean. It is possible to use a mobile mean to “smooth out” the series. Tests on the successive groups of points on the charts These tests can detect the presence of a phenomenon which might be abnormal: – nine successive points on a same side of the mean; – four points among five successive ones higher than 1 σ; – two points among three successive ones higher than 2 σ; – seven points higher and lower than the mean successively; – six increasing or decreasing points successively; –15 points lower than 1 σ. The log book of the measurement process If it is used with an intention to progress, the method of the control charts shall be accompanied with information to understand and explain the “abnormal” points which are bound to appear during the life of the instrument and the process. You cannot, without information, connect the appearance of an abnormal value with an event in the measurement process. 6.2.3.3. Use of the monitoring methods The methods of monitoring make it possible: – to know and control the measurement processes; – to protect oneself against the malfunction of the instruments and, more generally, the measurement processes; – to monitor the environment parameters of the measurement process (influence quantities); – to provide formal evidence that the results of measurement are under control; – to have a particularly efficient tool available to adapt the calibration intervals permanently and thus cut down the firm’s metrology expenses.
158
Metrology in Industry
6.3. The determination of the calibration intervals The importance of establishing appropriate calibration intervals for each instrument is well-recognized in international and European standards. For example, the EN ISO 9001 requires that measurement and testing instruments should be periodically confirmed through calibration. The same concept is extensively reformulated in the ISO 10012.
ISO 10012 section 7.1.2 – Intervals between metrological confirmation The methods used to determine or change the intervals between metrological confirmation shall be described in documented procedures. These intervals shall be reviewed and adjusted when necessary to ensure continuous compliance with the specified metrological requirements.
When a measuring instrument is found to be outside the limits of permissible errors, some corrective actions need to be taken on the production process which was measured by the instrument since the last positive calibration check. Consequently, considerable resources are meant to be paid. In order to reduce such costs, it is fundamental to establish a system that carefully watches the instrument calibrations. Some documents or standards give estimates as to the calibration interval. A very good guide is the NCSL RP – 1 document; other suggestions can be found in Document 10 of the Organisation Internationale de Métrologie Légale. The calibration interval is often determined by observations on a large group of similar instruments and estimations of their “average” behavior. However, in some cases, deciding upon an optimal calibration interval for a given type of instrument may be worthwhile if for instance the item has a particular importance in the firm’s production and quality system. Let us assume that the calibration condition of a particular instrument can be monitored by an observable parameter, s(t), whose possible variation is bounded by predefined limits of permissible errors ± a . Due to the very different causes that affect the calibration requirements, such as environmental, chemical, mechanical, human and electromagnetic fluctuations, it is justified to assume that the calibration condition varies according to random steps, whose accumulated effect degrades the calibration condition until it reaches an assigned threshold of permissible error, after which the instrument is considered “out of calibration”. Figure 6.2 describes an example of the evolution in time of the calibration condition.
Calibration Intervals and Methods
159
s(t)
+ a
0 tim e
-a in itia l a d ju stm e n t
n e w a d ju stm e n t
n e w a d ju stm e n t
Figure 6.2. Example of the evolution in time of the calibration condition with some adjustments
Calibration means the passive observation of the calibration status without any action. Setting the calibration status to zero or to some other conventional value is called adjustment. Nevertheless, once a calibration is performed, the calibration error is kept in due consideration either by a physical adjustment or by a software a posteriori correction of the successive measures. Therefore, we speak of calibration interval independently if it is followed by a physical or software adjustment, meaning the time interval at which the calibration condition of a measuring instrument is measured and its value is taken into account for the subsequent measurements. Let us suppose that the measuring instrument at hand is kept under stochastic control, according to the methods explained in the previous sections, by means of repeated measurements of check standards, whose results are registered on a control chart as in Figure 6.2. Two calibration interval determination policies can be considered. In the first policy, the measuring instrument is always kept under stochastic control and the calibration condition is almost continuously monitored on a control chart. The recalibration is only performed when deemed necessary, for example when the calibration condition exceeds an alert threshold ± m , which is fixed below the limit of permissible errors ± a , as depicted in Figure 6.3.
160
Metrology in Industry
+a +m
0
time
∆t
-m -a
initial adjustment
new adjustment
new adjustment
new adjustment
Figure 6.3. Example of the evolution in time of the calibration condition with alert thresholds
On the other hand, when using the second policy, the monitoring of the calibration condition can only be performed for a certain learning period useful to identify a stochastic model suitable to describe the evolution in time of the calibration condition. Such a learning period can be sufficient to evaluate the risk of using the measuring instrument “out of calibration” when it is used at a certain time after calibration. The optimal calibration interval is then identified by time interval which guarantees that such a risk does not exceed a certain fixed level. In this case, the control chart and check standard are only used for a limited period, after which a certain reasonable rule is deduced and the calibration interval is determined. The use of stochastic processes to model the degradation in time of the calibration condition of a measuring instrument or standard proves to be very effective in estimating the probability that, at a certain time after calibration, the calibration condition exceeds the tolerance threshold. Simple stochastic processes as a random running or a Wiener process can be physically justified by considering that the degradation of the calibration condition can be due to the accumulated effect of minor random variations. These processes have been examined and some of their properties can be expressed by known analytical expressions, for example, the probability that the calibration condition exceeds a threshold level at a certain time after calibration. These analytical expressions are then useful to fix the calibration interval according to a predefined risk level, but they can also be used in the case of the continuous calibration condition monitoring described above. In that case, it can be necessary to estimate the probability that the calibration condition has not
Calibration Intervals and Methods
161
exceeded the alert threshold m at a certain check, but that it has exceeded the limits of permissible errors before the next check, therefore leading to the unpleasant situation of an instrument out of calibration before the adjustment is performed (see last example in Figure 6.3). In addition to the criteria in section 6.2, a cost function can be added by inserting the cost either of the use of an instrument out of calibration, or of repeated calibrations, with the aim to minimize the total cost. For safety, one should choose a brief calibration interval, which means calibrating very frequently, to reduce the risks and the costs of using an instrument out of calibration. On the other hand, the cost of calibrations depends on the operations themselves, either if calibration is performed internally or by an external body, the cost of instrument unavailability during the calibration, plus other costs as standard breakage or their equivalent. Therefore, to reduce calibration costs, one is led to calibrate very seldom, which increases the calibration interval. These two contrasting tendencies can be formulated by a suitable annual cost function, whose minimization leads to the identification of the optimal cost saving calibration interval.
6.4. Bibliography NF EN ISO 9001: Quality management systems – requirement (2000) ISO 10012, Measurement management system – requirement for measurement processes and measuring equipment (2002) ISO 7870, Control charts – general guide and introduction (1993) ISO 8258, Shewhart control charts (1991/Cor1 1993) AFNOR, FD X 06-030, Application of statistics – guide for the setting up of the statistical control of processes (1992) AFNOR, NFX 06-031, Application of statistics – control charts Parts 0 to 4 (1995) AFNOR FD X 07-014, Intervals of metrological confirmation (not yet published) OIML (International Organization of Legal Metrology): “Advice for the determination of the intervals of recalibration of the measuring equipments used in testing laboratories”, International Document no. 10, (1984) NCSL, “Establishment and adjustment of calibration intervals”, Tech. Rep. National Conference of Standard Laboratories RP1 (1996) R. Kacker, NF Zhang, C Hagwood, "Real-time control of a measurement process”, 33 Metrologia, pp. 433-445 (1996)
162
Metrology in Industry
A. Bobbio, P. Tavella, A. Montefusco, S. Costamagna, “Monitoring the calibration condition of a measuring instrument by a stochastic shock model”, IEEE Trans. Instr. Meas., Vol. 46, no 4, pp. 747-751 (1997) DR. Cox, HD. Miller, The Theory of Stochastic Processes, Science Paperbacks, London: Chapman and Hall (1965) Carroll Croarkin Measurement Assurance Programs implementation, NBS Special Publication 676-II (1984)
Part
II,
Development
and
John Mandel, Measurement and Statistics, Quality Progress (1981) Statistical Methods in Research and Production, 4th ed, edited by Owen L. Davies and Peter L. Goldsmith, London and New York: Longman (1984) Esa Vitikainen “When do we need calibration of equipment used in testing laboratories?” – Nordtest Report 226 (1994) Jean-Luc Vachette, “Continuous improvement of quality”, Editions d’Organisation (1990) Gérard Brunschwig and Alain Palsky, “Statistical control of processes (MSP) – Utilization of control charts”, Techniques of the Engineer R - 290 Marc Priel and Christian Ranson, “Let’s make sure of the quality of our measurements”, International Metrology Congress, Lyon (1991)
Chapter 7
Measurements and Uncertainties
7.1. Introduction Measurement results are necessary to make decisions. As a rule, it can be considered that the result of measuring constitutes a piece of technical information which gets passed over to a user. Aware of this information, the user will then be able to make a decision about: – the acceptance of a product (when measuring its characteristics or performances toward establishing conformity to a specification); – the setting (or adjustment) of a measuring instrument; – the validation of a process; – the setting of a parameter as part of the control of a manufacturing process (servo-control); – the validation of a hypothesis in the framework of a development; – the protection of the environment; – the definition of safety conditions for a product or a system; – the medical diagnosis. All these decisions work toward the quality of products or services. Whether the decisions taken are apt and wise directly depends on the quality of the received information, that is to say, on the results of the measurements.
Chapter written by Marc PRIEL – Laboratoire National de Métrologie et d’Essais (LNE).
164
Metrology in Industry
The quality of a measurement result can be described by its uncertainty. Uncertainty: a quantitative indication of the quality of a measurement result Together with a measurement result, the uncertainty makes it possible to provide a quantitative indication about the quality of the result. This piece of information is vital for the users of this result so that they are able to assess its reliability. Without the uncertainty, the measurement results cannot be compared either: – between themselves (values obtained by two laboratories, comparison of successive calibration results with a view to the possible modification of a correction); or – in relation to the reference values stated in a norm or a specification (measurement results are used to prove the product conformity). The firm makes its decision on the basis of the information.
7.2. Measurement of physical quantity Measuring amounts to attributing a numeric value to an observed property by directly or indirectly comparing it to a standard. A physical quantity is an observable property specific to an object, a system or a physical state. The mass of a body is characteristic of its inertia; pressure and temperature are characteristic of the thermodynamic state of a gas. Three inextricably connected elements are included in the expression of a physical quantity: – a numerical value; – a unit; – an uncertainty. The quantity to be measured is called the measurand. The systems of observation and comparison and the standard make up the measurement system. Note: the words “error” and “uncertainty” which stand for two different concepts must be carefully distinguished; they must not be confused or interchangeably used.
Measurements and Uncertainties
165
Uncertainty: a new concept introduced The concept of uncertainty is comparatively new in the history of measurement. For decades it was error that was calculated, but the fundamental difference between the concepts of error and uncertainty must be clearly defined. It is now admitted that once all the known or suspected components of the error have been assessed and the adequate corrections have been made, there still remains an uncertainty about the value of the stated result (the correction is done as accurately as possible, but it is never perfect).
probability uncertainty
error
value 1
value 2
result
value 3
true value
values that could be attributed to the mesurand
Figure 7.1. Presentation of the concept of uncertainty
The metrologist’s aim is to get a result close to the right value. In order to reach this goal, he will reduce systematic errors by applying corrections and random errors by repeating his measuring process. This new approach was initiated in 1980 by a working party formed within the context of the International Bureau of Weights and Measurements (BIPM). It has resulted in the publication of an ISO guide in 1993 entitled “Guide for the expression of measurement uncertainty”, also known under its acronym, “GUM”. This chapter is based on the concepts and notations written in the 1993 ISO guide. The GUM is referred to in numerous national norms.
166
Metrology in Industry
Notations used in the GUM Classical notations of statistics: – variance of X: V(X); – standard deviation of X: s (X). when these quantities are used to express uncertainties the following notations will be written: – variance of X: u 2 ( X ) – standard uncertainty of X: u ( X ) = u 2 ( X ) – covariance of X 1 and X 2 : u ( X 1 , X 2 ) – linear correlation coefficient: r ( X 1 , X 2 ) =
u(X 1, X 2 ) u ( X 1 )× u ( X 2 )
– combined uncertainty: uc ( y ) – expanded uncertainty: U = kuc ( y ) with k as coverage factor. It will be noticed that the u symbol found in the notations is the initial letter of the word uncertainty.
7.3. Analysis of the measurement process To make the analysis of the measurement process correct is most likely the toughest and trickiest task in the assessment of uncertainties. This analysis demands some technical abilities, an inquisitive mind and a sense of analysis. It can only be performed by somebody who perfectly masters the technique of measuring. Two methods can be recommended for the analysis of measurement processes: the cause and effect diagram method or the method which consists of using the list published in the GUM.
7.3.1. The cause and effect diagram method Finalizing the mathematical “right model” requires to have minutely analyzed the measurement process in order to identify the possible causes of uncertainty. There is a technique called the “cause-effect diagram”, which makes it possible – with some racking of one’s brain and a very good knowledge of the measurement process – to deduce all of the causes.
Measurements and Uncertainties
MEANS
167
METHOD
RESULT OF MEASUREMENT
MATERIAL
MEDIUM
MANPOWER
Figure 7.2. Cause and effect diagram method
Successively, the contribution of the means, the method of measurement, the medium (temperature, pressure, hygrometry, etc.), the operator and the measured object (measurand) will be analyzed.
7.3.2. Using the list published in the GUM (section 3.3.2) The following list (from the GUM) can also be used in order to have as exhaustive a list as possible: a) incomplete definition of the measurand; b) imperfect realization of the definition of the measurand; c) non-representative sampling: the measured sample may not represent the defined measurand; d) inadequate knowledge of the effects of environmental conditions on the measurement or imperfect measurement of environmental conditions; e) personal bias in reading analogue instruments; f) finite instrument or discrimination threshold; g) inexact values of measurement standards and reference materials; h) inexact values of constants and other parameters obtained from external sources and used in the data-reduction algorithm; i) approximation and assumption incorporated in the measurement method and procedure; j) variations in repeated observations of the measurand under apparently identical conditions.
168
Metrology in Industry
7.3.3. Errors Any measuring operation is inevitably marred by errors. Two origins for these mistakes are: – the measuring system; – a poor definition of the measurand.
The measuring system The measuring system is never perfect. It can be sensitive to the environment (effects of temperature, pressure, etc.); it is definitely not reliable (since a dispersion of values is observed when observations are repeated); even the standards used for its calibration are not exact. In fact, the primary standard is an imperfect materialization of the definition of the unit it is supposed to represent. The unit is conventionally defined by the International Committee of Weights and Measures (CIPM). When a standard is being established, the best to be done is to reproduce the definition as precisely as possible, without totally managing to.
An imperfect definition of the quantity is itself is a source of errors Simply consider the numerous details it would be necessary to give to obtain an exhaustive definition of the quantity to be measured. Let us take a simple example: an observer is asked to measure the length of a 1 meter standard gauge. Have we given precise enough details? Definitely not: the temperature at which we wish the result to be expressed has not been mentioned; but is that sufficient? If the system of observation is accurate and reliable within a micrometer, it will probably be sufficient. If its performances are 100 times higher, the position of the gauge in relation to the direction of the acceleration of the gravity will have to be given, because vertically the length of the gauge is shorter than if it is lying horizontally on a plane (it gets smaller under the effect of its own mass). If the gauge rests on supports, we know from the mechanics of continuous environment that its length will depend on the position of the supports. The error concept is ideal and errors cannot be known. The objective of the metrologist is to declare a result as close as possible to the true value. So, for this purpose, he has to reduce the errors.
Measurements and Uncertainties
169
7.3.4. Cutting down errors The terminology defined in the international vocabulary of basic and general terms in metrology (VIM) (International Vocabulary of Basic and General Terms of Metrology, BIPM, CEI, FICC, ISO, OIML, UICPA, UIPPA, 2nd edition, 1933) makes it possible to write the following equation:
Result of measurement = true value + error
It is always possible to split up the error into a systematic error and a random error.
Random error (VIM 3.13) A random error is obtained by a measurement minus the mean which would result from an infinite number of measurements of the same measurand carried out under repeatability conditions. Note 1: the random error is equal to the error minus systematic error. Note 2: because only a finite number of measurements can be made, it is only possible to determine an estimate of the random error.
Systematic error (VIM 3.14) A systematic error is the mean that would result from an infinite number of measurements of the same measurand carried out under repeatability conditions, minus a true value of the measurand. Note 1: systematic error is equal to the error, minus random error. Note 2: like true value systematic error and its causes cannot be completely known.
170
Metrology in Industry
Value obtained with an infinite number of repetitions
Measuring Result
True value
systematic
random
error
Figure 7.3. Random and systematic error
The following equation can then be written:
Result = true value + random errors + systematic errors
The objective of any metrologist is to provide a result close to the true value; hence the need to cut down the errors. How can these errors be cut down?: – generally random errors are cut down by repeating the measurements and calculating the arithmetic mean of the readings; – systematic errors are cut down by applying corrections. 7.3.4.1. Cutting down random errors by repeating measurements A random error probably results from unforeseeable or stochastic temporal and spatial variations of influence quantities. The effects of such variations, hereafter called random effects, entail variations for the repeated observations of the measurand.
Although it is not possible to compensate the random error of a measurement result, this error can generally be reduced by making a greater number of observations. Its mathematical expectation or expected value is equal to zero.
Measurements and Uncertainties
171
Note: the experimental standard deviation of the arithmetic mean or average of a series of observations is not the random error of the mean, although it is so designated in some publications. It is instead a measure of the uncertainty of the mean due to random effects. The exact value of the error in the mean arising from these effects cannot be known. (GUM section 3.2.2) 7.3.4.2. Cutting down systematic errors by applying corrections This is unquestionably the hardest operation for the metrologist because it requires a keen sense of analysis. The measuring process is to be scrutinized in order to identify as many causes of errors as possible, then the necessary corrections likely to compensate the assumed errors have to be assessed. A vast knowledge of the measuring process and of the involved physical principles is very often necessary to imagine the factors which may influence the result of the measurement. In practice, many sources of error can slip in: – effect of influence quantities (temperature, pressure, etc.); – bias of the instruments; – position of the measured object (warped mechanical part, depth of immersion of a thermoelectric couple, etc.); – perturbation of the measured quantity by the presence of the measuring instrument; – faulty correction of a result; – error in an algorithm of measuring results processing; – error brought in by the measuring method; – error brought in by the measurement procedure; – etc. Let us consider a very simple case: an operator uses a glass liquid dilatation thermometer. He has it calibrated by a laboratory which gives it back with a calibration certificate indicating a correction (appropriate around 20°C) equal to +0.3°C. The operator takes the temperature of a bath and he reads it as 19.3°C; the numeric value of his measurement result is then:
y = x + Ce y = 19.3 °C + 0.3 °C y = 19.6 °C
172
Metrology in Industry
where: – y is the numeric value of the measurement result; – x is a one reading (or the mean of readings if measurement process has been repeated); – Ce is the calibration correction. Generally, numerous corrections are made, either to attempt to make up for assumed errors or to express the results in standard conditions. The corrections can be grouped together in three categories: – corrections of calibration: determined by calibration and appearing in calibration certificates; – corrections related to the environment: compensate the effect of influence quantities such as pressure and temperature; to make these corrections you have to know the coefficient of sensitivity of the instrument to the different influence quantities; – corrections to bring the results back to standard conditions: it is customary in some fields of metrology to express the values of the quantities in normalized conditions. For example, in dimensional metrology, the values of lengths are usually expressed at 20°C.
7.4. Modeling of the measurement process 7.4.1. Measurement procedure and model of the measurement process When the process of measurement has been thoroughly analyzed and a certain number of causes of error have been identified, do not boast, but think of all those forgotten ones. Corrections making it possible to compensate for errors will be applied to the identified errors. These corrections will be as good as possible, but there will remain a doubt, an uncertainty concerning the value of the corrections. The next paragraph will examine how these different doubts combine. In relation to random-type errors, it will be decided, for example, to repeat the observations so as to cut down these errors, and a number of repetitions will be decided upon. This leads to the development of a measurement procedure. The process of putting this measurement procedure into a mathematic form is called the modeling of the measurement process. In other words, modeling the process means transcribing in a mathematical formula the way the experimenter uses all the
Measurements and Uncertainties
173
information at his disposal to calculate the measurement or test result he gets: for example, a series of readings of the instrument, the value of a correction read in a calibration certificate, the value of a quantity obtained from a book, the measuring or the assessment of the effects of an influence quantity, etc. On the other hand, the measurand Y is usually not measured directly; it is determined from N other quantities X 1 , X 2 ,....., X N through the functional relation f; the model for the process is then: Y = f ( X 1 , X 2 ,...., X N )
The corrections (or corrective factors) appear among the Xi, as well as some quantities which take all the other sources of variability into account: the different observers, the instruments, samples, laboratories and times of the measurements. Therefore, the function f does not merely refer to a physical law, but to the process of measurement or test; in particular, the function must consider all the quantities that significantly contribute to the uncertainty of the final result. When several input quantities X i , X j are contributory to a same quantity t, it is sometimes useful to write the developed mathematical model. Meanwhile the input quantities are made explicit according to that same quantity t in order to avoid the introduction of terms of covariance into the application of the law of propagation of uncertainties later on. See section 7.6.2 for an example of the application of the realization.
7.4.2. An essential stage for the assessment of uncertainty: modeling the measurement
You must be aware that the most critical phase of the evaluation of the uncertainty of a result happens when the mathematical model describing the measurement is being written. If you omit to introduce a correction into the model (even if it is estimated equal to zero, due to ignorance), you will forget about it when the law of propagation of uncertainties is applied. That is why the stage during which the measurement process is analyzed and as thorough as possible an assessment of the causes of error is made is the key part of the estimation of measurement uncertainties. Optimization of the number of measurements It is often possible to decrease the effect of random errors by increasing the number of repetitions. It is, however, useless, and even illusive, to increase it rashly.
174
Metrology in Industry
It is often possible to express the combined standard uncertainty by an expression such as:
uc =
s2 + u2 n
in which s represents the variance of repeatability of the measurement process and n the number of measurements defined in the measurement procedure; the result is referred to the arithmetic mean of n observations. Therefore, the uncertainty is the result of the combination of two terms; it can be admitted that in order to optimize the number of observations n, the two components have to be of the same quantity.
Uncertainties
Application: – let us suppose that s = 5 and u = 3; – let us estimate the optimal number of measurements; – the diagram below illustrates the situation; – the curve s / n shows the decrease of the variable part of the uncertainty as a function of n and the curve u = 3 shows the invariable part. As you watch the curve of the compound uncertainty you can observe that increasing the number of measurements n does not make the uncertainty drop dramatically, for n > 5 for example. 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Number of repetitions "n"
Figure 7.4. Evolution of the uncertainty as a function of the number of repetitions
7.5. Assessment of the uncertainty of the input quantities
When the model of the measurement process has been established, the contribution of each one of the input quantities to the uncertainty of the announced measurement result will have to be assessed.
Measurements and Uncertainties
175
In every process of assessment of the measurement uncertainty, the standard uncertainties u(xi) or the corresponding variances u2(xi) of each one of the components occurring in the combined uncertainty will have to be assessed. Two methods can be used to estimate the numeric value (standard deviation or variance) of each one of the components: Type A method and Type B method. If there are enough resources, all the components can be estimated with Type A methods. Type B methods require experience and technical skills.
7.5.1. Type A methods
Type A methods are based on the application of statistical methods to a series of repeated determinations. They are chiefly used to quantify the repeatability uncertainties of the measurement processes. When a measurement process is repeated while keeping (as well as possible) the same conditions, a scattering of the measured values is generally observed, if the measurement process has a good enough resolution. With n independent values xi , the best estimator of the standard deviations is given by the arithmetic mean of the individual values xi . The best estimator of the expectation of the population is given by the arithmetic mean of the individual values x. The estimator of the expectation is given by:
x=
1 n ∑ xi n i =1
The estimator of the standard deviation (experimental) is given by: s=
1 n ∑ (xi − x ) n − 1 i=1
As in the past, the operator often performs numerous series of measurements (the number of measurements in the series can be different) with the same method, the same procedure, the same instruments and in similar conditions. These different series will enable him to calculate some estimators of the variance of the population: s12 , s 22 ,..., s 2k .
176
Metrology in Industry
The number of measurements in each series being (n1, n2 ... nk), a better knowledge of the variance of the total population can be obtained by combining the different estimators (pooled variance): s2 =
(n1 − 1)s12 + (n 2 − 1)s 22 + ... + (n k − 1)s 2k (n1 − 1) + (n 2 − 1) + ...(n k − 1)
which can also be written depending on the number of degrees of freedom υi = n i − 1 : s2 =
υ1s12 + υ 2s 22 + ... + υk sk υ1 + υ 2 + ... + υ k
2
The application of this reasoning makes it possible to calculate the component of repeatability u 2 (x ) , for example: u 2 (x ) =
s2 n
Note: this method of calculation (pooled variance) enables a better assessment of the variance of repeatability of the measurement process because the estimator bases itself on a significant number of observations. The operator can then use this value to assess the variance of the average of his observations in his usual measurement process. For example, if in his routine measurement process the operator performs only one measurement, then u 2 (x ) = s 2 , because in this case n = 1. If the (routine) measurement procedure had planned five observations, then the variance would have been divided by five. This highlights the advantage of assessing the repeatability of the measuring process with preliminary tests (implementing the highest number of causes of variability of the measurement process) before starting the operation. 7.5.2. Type B methods
These methods are used to quantify the uncertainties of the different components occurring in the model of the measurement process: uncertainty about the calibration corrections, uncertainty about the environment corrections, etc. Type B methods are used when you cannot or you do not want to use statistical methods. These Type B methods will be based on the experience of the operators, on some tests, and on the knowledge of physical phenomena.
Measurements and Uncertainties
177
For each one of the Xj occurring in the model describing the measurement process, the corresponding standard uncertainties will be “assessed” by using all the available technical information (extent and a priori distribution of possible values). Example 1: a correction must be made in a measurement process, but this correction (xi) is not completely known; the only information you have is that between two limits (lower (aii) and upper (ais)), the value of the correction will be estimated by: xi =
1 (aii + ais ) 2
and the estimator of the corresponding variance will be: s x2 = i
1 (aii − ais )2 12
If the difference between the two limits (lower and upper) is noted 2ai, the equation above can be written: 1 s x = ai2 3 i
( )
from which the standard uncertainty of xi can be assessed as u xi =
a
. 3 These calculations correspond to a rectangle distribution, which means that xi is as likely to take some value or other in the interval [aii , ais]. Example 2: a standard mass is returned after calibration with its calibration certificate which specifies its deviation from the nominal value and a calibration uncertainty expressed as follows: U = 0.006g (k = 2); the standard uncertainty about the correction will be very simply assessed by dividing the expanded uncertainty U by the coverage factor k, that is u(Ce) = 0.006/2 = 0.003g. The following table sums up various practical cases. The first column specifies the type of the component, the second the a priori selected distribution law and the third indicates which calculations to make.
178
Metrology in Industry
Component Resolution device. Hysteresis.
of
an
indicating
A priori distribution rectangle
Calculation method If the resolution is b u = b / 12
rectangle
If the maximal difference between the indications obtained by increasing and decreasing values is b, then u = b / 12 Effect of influence quantities derivative of sine If the variations of the temperature arc varying between two extrema in a are referred to by ± a, then u = more or less sinusoidal way, for a/1.4 example, the temperature of premises whose temperature is regulated. Drift of a measuring instrument. If the analysis of the results of the successive calibrations reveals a tendency that can be modeled, then a correction is made. The uncertainty about this correction is assessed, for example by a regression technique.
Asymmetric components of the type: error of parallelism between the measured object and the standard in dimensional metrology; or pouring out the contents of a phial in chemistry (the quantity poured out is always smaller than the contents of the phial). Correction not done.
Instrument verified and conformity with a class.
in
right-angled triangle
If the examination of the results of the successive calibrations does not show any tendency, you can not talk of drift, but of reproducibility. If the process is under statistical control, a Type A method is used. If the basis of the right-angled triangle is equal to d, then u= d 18
rectangle
You make an error if, knowing about it, you do not do a correction. This has nothing to do with assessing uncertainties. Nevertheless, the GUM suggests a solution: see sections 6.3.1 and F.2.4.5. If the class is defined by ± a, then u =u=a/ 3
Table 7.1. Example of Type B evaluation of uncertainty
Measurements and Uncertainties
179
In summary, the Type B methods are based on: – the choice of a form of the distribution of Xi; – the assessment of the limits of the variation of Xi (the extension of Xi). 7.5.3. Comparing the Type A and Type B methods
The following table compares Type A methods and Type B methods. Type A methods
Type B methods
Experimental results, series of measurements.
Results of previous measurements-makers’ data, data obtained from calibration certificates or books. Assessment of a standard deviation from an extension and the choice of a form of distribution. Type B methods require some experience and scientific knowledge.
Use of statistical methods, techniques of assessment of statistical parameters. Type A methods require resources to perform experimental tasks.
Table 7.2. Comparison of the Type A and Type B methods
To conclude, too great a stress should not be put on the differences between these two approaches; a Type B method based on long experience is preferable to a repetition of observations that would not implement all the causes of variability. Conversely, when you have little experience, repetitions make it possible to get closer to the uncertainty. In section 7.7, an example will be found in which the uncertainty is assessed by only using Type A methods; the norm ISO 5725: “Accuracy of results and measurement methods” is put into practice. Note: the classification in A or B types applied to uncertainty is not a substitute for the word “random” or “systematic”, formerly used to classify the uncertainties. The expression “systematic uncertainty” must not be used. Example: alternative use of the Type A or Type B methods: an operator wants to study the effect on his measurement process of the influence quantity “temperature”. He has two options: either he measures the temperature at regular intervals, then he calculates the mean of the values and their standard deviation (Type A method), or he consults the characteristics of the air-conditioning system; if, for example, the system is set for a prescribed temperature of 20°C ± 2°C, all he will have to do is divide the half range (2°C) by root of 2 to assess the standard deviation (Type B method).
180
Metrology in Industry
7.6. Calculating the combined uncertainty on the result
Once the model has been worked out and the standard uncertainties of the input quantities of the model have been assessed, the law of propagation of uncertainty can then be used to calculate the combined uncertainty on the measurement result. The law of propagation of uncertainty makes it possible to calculate the “combined” uncertainty of y, uc(y), or rather its variance uc2 ( y ) :
u (y) = 2 c
N
∑ i =1
2
N −1 N ⎡ ∂f ⎤ 2 u (xi , x j ) ⎢ ⎥ u (xi ) + 2 i =1 J =i +1 ⎣ ∂xi ⎦
∑∑
The law of propagation of uncertainty, which in its general application may seem a bit complex, does in many cases get simpler. Note: it will be noticed that the partial derivatives represent the “coefficients of sensitivity of the result” to the different input quantities. For example, if in the mathematical model the temperature is mentioned as an influence quantity, then the corresponding partial derivative may represent the coefficient of temperature of the measuring instrument. 7.6.1. Situation when all the input quantities are independent
In this case the terms of covariance are zero and the law of propagation is more simply written: 2
⎡ ⎤ ( y ) = ⎢ ∂f ⎥ u 2 (xi ) ∂x i =1 ⎣ i ⎦ N
uc2
∑
or, introducing the coefficients of sensitivity c i : N
u c2
( y ) = ∑ ci2 u 2 (xi ) i =1
7.6.1.1. Situation when the input quantities are independent and the model is a sum
y = x1 + x2 + ..... + x N then:
uc2 ( y ) = u 2 (x1 ) + u 2 (x2 ) + ..... + u 2 (xN )
Measurements and Uncertainties
181
7.6.1.2. Situation when the model is a product y = x1 × x2 × ....... × x N
This type of model is frequently seen in chemistry. In this case, the relative variance of the result Y is the sum of the relative standard uncertainties for the different input quantities xi of the model: uc2 ( y ) y
2
=
u 2 (x1 ) x12
+
u 2 ( x2 ) x22
+ ...... +
u 2 (x N ) 2 xN
7.6.2. Situation when the input quantities are dependent
In this case the terms of covariance will not be zero any more. The covariance u (xi , x j ) can be assessed; three methods of assessment are possible: 7.6.2.1. Assessment of the covariances by assessing a coefficient of correlation r(xi, x j ) You can write: u (xi , x j ) = r (xi , x j )× u (xi )× u (x j ) A practical solution will consist of varying r for the extreme values, -1, 0, +1, and watching the values of the uncertainties on y and for safety’s and caution’s sake keep the utmost value of the uncertainty. It is also possible, through reasoning based on physics, to evaluate r, but this requires much experience. 7.6.2.2. Assessment of the covariances by calculating the terms of covariance In a case where you have two connected input quantities X i and X j , assessed by their means X 1 et X 2 , determined from n independent pairs of repeated simultaneous observations, the terms of covariance are expressed by u xi , x j = s xi , x j with:
(
(
)
s xi , x j =
1 n(n − 1)
) (
)
n
∑ (xi, j − xi )(x j,k − x j ) k =1
7.6.2.3. Assessment of the covariances by considering the terms common to two input quantities Suppose two input quantities X 1 and X 2 assessed by x1 and x2 are dependent on a set of unconnected variables Q1 , Q2 ,.....QL . In such a way as X i = F (Q1, Q2 ....QL ) and X j = G (Q1 , Q2 ,...., QL ) some of the variables possibly only appearing in one or the other function; if the variance associated with the estimation q k of Qk is noted u 2 (qk ) , then the covariance can be calculated by the following expression:
(
L
) ∑ ∂∂qF ∂∂qG u 2 (qk ) k k
u xi , x j =
k =1
182
Metrology in Industry
See the GUM section F.1.2.3. The example below uses this expression of the covariance to calculate the uncertainty about the sum of the two masses; the common term comes from using the same standard. The way the model is written may lead to simplifications when the law of propagation of the uncertainty is applied. When you write a model describing a measurement process, it can be randomly developed. Experience teaches us that it is advisable to develop written models. The fact is that if different input quantities are dependent on another quantity, these quantities are connected. Making clear their relations with the third quantity when writing the model makes it possible to avoid introducing terms of covariance. Let us take the following example: two masses A and B, whose nominal value is 50 g, are compared to a same standard E. Then, A and B are used together to make a 100 g standard: what is the uncertainty on the 100 g mass thus obtained y = A + B? The mathematical model can be written as follows: A = E + x1 B = E + x2 y=A+B in which x is the result of the comparison of the mass A to the standard E, and x2 the result of the comparison of the mass B to the standard E. If you apply the law of propagation of uncertainty directly, you get the following equation in which there is a term of covariance. This term comes from the fact that A and B have been calibrated in relation to the same standard E: uc2 ( y ) = u 2 ( A) + u 2 (B ) + 2u ( A, B )
If you take the precaution to simplify the model it can be written as follows: y = E + x1 + E + x2 If then you apply the law of propagation of uncertainties, you get an equation in which there are no more terms of covariance: uc2 ( y ) = u 2 (E ) + u 2 (x1 ) + u 2 (E ) + u 2 (x2 )
Of course the same result will be obtained if you consider that 2u ( A, B ) = 2u 2 (E ) (see GUM section F1.2.3); it will be noticed that the covariance of A and B is the variance of their common terms. What is common to A and B is the standard E. To conclude, it is better to use a developed written model of processes; it avoids introducing covariance terms.
Note: when the non-linearity of f becomes significant, you should include terms of a higher order in Taylor’s development for the expression of u c2 ( y ); see the GUM sections 5.1.2 and H.1.7.
Measurements and Uncertainties
183
7.7. Use of the performances of the method (repeatability and freedom of bias) to assess the uncertainty of the measurement result
The method developed in this section constitutes a means which supplements the procedure of the GUM (see Chapter 8) when you do not know how to, or you do not want to, write or use the mathematical model to describe the measurement process. This method is based on the idea that information can be drawn from the results of interlaboratory tests or intra-laboratory tests to assess the uncertainty. The method is described in the fascicule of documentation AFNOR X 07 - 021: Assistance to the process of assessment and use of the uncertainty of measurements and test results (1999). This idea has been taken up at the ISO level by the “statistical methods” 69 Technical Committee; they are the subject of the ISO TS 21748 publication “Guide to the use of repeatability, reproducibility and trueness estimates in measurement uncertainty estimation”. The publication “Guidelines on the expression of uncertainty in quantitative testing – EA 4/16” also develops this approach for the domain of testing activities. There are numerous situations for which the method for obtaining the result is complex enough to make it impossible to model it. This situation is particularly found in some test processes. In order to ensure a reproducibility of the results, the conditions in which the test method is implemented are vital and must be perfectly controlled. The quality of a test method is judged by its accuracy (see ISO 5725): – trueness of agreement between the average value obtained from a large series of test results and an accepted reference value; – precision of agreement between independent test results obtained under the stipulated conditions. Precision corresponds to a characteristic which quantifies a performance of a method; it means that a method is appropriate to supply test results which are very close to each other when the same product is tested several times with the respect of the test conditions defined by the method. You are always situated between two extremes: – repeatability (r): precision under repeatability conditions, where independent tests results are obtained with the same method on identical test items in the same laboratory by the same operator using the same equipment within a short period of time;
184
Metrology in Industry
– reproducibility (R): precision under reproducibility conditions where test results are obtained with the same method on identical tests items in different laboratories with different operators using different equipment. Other characteristics of the method (e.g., linearity, robustness, etc.) can also contain some interesting information to assess the uncertainty. The ISO/CEI 17025 norm provides that the validation data can be used to evaluate the uncertainty of the measurement result: Reasonable estimation shall be based on knowledge of the performance of the method and on measurement scope and shall make use of, for example, previous experience and validation data (section 5.4.6.2).
The most quoted characteristics which quantify the performances of the method are “detection limit, selectivity of the method, linearity, repeatability and/or reproducibility, robustness against external influence, etc.”. All these characteristics matter when making sure that a method is capable of meeting the needs of the customer of the test, but not all are useful for assessing the uncertainty. In general, knowing the repeatability, the reproducibility, the robustness of the linearity and the freedom of bias are sufficient to assess the uncertainty of the result.
7.7.1. Intra- or interlaboratory approaches
Several approaches are possible to assess the characteristics of a method; one is an intra-laboratory approach: the characteristics will be determined exclusively by tasks done within the framework of a laboratory. A collective approach (called interlaboratory) can also be conducive to the evaluation of the characteristics of the method. The collective approach is the richest in information since the sources of variability of the result are more numerous: different laboratories, different equipment and personnel, etc. In an intra-laboratory approach, it will be necessary to make sure that the largest number of causes of variability can be expressed during repeated tests so that the dispersion of the results is representative of the uncertainty. Those readers who might find it difficult to connect this with the traditional application of the GUM can imagine a “Type A super method”, which alters all the identified factors as having an effect on the measurement result. Figure 7.5 illustrates the possible different approaches by repositioning the procedure of the GUM described in Chapter 8. The branch entitled “analytical process” represents the classical approach developed in the preceding sections and
Measurements and Uncertainties
185
summarized in Chapter 8. The other branches present the channel “use of the method’s validation data”; this channel can be activated either by an intra-laboratory approach or by an interlaboratory approach.
Definition Definition of the measurand, Measuran , List of uncertainty components
Intra-laboratory Intra laboratory approach
Yes Analytical method
interlaboratory Inter laboratory approach
Physical model model? ? Including correction
No
Statistical model
Proficiency testing Proficiency testing Organization of Organisation of repetitions, repetitions, validation validation method method
Method accuracy Iso 5725 ISO 5725
Adding Adding others other uncertainty uncertainty factors e.g. factors e.g.uncertainty bias uncertainty on on thethe bias
Use of values already published Published + Uncertainty on on the thebias bias and factors not taken into account during interlaboratory study
Evaluation Evaluation of standard standard - uncertainties of uncertainties
Use of propagation law of uncertainty GUM
ISOguide guide43 43 Iso + Iso/Dis ISO/Dis13528 13528
Variability + Uncertainty on the bias and factors not not taken taken into account during interlaboratory intelaboratory study
Figure 7.5. Diagram of the different possible approaches for the evaluation of the uncertainty
7.7.2. Intra-laboratory approach
Although there is no physical model that describes the measurement process, there is still a statistical model for the data processing. This model can be written as: y = m +C Jus +C Lin +
∑c x + e i i
i
where: y = result of the measurement; m = true value; C Jus = correction of freedom of bias of the method; C Lin = correction of linearity; ∑ ci xi = corrective terms for robustness, sampling, time, the operator; i
e = residual error (repeatability).
186
Metrology in Industry
The law of propagation of variances is then applied to this statistical model to assess the variance on the final result y:
u 2 ( y ) = u 2 (cJus ) + u 2 (cLin ) +
∑ c u (x ) + S 2 2 i
i
2 r
The methods of evaluation of the different components will be presented in section 7.7.4 below.
7.7.3. Interlaboratory approach
Just as a statistical model has been established for the intra-laboratory approach, the same thing can be done for the interlaboratory approach with:
y = m +δ + B +
∑c x + e i i
i
where: y = measurement result; m = true value;
δ
= freedom of bias of the method;
B = laboratory effect;
ci xi = corrective terms for not included effects at time of interlaboratory tests; e = residual error (repeatability). The variance of reproducibility is the sum of the variance of repeatability and the intra-laboratory variance:
S R2 = S L2 + S r2 and the variance of the result will be noted:
u 2 ( y ) = u 2 (δ ) + S R2 +
∑ c u (x ) 2 2 i
i
The methods of assessment of the different components of the uncertainty of the result y will be presented in the next section.
Measurements and Uncertainties
187
7.7.4. Data processing for intra- and interlaboratory approaches
7.7.4.1. Assessment of the repeatability and the reproducibility The processing methods, whether for an intra-laboratory approach or an interlaboratory approach, will be similar for assessing the repeatability and the reproducibility. If you plan a test (for a level of the quantity) you should use a table in the following form: Measurements
Laboratories
1 : : i : : : p
Position
dispersion
y1
s1
y1
si
yp
sp
y11..........y1n1
yi1..........yin1
yp1..........ypnp
If the approach is intra-laboratory, the experiments will not be repeated in different laboratories, they will be repeated in the same laboratory. Two statistical tests (Grubbs and Cochran tests) will then be used (homogeneity test and elimination of ouliers). After checking the validity of the data, the average level will be calculated; it is the arithmetic mean of the different values: p
y=
∑y
i
i =1
p
then the standard deviation of repeatability sr: p
∑s
i
sr =
i =1
p
and the standard deviation of reproducibility sR: sR =
1 p −1
∑ (y − y ) p
2
i
i =1
+
n −1 2 sr n
188
Metrology in Industry
If the laboratory has applied the test method correctly and the obtained results could have been partly used in the interlaboratory test, it may first be satisfied when its uncertainty can related to the reproducibility as follows: uc ( y ) = s R
This statement is not quite correct, because if you adopt this solution, you actually modify the definition of the measurand; you are no longer concerned by the value announced by a laboratory, but by the average value resulting from the tests of all the laboratories. Considering that the standard uncertainty is equal to the standard deviation of reproducibility may lead to overestimating the uncertainty, which is being cautious, but it entails drawbacks, namely a standardization of uncertainty. This practice may conceal real differences of quality between different laboratories. It is preferable to give an attention to the intermediate repeatability. 7.7.4.2. Assessment of the freedom of bias (trueness) References must be available to be able to assess accuracy. Reference values may come from certified reference materials, values obtained from a reference method, values from an interlaboratory aptitude test, but you have to check that the reference value is traceable to the International Units System (SI). Corrections of bias are seldom applied in some fields (e.g., analytical chemistry); it is customary to improve the accuracy of the method until it is acceptable. This procedure is developed in chemical analysis; you have to be able to decide whether the bias is acceptable and the following test can be used. To calculate the normalized error En, if this quantity is lower than 2 the deviation from the reference is regarded as negligible: EN =
xi − xRe f 2 ui2 + uRe f
However, even if the deviation is not significant, the uncertainty of the reference will come into this process and at least it will be necessary to consider that the uncertainty due to the bias is equal to the uncertainty about the reference used:
u 2 (CJus ) = uRe2 f
Measurements and Uncertainties
189
7.7.4.3. Evaluation of the linearity To evaluate the linearity on the studied domain of measurement, n measurements are to be repeated at k levels of the quantity, then the calibration line will be estimated by the method of least squares. The deviations from the line are calculated (deviation between the value experimentally obtained and the value obtained by the model); these deviations are then tested by comparing them to the repeatability to determine whether they are significant. The following equation can be used as an uncertainty component related to the lack of linearity. In this equation, the maximal residual constitutes the largest deviation between the experimental points y and the modeled points y, by the calibration curve drawn by the method of the least squares: U (C Lin ) =
Residual Max
7.7.4.4. The terms
3
∑
ci u 2 (xi )
i
The reader has noticed that the terms of this type appear in the intra- or interlaboratory approach. They represent all the contributions to the uncertainty of the result which it has not been possible to implement, or that were not used when the tests were being repeated. For further details, see the norm ISO TS 21748.
7.8. Reporting of the measurement result
Applying the law of propagation of uncertainties makes it possible to assess a combined standard uncertainty uc ( y ) . For diverse reasons, the expanded uncertainty U has to be written as: U = kuc ( y )
in which k is the extending factor. The value of the extending factor k is chosen according to the level of confidence requested for the interval y - U, y + U; generally k = 2 or 3. Choosing k = 2 is the same as considering an interval with a confidence level of approximately 95%. The numerical values of the estimation Y and its standard uncertainty u (y) or U must not be given with an excessive number of digits. Two significant digits are usually enough for the standard uncertainty and the expanded uncertainty: Y=y±U
190
Metrology in Industry
As for the numerical value of the result, the last figure to retain is the one which holds the same position as the second significant figure in the expression of the uncertainty. The estimate of the measurand has to be rounded according to its uncertainty: for example, if y = 10.057 62 Ω with u c (y ) = 27 m, u c (y ) has to be rounded up to 10.058 Ω. 7.9. Example
Calibration of a mass: nominal value 10 kg (from an example published in the EAL R2 document, supplement 1) The calibration of an OIML M1 class, 10 kg nominal value mass is carried out comparatively to an OIML F2 class reference mass with the same nominal value by using a mass comparator whose characteristics have been determined beforehand. E1: Analysis of the measurement process The analysis of the measurement process shows the following causes of error: – value of the standard mass; – drift of the standard (durability of the standard); – repeatability of the comparator; – effect of the off-centering of the mass on the pan of the comparator; – thrust from the air. E2: Measurement procedure In order to eliminate the phenomenon of drift during the weighing process, a method of substitution called Standard Mass Mass Standard (SMMS) will be used: the standard, then the unknown mass, then the mass again, and finally the standard are placed on the pan of the comparator. In order to reduce random errors, the weighing process is repeated three times. E3: Mathematical model of the measuring process mx = ms + δmD + δm + δmc + δB
where: m x : value of the unknown mass (conventional mass); ms : value of the standard mass (conventional mass);
δmD : drift of the standard mass since the last calibration;
Measurements and Uncertainties
191
δm : difference observed between the unknown mass and the standard; δmc : correction to make up for the error due to the off-centering of the mass; δB : correction of thrust from the air. E4: Estimation of the standard uncertainties on the input quantities of the model – Reference standard ( ms ): the calibration certificate indicates the value of 10,000.005 g with an expanded uncertainty of 45 mg (extending factor k = 2). 45 = 22.5 mg. Therefore, the value of the standard uncertainty is u (m s ) = 2 – Drift of the standard ( δmD ): the drift of the value of the standard mass is inferred from previous calibrations; its value is considered equal to zero with variations of ± 15 mg. If a rectangular distribution is surmised, the value of the corresponding standard uncertainty is: u (δmD ) =
15 3
= 8.66 mg
– Comparator ( δm , δmc ): a previous evaluation of the repeatability of comparison of two masses having the same nominal value of 10 kg has resulted in a variance (accumulated; see section 7.5.2) of 625 mg2. No correction is applied to make up for the variations due to the off-centering of the masses on the pan, but it is considered that these effects result in a maximal variation of the indications of the comparator of ± 10 mg; if a rectangular distribution is surmised, the corresponding standard uncertainty is: u (δmc ) =
10 3
= 5.77 mg
– Thrust of the air ( δB ): no correction is applied to make up for the effects of the air thrust. The limits of the possible variations are estimated to be at most ± 1 x 10-6. – Correlations: a survey of the different input quantities of the model does not show any correlations. E5: Making the measurements Three observations of the difference between the value of the unknown mass and that of the standard mass are made by using a substitution method whose sequence is SMMS.
192
Metrology in Industry
Series no.
Mass
Readings
1
Standard Unknown Unknown Standard Standard Unknown Unknown Standard Standard Unknown Unknown Standard
+0.010 g +0.020 g +0.025 g +0.015 g +0.025 g +0.050 g +0.055 g +0.020 g +0.025 g +0.045 g +0.040 g +0.020 g
2
3
Differences observed
+0.01 g
+0.03 g
+0.02 g
The arithmetic mean is δm = 0.020 g .The estimator of the standard deviation of repeatability of the weighings (estimated by an accumulated standard deviation of tests carried out earlier; see section 7.5.1) is s p (δm ) = 25 mg . Thus, the standard uncertainty on the mean of the three measurements is u (δm ) = s(δm ) = 25 mg = 14.4 mg . 3
E6: Calculation of the combined uncertainty, application of the law of propagation of uncertainty The mathematical model of the measurement process is written: mx = ms + δmD + δm + δmc + δB
The law of propagation makes it possible to calculate the variance on the value of the unknown mass: uc2 (mx ) = u 2 (ms ) + u 2 (δmD ) + u 2 (δm ) + u 2 (δmc ) + u 2 (δB ) – Synthesis table Contribution to uncertainty
Quantity
Estimator
Standard
Xi
xi
u (x i )
ms
10,000.005 g
22.5 mg
normal
1.0
22.5 mg
0.000 g
8.95 mg
rectangle
1.0
8.95 mg
0.020 g
14.4 mg
normal
1.0
14.4 mg
0.000 g
5.77 mg
rectangle
1.0
5.77 mg
0.000 g
5.77 mg
rectangle
1.0
5.77 mg
δmD δm δmc δB mx
10,000.025 g
Probability distribution
Sensitivity uncertainty
Ci
ui (y )
29.3 mg
Measurements and Uncertainties
193
It will be noticed in this table that the sensitivity coefficients (partial derivatives) are equal to 1; this comes from the fact that the mathematical model of the measurement process is a sum. E7: Expression of the final result and its uncertainty Expanded uncertainty: U = k × u (mx ) = 2 × 29.3 mg ≅ 59 mg U = k × u (mx ) = 2 × 29.3 mg ≅ 59 mg
Final result The fiducial value of the 10 kg nominal value mass is: 10,000.025 kg ± 59 mg (k=2).
7.10. Bibliography Norms and general documents Guide to the expression of uncertainty in measurement ISO (1993) Accuracy (trueness and precision) of measurement methods and results, ISO 5725 Metrology and application of statistics – help for the process for the estimation and the use of measurement and test results uncertainty, AFNOR X 07 - 021 (1999) Guide to the use of repeatability, reproducibility and trueness estimates in measurement uncertainty estimation, ISO/TS 21748 The expression of uncertainty and confidence in measurement, NAMAS M 3003 Barry N Taylor and Chris E Kuyatt, “Guideline for evaluating and expressing the uncertainty of NIST measurement results”, NIST Technical Note 1297, 1994 edition Stephanie Bell, “A beginner’s guide to uncertainty of measurement”, Measurement good practice guide No 11 (1999) National Physical Laboratory, Teddington, UK Magnus Holmgren et al. Measurement uncertainty leaflet (SP INFO 2000 27 uncertainty pdf), SP Swedish National Testing and Research Institute Quantifying uncertainty in analytical measurement EURACHEM/CITAC Guide CG4, 2nd ed, QUAM: 2000. 1 Expression of the uncertainty of measurement in calibration, EAL - 4/02 (December 1999) EA Guidelines on the expression of uncertainty in quantitative testing – EA-4/16 Eurolab technical report no. 1/2002 June 2002 Measurement uncertainty in testing ILAC – 17: 2002, Introducing the concept of uncertainty of measurement in testing in association with the application of the standard ISO/IEC 17025
194
Metrology in Industry
Books Twenty-seven Examples of Evaluation of Calibration Uncertainty, Collège Français de Métrologie, (1999) Christophe Perruchet, Marc Priel, Estimer l'incertitude – Mesures Essais (Assessing uncertainty – Measurement and tests), Afnor (2000) ISBN 2-12-460703-0 Ignacio Lira, Evaluating the Measurement Uncertainty: Fundamentals and Practical Guidance, Institute of Physics Publishing (2002) ISBN 0-7503-0840-0
Chapter 8
The Environment of Measuring
This chapter might be summed up as: – “It is not because a measuring instrument is new that it is good.” – “It is not enough to use good equipment to make good measurements.” The result of a measurement is the conclusion of a process which is comprised of: – the implementation of a method; – the utilization of measuring equipment; – the intervention of operators; – a physical environment (temperature, etc.); – a measurement procedure. All these elements have an influence on the result. Thus, it is essential to make a few general points: – depending on the expected accuracy, the place where the instrument is used is analyzed in order to reveal any possible significant interactions; – so as to make sure of the quality of the measurement results, the qualification of the operators has to be checked and ascertained. As in any field, it is important to ensure the suitability of both the manpower and the function; – it can be difficult to guarantee the quality of these activities without a good document which describes, among other items, the measurement procedures. Chapter written by Jean-Yves ARRIAT – Ascent Consulting, and Marc PRIEL – Laboratoire National de Métrologie et d’Essais (LNE).
196
Metrology in Industry
These are, according to us, the main points that should be taken into account and are what we define as the “environment of measuring”.
8.1. The premises In order to successfully carry out the operations of measurement, calibration or verification, as well as of storing the instruments when they are not used, there are a certain number of processes: – to define safe storing areas, fitted out so as to prevent damage to or premature deterioration of the equipment; – to define the procedures for the reception and the dispatch of the material (when instruments are sent away for maintenance or calibration); – to define appropriate instructions to ensure that the premises are kept clean; if particular conditions of hygiene and cleanliness are required and specified for the measurement procedures, the cleaning and maintenance must be thorough; – to define appropriate instructions about maintenance and protection (against corrosion, for example). For equipment which requires periodical maintenance, some instructions must indicate how to deal with this maintenance. It may sometimes happen that the cleansing products are not compatible with the measuring premises (for example, emanation of alcohol or chlorine, etc.); – to know and control the environmental conditions as well as the influence quantities which should be taken into consideration. These environmental conditions are of differing natures; depending on the measurements required they can be: – the average temperature and its variations as a function of time and space; – the atmospheric pressure; – the relative humidity of the ambient air; – the quality of the air, the dust and the drafts; – the shocks and the vibrations; – the various fluctuations related to the supplies (power, fluids, etc.); – the radioelectric disruptions. These main parameters cannot be completely controlled and kept independent from the outside environment, so they will have to be maintained within certain limits (defined in accordance with contractual demands). Thus, the user determines these limits depending on the uncertainty on the measurement results that is sought. It is advisable to record the evolution of these parameters over time.
The Environment of Measuring
197
It will be necessary in some firms to reserve a place specifically for calibrating and verifying the measuring equipment. For more information about the creation of a calibration laboratory, see the bibliography in section 8.4. The requirements about premises depend on: – the physical parameters (for example, thick lead walls and remote controls are necessary for the measurements of ionizing radiations); – the uncertainties (for example, in the national metrology laboratories, the calibrations of gauge blocks are taken with interferometers whose temperature is known within a few hundredths of a degree Celsius; for the measurements of components with margins of a few hundredths of a millimeter, variations of a few degrees in the workshop will be acceptable). Based on our experience, we would suggest that: – north-facing exposures are preferable; – an indoor curtain insulates from the light and an outdoor curtain insulates from the sun so that the room cannot become warmer; – personnel and equipment require sufficient space so that two operations do not influence each other; – external disruptive activities should be avoided (arc-welding instruments for stamping press, etc.); – electric wiring should be up to the norms with an earth plug adapted to instruments of measurements; – smoking should not be allowed. We will now take a closer look at some of the parameters. We suggest that the reader carefully note each one of them. If the reader thinks that some of the parameters do not concern him, he will be wrong. By way of example, the remarks about electric measurements concern most laboratories as there are electronic devices which can be sensitive to radioelectric disturbances in all measuring instruments instruments.
8.1.1. Ambient temperature This is not subjected to any particular requirement (except, of course, contractual requirements). Nevertheless, the engineering industries work at around 20°C ±2°C and 65% RH ±10% HR, complying with the recommendations of the international ISO no. 1 standard which sets the reference ambient temperature at 20°C. Electricians prefer to use the value 23°C ±1°C and 50 % RH ±10% RH as reference temperature, which complies with the criteria of the ANSI D 2865 and D 3865.
198
Metrology in Industry
Recently, attempts have been made to standardize the reference temperatures (a change from 20°C to 23°C for various reasons: comfort of the operators, standardization in firms with mechanical and electronic activities, and also decreasing costs of air conditioning in tropical countries). As a change in reference temperature would result in a large number of changes in many companies (new plans, checking tools, etc.), as well as in costs of such changes, it has been decided to maintain the status quo. In many cases, keeping the temperature at about 2 to 3°C will be satisfactory. Depending on what uncertainties are sought, a fluctuation of 0.4°C to 0.6°C will also be satisfactory. We would like to emphasize the approach a firm should adopt: the specifications on the conditions of temperature have to be established according to the margins on the manufactured items, the uncertainties of measurement required to master the manufacturing processes and the uncertainties of measurement which establish the conformity of the manufactured products. It is one of the firm’s responsibilities to check their implementation. We would draw the reader’s attention to a very important point: the cost of installation. You may wish to have a very hi-tech installation to make life easier, but you must also have the means to ensure its maintenance; it is not enough to have the funds to buy it, you also have to keep it functioning over time. In metrology, good working organization and an ability to meet deadlines and under pressure are usually the required qualities.
8.1.2. Relative humidity Regarding causes of error, in practically all the fields of measurement, relative humidity (RH%) has comparatively little influence. It can nevertheless generate the following problems: – too low a rate of RH% causes discomfort to the personnel who have to remain on the premises; – big or sudden variations in time between the place of use and the place of calibration can generate abnormal drifts at the level of the supports of the standards of resistor – or capacitor – standards; – if the RH% value is too high, it can cause damage to the measuring equipment due to oxidation of the contacts, variation of the insulation resistance, corrosion, etc. Too high an RH% must be subjected to a measurement procedure when the laboratory is situated in an area where the humidity rate is high. A stocking time
The Environment of Measuring
199
must be determined before the instrument is plugged in, in order to avoid harmful condensations; likewise, when the equipment is temporarily stored after calibration, means have to be found to make it possible to control the environmental conditions as well as possible.
8.1.3. Handling of the air conditioning systems Particular attention must be given to the handling and the maintenance of air conditioning systems; in some metrology laboratories, temperature is one of the essential components in the budget of uncertainty. It could be considered that air conditioning should be looked upon as a measuring instrument and be as well looked after. Technical files with the recordings about all the maintenance operations and adjustments, and charts of the temperature readings should be kept.
8.1.4. Power network The fluctuations of the voltage of power supply may affect the performances of the electrical measuring equipment. The variations of the effective voltage may appear in two ways: – slow variations of voltage, which are generally attenuated by the equipment itself; – rapid variations of voltage, which require an external adjustment. It may be useful to dispose of several power supplies, when it is justified by the activity of the laboratory, i.e.: – a general circuit (lighting, air conditioning, various equipment); – a voltage – regulated and filtered measurement – circuit; – an emergency circuit: supply of the reference standards which need to be working permanently, e.g. the battery cases and, to a lesser extent, the thermostatcontrolled baths where the standard resistances are kept.
8.1.5. Radioelectric disturbances One should consider taking certain precautions in order to protect the measuring equipment from the influence of these disturbances, especially in those laboratories that are close to a strong source of radioelectric disturbance (radiodiffusion transmitter, etc).
200
Metrology in Industry
The “electrical earth” or “grounding” must be the object of the precautions; it is sometimes useful to have one earth specifically connected to calibration equipment. The safety regulations require that the personnel must not be able to access to two different earths. Thus, precautions should be taken at the time of the implementation when the equipment is installed to ensure the security of the operators. In some geographical areas, or when carrying out some measurements, it is necessary to protect oneself against the radiation that is emitted. Consequently, a Faraday cage should be available, or all the laboratory, or part of it, should be screened. 8.1.6. Measurements on-site In many cases, the firm must calibrate the measuring equipment on-site where they are used, either because the instruments cannot be transported or because once they are installed they are not easily dismantled. The calibration equipment used has to be specifically developed for that use (robustness, container for the transportation, autonomy, etc.). The factors that influence the different environmental parameters likely to be found on the site have to be assessed. A specific procedure for the assessment of the uncertainty should be prepared. It should, in particular, take into account the “sensitivity coefficients” of the instruments to the different influence quantities (see Chapter 7). 8.2. The personnel 8.2.1. The connection to the metrology function It is necessary to secure independence for the metrology function; it is often connected to the quality manager. When it is connected to quality, this type of organization provides the metrology function with: – the authority it needs to do its work; – the independence from the other services which makes it possible to avoid the pressures (in particular, from production) that might influence the judgment and the work of the personnel concerned. Assuming that its metrology function does not automatically result in a company creating a laboratory equipped with expensive material, the company can simply obtain a few references such as boxes of gauge blocks, of smooth rings, of reference temperature gauges, etc. These references will then be used to check such measuring means as calipers, micrometer screws, air-conditioned chambers, etc.
The Environment of Measuring
201
The metrology function can subcontract out of the firm all or part of its activities, or delegate some to other sectors of the firm (especially if it is an industrial firm) but the person in charge of the metrology service remains responsible for the metrology function and continues to manage it.
8.2.2. Staff involved in the metrology function Metrologists must have the technical competence required to do their job. Their job is precisely defined. The person responsible for the service ensures that the qualification and experience of the personnel is maintained at an appropriate level through continuing education. There are different ways to achieve this: – the circulation of scientific and technical journals; – information and training meetings; – the participation in the work of vocational groups; – training courses, etc. The basic need for technical knowledge must not ignore certain useful human qualities such as precision, which is not the least of them. Training-activity records should be permanently available, and should include, among other pieces of information, the results of the activities. The training gets started according to pre-established schedules; it should disturb the metrology activities as little as possible. The metrology function also often takes the role of adviser about the choice of measuring instrument and it participates in the training of the personnel who use the equipment. Therefore, it should be aware of the need for information and should inform the other people in the firm about the existence of courses that are in their fields of activity, or likely to interest them. The metrology function puts them in touch with different working entities which can answer their queries as far as possible. Inexperienced or temporary personnel can undertake measuring operations, but only if this does not entail any risk of prejudice to the quality of the measurements. Such personnel should not be left on their own and there should be more experienced personnel than inexperienced personnel.
202
Metrology in Industry
8.2.3. The qualification of the personnel Some regulated activities require a certification; it can be obtained through organizations approved by the state authorities. The fact is that for delicate operations, the operator has to demonstrate his skill and it should be approved. Even if there were no formal requirement to do so, it would seem sensible to adapt the knowledge of the personnel to the demands of the activities they undertake. The personnel are the motor of the firm; their training makes it possible to ensure that: – their abilities are appropriate to the needs of the firm; – the abilities evolve and adapt to the technology and to the requirements of the markets; – there are faster and safer initiatives and decision-making. Training structures for and qualification of the operators make it possible to define the types of measurement or calibration the personnel are able to undertake. Records of all training and qualifications are indispensable. See Chapter 12 for further information about the metrological profession.
8.3. The documentation 8.3.1. Filing of the documents Given the number of documents that exist in a firm in relation to the metrology function and their diversity, it is important to be well-organized, precise and methodical when dealing with them. There are two categories of documents. 8.3.1.1. Documents dealing with the quality system These documents, which are reference documents, define the criteria that the firm (or the laboratory) has set up to deal with quality and, consequently, with metrology. They make up the firm’s “reference system”. Different notions have to be taken into account regarding these documents: – the national and international norms; – the firm’s internal norms, if it has created any; – the technical documents (directions, guidebooks, etc.) from outside the firm (suppliers, clients, etc.); – the internal documents (programs, procedures, measurement procedures, instructions, etc.);
The Environment of Measuring
203
– the files related to the measuring equipment which can include specifications, as well as the copy of the order, the report of revenue, the documents about maintenance, about calibration and verification, etc. These documents should be easily accessible. 8.3.1.2. Records regarding quality This second category (the documents concerning measurements) makes it possible to preserve the primary results of the measurements so as to be able to repeat all the investigations that might be needed in the future. Preserving these data also makes it possible to show that the measurements have actually been taken, and to build up confidence between the client and the supplier of the measuring operations. In addition, a clever use of this data makes it possible to be more accurate about the intervals of calibration and to extract information on the quality and the condition of the different materials. These documents include, for example: – the measurement records; – the calibration certificates; – the identification sheets of the metrological means; – the monitoring cards of the measuring instruments. These documents should be handled and set up with great care before they are used. The time spent considering and specifying what you want is seldom wasted. It often avoids later corrections, adjustments or alterations which are real problems for quality. The documents in which the measurement results are saved must be clearly presented. Presentation must be given special attention and care, especially on the transcription of the parameters and the measurement results; and remember to make the documents reader-friendly. Similar documents should be as uniform as possible. The results must be laid out accurately, clearly, unambiguously and in full, in conformity with instructions which may be part of the method of measurement. The results are given with their uncertainty, either calculated or estimated. If anything has to be added to, or if corrections have to be made in, the measurement files, this has to be done with clarity and must be ambiguous. The signatories of the documents and the meaning of their signatures or initials must be explained in a separate document. The importance of the signature must be emphasized; it creates an awareness of responsibility for any metrological action. Every document must be dated.
204
Metrology in Industry
8.3.2. Management of the documents The management of the documents is based on different stages. It is vital for the reader that he or she should not to forget that a document is not created for the personal satisfaction of its author, but to satisfy a need. The phase of creation is fundamental, for it meets a need that has been expressed. Every document has to be checked, preferably by an outsider, to facilitate the detection of errors. Depending on the importance of the document, the verification will or will not be done by the signatories. These operations are done by different members of the personnel. In general, the approver who is at least as competent is not the drafter. How many people should sign the document? Not too many; two signatures (the drafter’s and the approver’s) are likely to be enough. The verification may entail some modification. If so, the document will be re-examined after the modification, prior to approval. Only the documents created inside the firm need to be submitted for approval. Documents such as work instructions have to be read by the users (there is nothing against involving the users in the drafting; sometimes it is advisable) so that they can give their opinion before the final approval. This should make the integration of the documents easier as ownership of the documents will have been given to the users; the users will not be in a position to reject a document that they do not know. Controlling the circulation of documents makes it possible to have the relevant editions of the appropriate documents at all the necessary places; who should receive a document should be determined at the time of its drafting. It is useful to put the documents into charts with the following information (these lists can preferably be computerized): – the sources of the documents; – the titles of the documents; – the category of readers the document target; – the name of the signatory persons; – the latest edition in use; – the frequency of revision, etc. These charts make it possible to know at all times the titles of the documents in use and the name of their present readers.
The Environment of Measuring
205
A system of “acknowledgement of receipt” proves that the documents have been received. It is also necessary to ensure that old editions have been regularly withdrawn from the circulation except those that are retained for the archives. The people who use the documents should immediately say if they do not understand a document or if a document is outdated. Any irrelevant document should not remain available; it could lead to errors and a loss of credibility in other documents. Doubt is a generator of chaos. The reference documents should be regularly revised, according to a scheduled frequency. Modifications may be necessary following: – new needs of the users; – a change in the contents of the documents; – internal audits of the services that use the documents. If modifications are necessary, a new edition must be brought out. Depending on how important the changes are, it can mention the significant modifications that have occurred since the previous edition. As a rule, any modification must entail a re-examination and approval by the metrology functions who originally approved the documents. Some documents dealing with contractual requirements or security have to be archived in special conditions and for minimum periods. It is then necessary to make arrangements for this.
8.4. Bibliography National Conference of Standards Laboratories, Recommended Practice – Laboratory Design (July 1986) National Conference of Standards Laboratories, Recommended Practice RP-3, Calibration Procedures (January 1990) Monograph no 7 of the BNM , CHIRON publisher M. Priel and B Schatz, “Organisation d'un laboratoire d'étalonnage” (Organization of a calibration laboratory) Techniques de l'ingénieur – R 1215 – France L. Erard, “Constitution type d’un laboratoire de référence en métrologie électrique” (Typical constitution of a reference laboratory in electrical metrology) Techniques de l'ingénieur – R 925 – France
206
Metrology in Industry
8.5. Appendix Major elements applicable to metrological activities (calibration, verification, etc.) to be taken into account when drafting a procedure When you write a procedure, you must include a certain amount of information. The level of information must be suitable to the level of knowledge of the potential readers. The following are the main headings that you ought to consider, even if all of them are not used: 1. Purpose and scope of application of the procedure 2. Physical principle of the method of measurement 3. Reference to the norms in use, bibliography 4. Limitation of the method – scope of measurement – uncertainty of measurement – types of equipment concerned by this method (category and main characteristics) – satisfactory environmental conditions (considering what uncertainties are expected) 5. Reference materials (related to national standards) – draw up the outlines of the traceability to the national standards 6. Maximum errors permissible, or uncertainties 7. List of the equipment and accessories to implement – diagram of assembly – special instructions about the use of the material 8. Preliminary operations The purpose of these operations is to guarantee the validity of the process after you have ensured that the instrument works correctly; the description of these operations can be found in specific documents. The operations have to be realized so that the validity of the verification, or of the calibration, can be ensured. The following are examples of these operations:
The Environment of Measuring
207
– the stabilization of the temperature of measuring instruments – the setting of the mechanical zero of analog instruments – the cleaning (and demagnetization) of the gauge blocks – the switching on beforehand of the electrical measuring instruments, etc. 9. Applicable measurement procedure The mode of operation is the main part of the procedure, so it should be welldeveloped. The description of this mode of operation can be found in specific documents. The written procedure must precisely define the sequences of the different operations and, when necessary, refer to the instructions for the software that is used. The measurement procedure indicates the number of points of measurement to be undertaken and the predetermined values to take on the scale of measurement. This will be the largest part of the document; it contains the firm’s know-how and, as such, it is often confidential. The procedure should be adapted to the level of competence of the operator in charge of the work. The question of relation between mode of operation and procedure is often raised. From our point of view, the mode of operation is the paragraph of the written procedure that contains the detail of the operations. However, depending on how complex the procedure is and whether the operators have different levels of qualifications, several modes of procedure (more or less detailed) may have to be written for the same procedure. 10. How can the raw results be processed when necessary? 11. Assessment of the uncertainties of measurement – related to the method – related to the calibrated or verified equipment (short-term repeatability, resolution, discretion, etc.) See Chapter 7 on this particular point. 12. Presentation of the results 13. Criteria for decision-making when a verification is in question 14. Document of evidence (recordings about the quality as it is understood in the ISO 9000 norms)
208
Metrology in Industry
This document completes the procedure; it contains the results that have been obtained from the calibration or the verification. At least one copy should be kept to ensure traceability has been achieved. The document will be the calibration certificate if calibration has occurred. In the case of verification, the report of the verification will show which decision has been taken about the measuring instrument verified. Whether a calibration or a verification, the operation will appear in the instrument’s file and will be noted on the instrument’s identification sheet. For further information, you can consult the French documentation fascicle of AFNOR titled “Practical method for the drafting of the procedures of calibration and verification of measuring instruments”.
Chapter 9
About Measuring
9.1. Preliminary information 9.1.1. Physical quantity Set a problem correctly and it is half solved. Therefore, first of all, it is necessary that you should know well the physical quantity, or quantities, to be measured. In the easiest cases it is enough to determine one single quantity: a mass, a temperature, a length of time, an electric value, etc. In many applications some set of quantity has to be measured: – several dimensions of a component; – several electric features of an instrument; – the timing of several events. Finally, when the quantity measured is very sensitive to an “influential parameter”, it is essential to determine this parameter together with the considered quantity. For example: – the mass of a powder does not mean anything unless you know its water content; – measuring a Weston battery is of no use if its temperature is not known;
Chapter written by Claude KOCH – retired.
210
Metrology in Industry
– since the coefficients of expansion of metals is never equal to zero, the temperature of gauges is always taken when they are measured. The example of quartz is not so well-known. Even when set in air-tight bulbs, quartz is slightly sensitive to atmospheric pressure. So, even in relation to the best quartz oscillators, one should, strictly, take into account the atmospheric pressure at which they are used.
9.1.2. The object to be measured The choice of instruments, the methods and the precautions will vary depending on the object to be measured. Thus, the pressure exerted by a sensor to measure dimensions is acceptable if the part you examine is made of metal, but it must be rejected if the object is soft. The length of a material will raise other measuring problems. Finally, if the object the length of which you want to know is a red-hot metal ingot, you will have to use non-contact, then optical, methods. Another example: electric resistances with two, three or four terminals require different methods and measuring equipment.
9.1.3. Field of measurement The field of measurement is the set of values that the quantity to be measured can take; this field is entirely defined by the minimal and the maximal values of the quantity. The range of measurement is the difference between the minimal value and the maximal value. It follows from these definitions that the range can be deducted from the field, but not the reverse. Therefore, it is far more favorable to know the field rather than the range. Example: quantity in temperature In a catalogue, a manufacturer introduces five types of mercury thermometers with a resolution of 0.1°C covering the following fields: -20 to +10°C
0 to +30°C
+20 to +50°C
+40 to +70°C
+60 to +90°C
These five types have the same range of 30°C, but their various fields design them for totally different applications.
About Measuring
211
9.1.4. Four types of uses of measuring instruments For research, it is advisable to have accurate multifunction instruments at one’s disposal. Some electronic instruments can be supplemented by filing cards or drawers, which saves having too many instruments. On the other hand, it will not be necessary to go through a large field of working temperatures, or to be over-careful about shocks, as in a research laboratory the instruments are subjected to low variations of temperature and are not moved about much. In manufacturing, the most adequate type is the automatic monofunction instrument, which is very well-suited for these conditions of use; it is moderately robust; its precision and its price are limited. Digital display instruments do very well for manufacturing; – when the analog/digital conversion has been done, the displays can, if needed, be situated at a distance; – this type of display can be used unambiguously by anyone, whereas reading a non-digital dial requires interpretation from the operator; – digital instruments can be equipped with thresholds to automatically find out those results that do not fit in a given range; – digital measuring makes it easier to pass the measurements to a global control by computer. Nevertheless, a digital instrument is not to be used when the operator has to do an adjustment because then the display changes constantly and the operator cannot read its variation, nor even which way it varies. For a building site, you have to choose instruments that are automatic and multifunctional and the accuracy of which is limited; however, they must be watertight and very robust. For a metrology laboratory, monofunction instruments with high accuracy and resolution will be preferred. Whenever you have to make a choice, you should choose types without in-built references and purchase separate references, for example: – length comparators and separate gauges; – Wheatstone’s bridges and separate electric resistances in order to be able to use several references alternately, then you can go on working while some references are left with a calibration laboratory. Ease of calibration and verification should be taken into account.
212
Metrology in Industry
If you have to choose between adjustable references and fixed references, choose the latter because the traceability of a fixed element is easier to establish; an adjustable element may have been modified without it appearing in its file. For traceability, see Chapter 5. Metrology instruments should not be subjected to rough conditions of use; for example: – no shocks; – no vibrations; – restricted temperature field; – possibility of leaving the electronic instruments working permanently; – handling by qualified personnel. All this can be taken into account when selecting the types of instruments. Finally, it is no use choosing instruments that are automatic or equipped with a remote control, nor electronic instruments that reach their nominal characteristics after only a few minutes.
9.1.5. Influencing quantities Whatever the principles of measuring instruments, whatever the quality of their manufacture, it is impossible to make them proof against influencing quantities, especially the following two which almost always interfere: – Temperature, which dilates substances, alters the characteristics of electronic components, modifies the viscosity of fluids, etc. In order not to be affected by temperature variations, references are kept in air or oil thermostat-controlled chambers. The best thermostat-controlled oil baths that can be obtained limit temperature variations to ± 0.001°C when placed in ideal surroundings of 20.0°C. However, if the laboratory temperature varies by 0.5°C, that of the oil changes by about 0.002°C. Hence, the oil bath does not provide thorough protection against problems in the air-conditioning; it divides the fault by roughly 200. Therefore, using thermostat-controlled oil baths does not mean you can avoid using airconditioning. – Time (lapse), which modifies many quantities; for example, the frequency of oscillators and the characteristics of electronic instruments, especially soon after they have been plugged in. Consequently, it is advisable to leave metrological instruments working uninterruptedly, or to plug them in the evening before using them the following morning.
About Measuring
213
There are many other influencing quantities: the hygrometric level of air, electric and magnetic fields, shocks and vibrations and, it must be added, the location of the instruments in the area, which act as an influencing parameter. A measuring instrument should reach the accuracy stated by its manufacturer after a period of stabilization, when it is motionless, sitting on a horizontal surface, at the rated temperature – often +20°C – plugged in on the 50Hz mains at precisely 220V, or if it is battery operated, when the batteries are new. What happened the accuracy when these conditions change? In other words, how do influencing quantities interact? And how does one become free of them? This is a difficult problem because of the frequent lack of information in technical notices. It can be solved by making a list of the influencing quantities, finding their effects in order to get rid of them or compensate for them, or even assess their effects. As an example let us take the case of metal gauges that dilate when the temperature rises. It will be necessary: – to assess the interference of the influencing temperature quantity: this may entail finding out about the alloy of the gauges in order to know their dilatation coefficient; – to get rid of the influencing quantity, which – still in the same case – will imply a reduction of the variations of temperature affecting the references and elements being controlled; – to proceed by compensation: this will be possible if the element to be measured has the same dilatation coefficient as the gauges. If element and gauge are kept at exactly the same temperature, this temperature will then be able to vary without the comparison being affected; – to undertake some calculations: more generally, if the element in question and the gauges have different dilatation coefficients, it will be possible to measure the temperature at which the comparing of length is done and calculate the error resulting from the gap between the temperatures.
9.2. Choice of a measuring principle Before you make an inventory of the criteria of choice to consider for a measuring instrument, you have to choose a principle to apply. There are three main measuring principles. Each has specific benefits and drawbacks. Therefore, it is essential to make inquiries before any purchase in order to know which principle has been chosen for the instrument that one is considering buying.
214
Metrology in Industry
9.2.1. Differential measurement Differential measurement consists of comparing the unknown object to another object of the same nature by means of a measuring bridge, a comparator or a differential instrument. The issue will come down to measuring a length with gauges or with the aid of a tight link, a weight with a set of masses, or a chronometric magnitude by comparing the time of the studied phenomenon to a reference clock, usually a quartz one. Differential measurement is above all else the metrological procedure: the comparing instrument and the references are identified separately, which makes connecting easier.
9.2.2. Direct measurement In the case of direct measurement, the user does not have to proceed to any assembly. The user no longer has to bring together a comparator and separate references; he uses an instrument that immediately gives a result: a caliper to measure a length, a multimeter for a difference of potential, a frequency meter for a frequency. Contrary to what it seems, this measurement is also a differential measurement because there is in the instrument a reference of the same nature as the measured magnitude: the caliper “refers” to its graduated body which, representing a ruler, the multimeter compares the unknown voltage to that of its Zener diode, the frequency meter compares the unknown frequency to that of its internal oscillator. However, reference and comparator make up a whole, which leads one to forget that it is a comparison that is being made. In some cases, the same instrument enables the user to choose between differential measuring and direct measuring. This is true of digital frequency meters if the user can choose between the inbuilt quartz and an external synchronization signal. In fact, a direct measurement is a “masked” differential measurement. Direct and differential measurements have the same principle but set out in two different ways: – differential setup is preferable for metrology laboratories; – all-in instruments (masked differential measurements) are better suited for industrial uses.
9.2.3. Indirect measurement Indirect measurement is altogether something different. The point is to replace the measurement of the unknown quantity by determining another quantity proportional to it. Who has not, when a student, measured a crazy-shaped surface by materializing it with cardboard or metal sheet and comparing the mass of the sheet to that of one square decimeter of the same material? All industrial thermometers
About Measuring
215
proceed indirectly: with liquid-, thermocouple-, resistance-, quartz-thermometers you determine, respectively, a length, a potential difference, an electric resistance or a frequency linked to the temperature by a one-to-one relation. Likewise, a precise measurement of mass makes it possible to measure the volume of a liquid or a number of identical objects. Here are other examples of indirect measurements. Measuring a length by determining a length of time An echometer sends a brief impulse in a cable; this signal is “reflected” either at the end of the cable if the cable is sound, or at a fault if there is one. The time the electric impulse takes to go there and back is proportional to the distance covered and indicates the length of the undamaged cable or the position of the fault. In a similar way, the distance, within under one meter, from the earth to the moon was established by echometry with an ultra-brief light impulse. Measuring the velocity of a fluid by determining a temperature As in the hot-wire anemometer, the wire through which an electric current passes gets cooler at a rate dependent on the speed of the air flowing around it. In this case, the temperature of the wire is identified by the indication of the electric resistance. In principle, it is, in effect, a temperature that is to be found, but actually it is an electric resistance that is eventually measured. Indirect measuring is useful; it is most frequently used to “replace” the physical quantity to be determined, by a frequency or an electric quantity that is easily measured even from a distance. However, you must always keep in mind that indirect measuring instruments, more than any others, require calibration. 9.3. Practicing in metrology The problem is to take measurements in a metrological context. It is not sufficient just to read or record a physical quantity from a suitable apparatus; in addition you must: – directly or indirectly connect the instruments you use for references; – list, check and criticize the working conditions; – calculate the effect of the influencing quantities; and – determine the uncertainty of the results. All these extra actions transform mere measuring into a “metrological action”.
216
Metrology in Industry
9.3.1. Implementing the instruments Once chosen, the instruments have to be implemented, meaning they have to be set up in a suitable place, away from vibrations, temperature variations and, generally speaking, protected from any disruptive “agent”. It is highly desirable to have a large table at one’s disposal with nothing on it but what is necessary, that is: – the elements to control; – the instrument or the measuring assemblies, separate references included; – the accessories: calculator, recorder, printer, etc.; – the instructions for use of the main measuring instrument; – a laboratory notebook.
9.3.2. Precautions before measuring The secret of metrology lies in the saying, “More haste, less speed”. Indeed, to avoid a series of measurements turning out to be useless, a many precautions must be taken before starting: – check the measuring assemblies; – give sufficient time for the stabilization of the elements to be measured; – use a guide list that you have drawn up for each type of operation; – keep a laboratory notebook and write down all the information about the operations: date, time, identification of the elements controlled, operations undertaken, results, temperature, etc. Regarding notices concerning the stabilization of measuring instruments: if there is a lack of indications in technical notices – which frequently occurs – tests will have to be carried out in order to determine how long they should work to obtain the nominal characteristics.
9.3.3. Measurements Taking measurements may take little time compared with the preparation. However, the metrological spirit urges one to repeat the measurements again and again and to practice self-verification. As an illustration, let us take the measuring of a mass Mx of roughly 103 grams. It will be wise to do five determinations one after the other: – a mass marked 100 grams; – the mass Mx;
About Measuring
217
– the same 100 gram mass, again; – the mass Mx a second time; – a third determination of the 100 gram mass. The three determinations of the 100 gram mass (reference) may possibly reveal a systematic error. If, on the contrary, the three weighings are repeated correctly, it will be a sign of exactness and it will be a plus in the evaluation of the uncertainty in the two determinations of Mx.
9.3.4. Variations and their sign To measure is to compare an unknown element to a reference, for example. The result of any comparison is a measurement made up of two elements: an absolute value and a sign. The value is given unambiguously by the instruments, but the sign of the variation is dependent on the assemblies, the connections or some commutations; this demands much care from the metrologist. What would be the use of determining: – the variation within a nanosecond of two clocks; – the difference within one-tenth of a micrometer between two gauges; – the defect of a right angle within one second of an arc; – if a mistake were made about which way the variation goes, that is about the sign of the difference? You have to be all the more careful as all measuring benches are not based on the same principle. For example, to control digital voltmeters there are: – sources of reference providing round values of, for example, tension; for a tension (source) of 10.000V a given voltmeter will display 10.003V; – sources of reference that have to be adjusted until the voltmeter displays a round value; the voltmeter of the example will show 10.000V when the source supplies 9.997V. The two results 10.003V and 9.997V, seemingly conflicting, represent the same flaw in the voltmeter. The raw result of the control (10.003V or 9.997V) only means something if one knows the principle of the measuring bench used and the method applied.
218
Metrology in Industry
9.3.5. The time factor When a measuring problem is tackled for the first time, it is not unusual for the preparation to last 20 to 30 times as long as the execution of the measurements. Preparing actually means studying the problem, choosing a method and some instruments, setting these up in a stable thermal surrounding, testing them and critically assessing the results. At the risk of making the time spent on this preparation even longer, it is advisable to write a procedure, especially if the operations are exceptional and only irregularly performed.
9.4. Expression of the results In metrology, a set of measuring is completed when the values that have been found have been written, printed, recorded or committed to memory. The expression of the results must always indicate the two following elements: – the designation of “the object” that has been measured: identification of the instrument, of the subset, of the sample; – the date, and in some cases the precise time of day of the measurings. As for the actual result, it must include the three parts indicated in section 7.2 of Chapter 7: – the numerical value; – the unit; – the uncertainty. 1st example: a blade is 1.072 mm thick within ± 0.005 mm There are three parts in this result: – the numerical value
1.072
– the unit symbol
mm
– the uncertainty
± 0.005 mm.
2nd example: a kilogram of steel has a mass M = 999.999875 g within ± 5 µg The three elements of the result are: – the numerical value
999.999875
– the unit
gram (or its symbol g)
– the uncertainty
± 5 µg.
About Measuring
219
The value of this measurement, with its many repetitions of the number 9, is not easy to read, so it will be expressed differently. The value of the mass is at -125µg (which implies “with regard to the nominal mass”). In this form, the result includes more than three elements: – the nominal value (1 kg), which is implied; – the algebraic value of the variation (with regard to the nominal value); – the uncertainty; – the units, for the nominal value, the variation and the uncertainty. 3rd example: the frequency of a quartz measured with an atomic oscillator (cesium) is: F = 4,999,999. 999985 hertz (symbol Hz) To avoid using a great many 9s or 0s, the frequencies of oscillators are most frequently expressed by their relative variation with regard to the reference. In this example: reference quartz
Fo = 5,000,000.000000 Hz F = 4,999,999.999985 Hz.
The relative variation of the frequency is: F − F0 ∆F −0.000 015 Hz = = F0 F0 5, 000,000 Hz
=
− 3.10−12 (no dimension number)
It is customary to say this oscillator is at - 3.10-12 which implies “from the reference”. These three examples highlight the following principle: You can express a metrological result: – either by a value (a number and the unit you use); – or by an absolute variation with regard to a reference (expressed with the same unit as that of the quantity you study); – or by a relative variation with regard to a reference (the quotient of two quantities of the same nature, therefore a no-dimension number). A variation, absolute or relative, is an algebraic quantity made up of a value and a sign.
220
Metrology in Industry
9.4.1. Graphs Whenever possible, numerical results will be supplemented with a graph, the great benefit of which is to bring out discrepancies when any occur. Consider, for example, this series of results which should have had the same value: 0.704899 0.704901 0.704898 0.704892 0.704900 0.704899 0.704899 These values are apparently close, but a diagram in a proper scale immediately shows that indeed the results form a cluster, except the fourth result. Value
0.704902 0.704901 0.704900 0.704899 0.704898 0.704897 0.704896 0.704895 0.704894 0.704893 0.704892 0.704891 1
2
3
4
5
6
7
Number of measurements
Figure 9.1. Number of measurements
9.4.2. Histograms A histogram is a graph which for each value found gives the number of times it has appeared (frequency). For the series of 7 measurements, taken as an example in section 9.4.1 above, the histogram is as follows.
About Measuring
221
Frequency
3
2
1
0 892
893
894
895
896
897
898
899
900
901 Results
Figure 9.2.
The group of the six results on the histogram goes without any comment and the isolated value stands out. Two notes about the terminology: – The word frequency is used in statistics and means number of times an event happens; it must not be confused with the frequency of a signal or of a phenomenon, which is the number of cycles per unit of time and which is expressed in hertz (symbol Hz). – A histogram is a “bar-chart” that provides the frequency (that is, statistically) according to measurement results. In a histogram, all the results are mixed up together in the same diagram and the order in which the values appeared is lost. So, in spite of the definitions of two words being similar, the histogram conceals the history of the results. Let us remember that the word histogram comes from the Greek histos (texture, web), whereas the word history comes from the Latin historia (history, story).
9.5. What qualities does a metrologist require? Whatever physical quantity he may be dealing with, a metrologist must reason and behave in a way “adapted” to accurate measuring. Therefore, he must have many qualities.
222
Metrology in Industry
9.5.1. Be inquisitive First and foremost, a metrologist has to be curious, and his curiosity must take many forms, and be about everything. A metrologist must make inquiries: – about the instruments he controls; – about the proceedings; – about the influential quantities. But that is not all: he must also keep himself regularly informed of his firm’s activities that have a direct influence on measuring problems, current and future. He must visit laboratories and meet other metrologists.
9.5.2. Be tidy and methodical Frequently, measuring means comparing an unknown object to a reference. These comparisons will be worthless if they are not always performed in exactly the same way. For some complex operations, it would be advisable to write detailed procedures and faithfully follow a guide list rather than rely on one’s memory or on instinctive habits.
9.5.3. Be open to doubt A good metrologist ought to question everything: references, comparators, proceedings. Doubt will urge him to, for example: – criticize the processes in order to improve them; – check that the references implemented were calibrated when they were supposed to be; – check the proceedings are correct: right temperature, stabilization, creating a vacuum for measurements of absolute pressure, etc. Doubting inevitably leads to repeat measurings, preferably with several instruments and, even better, applying other methods. It is best to aim for a repeatability and reproducibility of measurings.
9.5.4. Be observant A keen sense of observation will enable a metrologist to avoid many mishaps, for example: – by noticing that an assembly has to be modified;
About Measuring
223
– by noticing that a 127/220V tension switch must be reversed; – by finding out the temperature of an instrument is not normal; – by noticing that a standard gauge is scratched.
9.5.5. Be honest Being honest for a metrologist means: – leaving a blank in a result table every time a determination has not been worked out because of a lack of time or any other cause; – writing down all the results without making any change, even unexpected values; odd values can be of great interest because they usually lead to significant results: unstable instrument, effect of an influential magnitude or, more commonly, a confusion between two elements to be measured. This is a long list of qualities, but do not let that worry you. Tackling metrology is the fate of those who intensely love measuring. You could almost assert that one takes up metrology as one takes holy orders.
This page intentionally left blank
Chapter 10
Organization of Metrology at Solvay Research and Technology
10.1. Presentation of the company Solvay is an international pharmaceutical and chemical group headquartered in Brussels; it has subsidiaries and joint companies in 50 countries and employs some 31,000 people. In 2002, its consolidated turnover reached €7,900 million coming from four areas of activity: chemicals, plastics, transformation and pharmaceuticals. Organized in “strategic business units” and in “competence centers”, the group is deeply involved in a policy of total control of quality for the benefit of its clients. Solvay Research and Technology is the major research center of the group. Its research programs take in Solvay’s activities, minus the pharmaceuticals sector. The site is located at Neder-Over-Hembeek (Brussels) and it stretches over 23 hectares. Close to 1,100 people from the scientific and technical services of the Solvay Group work there.
Chapter written by José MONTES – Solvay/Belgium.
226
Metrology in Industry
10.2. Organization of the metrology sector 10.2.1. Creation The creation of a metrology sector in 1995 was the result of a 1994 survey concerning the organization of the firm in conformity with quality insurance. The main conclusions of the survey revealed the urgent need for some divisions to join a quality system (ISO 9000, GLP-GMP) and the necessity to create a metrological organization of the basic quantities (temperatures, pressures, mass flow of gases, masses, time, etc.). The mission assigned to the organization was that it should be a center of competences in which the means and experience of the site were integrated, and in which the consistency of the management of the metrological requirements was secured. The metrology sector was naturally integrated to the group in charge of the activities concerning the instruments and the automation on the site; some of its personnel who were technically competent were recruited.
10.2.2. Missions These following missions are assigned to the metrology sector: – to ensure the development and the management of the working standards and their connection to national standard; – to carry out the plan technical tasks of calibration; – to take charge of the computerized management of the periodical verification of the measuring means; – to centralize and keep up-to-date the data of the supply of measuring instruments which are periodically checked; – to keep documents (draft the calibration certificates, archive, keep records, etc.); – to draft the necessary general and measurement procedures; – to provide internal clients with advice and technical support.
10.2.3. Organization The organization of metrology is dictated by the company’s internal rules, in agreement with the ISO 9000 or GLP-GMP rules followed then by the rules of the internal clients. One of the major requirements of these rules is to control the checking, measuring and testing equipment; the greatest part of this control consists of periodical calibrations.
Organization of Metrology at Solvay Research and Technology
227
The metrology sector has organized itself in such a way as to provide a technical competence which is adaptable to the needs of the client and to offer an administrative organization which is as homogeneous as possible for all the internal clients of the site. When he decides to set up a quality system, the client uses these services in order to define and organize the calibration operations. Once started, the process goes through the main following stages: – inventory of the representative measuring equipment and analysis of the metrological constraints is undertaken with the client; – identification of the measuring equipment in agreement with the codification that has been adopted and, using labels, marking the measuring equipment; – introduction of the data and the specifications of the measuring equipment into the database; – checking whether the measuring equipment is suited to the needs specified by the clients; – calibration of the measuring equipment; drafting the documents and handing them over; – periodical follow-up of the measuring equipment.
10.2.4. Geographic localization of the activities Calibration activities are carried out either at the laboratory of metrology where the instruments are returned, or directly on-site. Calibrating on-site makes it possible to consider the measuring equipment in their environment; it also favors direct dialog with the client. The metrology laboratory has air-conditioned premises, the temperature of which is regulated and the hygrometry of which is under control. In it are most of the working equipment, the standards, the data-processing tools, the documents and the archives.
10.2.5. Composition of the bank of measuring equipment The bank of the measuring equipment, which is periodically attended to, continues to grow and in December 2003 numbered 4,020 units. The figure below represents how the categories of measurements are distributed.
228
Metrology in Industry
Dimension 1%
Pressure 30%
Mass 8%
Flow 6%
Others 15%
Speed 2% Level 2%
Temperature 49%
Others 4%
Figure 10.1. Solvay R & T Park – metrology distribution of the measurements
10.3. Metrology 10.3.1. Identification The measuring equipment must be identified one by one. The identification attributed to the equipment at the time of the manufacturing process has priority and is maintained. In cases where the manufacturer did not identify the equipment, identification is determined from an internal general convention of engineering based on the ISA (Instruments Society of America) norms. The basic principle of identification has two parts: – the functional identification generated by the type measurement (for example, TE, PI, FT); and – the identification related to the geography location (building and premises) as well as to the type of instrument (viscometer, oven, etc.). The latter part also mentions the general identification of the equipment, and thus a coherent link is ensured. Self-adhesive labels mark the measuring equipment and instruments. 10.3.2. Connection of the standards The measuring instruments or equipment are calibrated with the help of working standards. In their turn, the working standards are periodically calibrated by laboratories accredited by the OBE (Belgian Organization of Calibration), which is itself a member of the EAL (European Cooperation for Accreditation of Laboratories).
Organization of Metrology at Solvay Research and Technology
229
This procedure guarantees the traceability of the measuring instruments or equipment through their connections to the national standards. Some equipment, which is not worth investing in expensive standards, is simply calibrated by a competent accredited laboratory.
10.3.3. Periodicity of the calibrations The periodicity of the measuring equipment calibrations, which are metrologically dealt with, and of the working standards is entered into the database. The periodicity is defined according to the manufacturer’s specifications, our experience with the equipment, the environment in which it operates and whether the client makes intensive use of it or not. The periodicity can be reviewed, as a result of a particular cause, or as a consequence of the results of several calibration cycles. It will be lengthened if the results prove to be stable and always within the tolerance interval; conversely, it will be shortened if drifts or systematic excesses are observed. The data concerning the measuring equipment are recorded in a file located in a share zone of the firm’s local area network. All the clients can access and read the file. The chief benefit of this organization is the updating of the source in real time and the ability of the client to use his part of the file for his own internal management.
10.3.4. Calibration operations The calibration schedule is subordinate to the dates which are obtained by confronting the requested checking periodicity with the date of the last calibration. A sliding schedule is drawn up at the beginning of the week and is used as a base for planning the interventions. The schedule includes the list of the measuring equipment that is due for verification in the week, as well as that due in two and four weeks’ time. The work done amounts to more than just a calibration; it is a certificate of verification as it declares the instrument to be in a state of conformity or nonconformity.
230
Metrology in Industry
Details of the work done are given in the calibration certificate; there are several steps (or stages): – the results of the calibration before corrective maintenance (adjustment or repair); they concern the time since the last calibration and make it possible to verify the possible impact of a measurement drift on the process; – the comparison with the specifications (tolerances); – the results of the calibration after, perhaps, corrective maintenance; the results concern the period to come, beyond the date of the calibration; – the final comparison with the specifications; – the ruling about whether the measuring instrument that has been checked is metrologically in conformity with the specifications.
10.3.5. Documentation of the calibration results The documentation of the calibration results is made complex by the diversity of the measuring equipment found on-site. The results of the measurements are recorded in a document which is addressed to the client. The document, it is called a calibration certificate or metrological control if the testing was of a secondary piece of measuring equipment that it is not really possible to adjust (for example, a drying-oven temperature). The document for the benefit of internal clients is always the same regardless of who the client is or what type of measurement has been made. The constraints inherent to each category of measurements (pressure, temperature, flow, mass, etc.) have led to develop a more complete specific worksheets for specific categories of measurement. The copies of the calibration certificate and of the worksheet are archived in the metrology laboratory and make it possible to keep track of the measurements. When you select the measuring equipment’s identification, the chief identification data of the measuring equipment are automatically transferred from the file to the worksheet. The synthetic results are then automatically transferred from the worksheet to the calibration certificate. The successive operations stated on the certificate sum up as follows: – selection of the equipment to calibrate; – automatic input of the main identification data from the file to the worksheet; – encoding of the results and intermediary automatic calculations on the worksheet; – automatic input of the identification data and of the calibration results on the certificate.
Organization of Metrology at Solvay Research and Technology
231
In order to make the transcriptions of the information dependable, the database is automatically updated after the documents have been edited. 10.3.6. Verdict of the metrological confirmation Metrology is responsible for the quality of the measuring equipment it has verified and its role is to guide the client into establishing the overall conformity of his equipment. In the end, it is up to the client to make the decision about conformity, after supplementing the results of the calibration certificate(s) and any other tests undertaken. The comparison of the calibration results with the instructions about measuring equipment (tolerances) leads to two possible types of decisions to be decided: – if the deviation is within the interval of tolerance, the measuring equipment is said to be conformable and brought back into service; – if the deviation is outside the interval of tolerance, the measuring equipment is said to be not conformable, which means one of the three following solutions: - adjustment or repair; both interventions require a new compulsory calibration before the measuring equipment returns to service, - downgrading; there will be a less demanding new prescription adapted to the new use, - scrapping; the instrument is judged to be unsuitable to measuring, it is scrapped and some parts, intended for the repair of similar instruments, are salvaged. 10.3.7. Indication of the state of the calibrations When a measuring instrument has been calibrated, its state is indicated by a calibration label clearly visible on the instrument. The label guarantees that the measuring instruments has been verified and tells how accurate it is. If there is no label – the label having been lost or deliberately removed – it means the state of calibration has been not conformable. The label mentions: – the identification of the metrology sector; – the identification of the measuring equipment (according to the file); – the dates of the calibration and of the next calibration (based on the determined interval); – the reference of the calibration certificate; – the initials and signature of the performing operator.
232
Metrology in Industry
10.3.8. Personnel and subcontracting The personnel of the metrology function organize its interventions according to plans dependent on the list extracted from the database. It is autonomous in the performance of its tasks and the production of its documents. Handling the amount of work and spreading it over the year is done in agreement with the clients. Qualified subcontracting personnel are used to carry out part of the activities. They work according to the procedures and with the documents established by the metrology function. The metrology function is responsible and answerable for the quality of the performances of the subcontracting personnel.
Chapter 11
Metrology within the Scope of the ISO 9001 Standard
11.1. Introduction The control of measuring equipment is based on the following observations: – you cannot know what quality you have obtained if you cannot measure it; – you cannot make measurements if you do not have the proper equipment for it; – you cannot trust your equipment if you do not have them under control, etc. and that is the object of metrology. This binding link between metrology and quality was taken into account by the quality directors who took part in the drafting of the ISO standard of the 9000 series on the “management of quality”. From the beginning, a chapter (out of the 20 of the original standards) was devoted to this theme. Its drafting by quality directors somehow raised problems for its implementation by metrologists, mostly regarding the strictness of their technique, particularly the specificity of the vocabulary.
Chapter written by Philippe LANNEAU – Management Services, and Patrick REPOSEUR – Comité Français d’Accréditation (COFRAC).
234
Metrology in Industry
On the other hand, the official metrological structure, at the national as well as international levels, has to be coherent with the requirements of the system of reference. This has brought the Comité francais d’accréditation (COFRAC) together with the National Metrology Institute (BNM), with all the partners concerned and circulated under the double stamp of COFRAC and BNM.
11.2. Introduction to the evolution of the standard The third version of the “quality” ISO 9001 standard (December 2000) presents noticeable evolution in comparison with the previous versions. The new output is more user-friendly, easier to read, because it is written in a more “everyday”, less normative, French. The concepts themselves – the ideas – are accessible to most readers. Finally, the approach is more general, less manufacturing industry-oriented; the aim is no longer to give the clients “the assurance of quality”, it is to “manage the quality” on behalf of the firm. Naturally, the client gets something out of it, so everyone is satisfied. Metrology will be one of the elements to “manage”, as part of the organization which has been set up.
11.2.1. The concept of continuous improvement Continuous improvement symbolized by the “(PDCA) cycle” proposed by E. Deming is familiar to quality managers; it is the basis of the structure of the new system of reference. As a matter of fact, it is proposed split it up into four phases which come one after the other in a logical order with the purpose of improving the functioning of the existent organization. The control of the checking, measuring and testing equipment (section 7.6) is explicitly mentioned in the phase which describes “the realization of the product” (Chapter 7). It is not without reason that metrology is positioned as one of the elements integrated into the firm’s central process. The elements which are necessary to control the measurements are found in the phases called “monitoring and measurement of the processes” (section 8.2.3), “product” (section 8.2.4), and “control of the production” (section 7.5.1d).
Metrology within the Scope of the ISO 9001 Standard
235
11.2.2. The process approach The phase of the process of “realization of the product” also proposes an original approach that seeks to put the functioning of the firm on a line which goes from the client’s request, to the delivery of the product (or service!) to the client. It is the process (sometimes called “client – client” process) that is positioned crosswise in comparison with the firm’s vertical hierarchical organization. In this context, our approach to metrology is defined in the ISO 9001 standard as a control of the “measurement process”1, the client being the user of the result of the measurement.
Measurement request
Measurement process
Measurement result
Figure 11.1.
This approach has consequences in the area of the process which concerns metrologists who are no longer satisfied by simply having their measuring equipment calibrated and affixing the appropriate labels. Metrologists become involved very much earlier, at the time of choosing of the equipment which means at the time of implementation, in order to meet the needs of user of the measurement (the client of the process). However, first of all, the “policy of the control of the measurement” has to be defined at the firm’s highest level; this makes it possible to make a decision that is suitable for the kind of risk the management has decided to take, that is either: – moderate control of its measurements for a low cost, but a high risk of internal malfunctions or of clients’ complaints; or – an intensive control, which means a higher cost for a greater security. 1 The ISO 10012 standard provides the elements of an explanation.
236
Metrology in Industry
It is obvious that something between these two extreme options would be preferred. The decision will be made “in accordance with each different case”, by analyzing the case’s need in measurement, its impact on the control of the firm’s general process or of the quality of the products. Given its implications, it is unquestionably up to the firm’s management to make this decision about a risk of such a level. It is one of the “management processes”. The function which assumes the responsibility of the “measurement process” will then have to implement the policy of the control of measurement. This evolution of the standard encourages the firm “to take itself in hand” by defining objectives without going into details or fixing the means necessary to reach the objective.
11.3. Measurement control process Let us start with the schematic representation to be found in the ISO 9001 standard (section 0.2) to situate the process of measurement control.
Continua l improveme nt of the quality mana gement system
Management responsibility Customers
Customers
Measurement, analysis and improvement
Resource management
Requirements
Input
Product realization
Product
Satisfaction
Output
Value-adding activities Inform ation flow
Figure 11.2. Model of a process-based quality-management system
Metrology within the Scope of the ISO 9001 Standard
237
The contents of the five steps of the measurement process are described as follows. Step 1 – expression of the need for measurement This step comes from the “customer” of the measurement, from inside the firm (the design department or the process service), or outside (the buyer or the consumer who sets the specification in his schedule). The characteristics of the need will be: – the type of measurement; – the range in which the expected results are to be found; – the tolerance of the measurement. Step 2 – analysis of the need for measurement This second step corresponds to the metrological competence’s taking responsibility for the process. It also makes it possible to specify and make clear the need, in agreement with the client. The control of the whole process depends on the quality of this cooperation with the client. From this step it will be possible to give a correct answer to the problem regarding: – which technique to implement; – the corresponding fitting range; – the uncertainty that goes with it. Step 3 – setting up of the appropriate equipment (the response) From the elements defined in the previous step, this step makes it possible to set up the measuring equipment. This includes the supply (purchase, or looking for what is immediately available), the receipt, the assembly, and also the realization of the “administrative” part of the control of the equipment. This last part consists of identifying the equipment (marking it, for example) and opening a file or an identification sheet (one can get ideas from the FD X 07-018). Step 4 – traceability In our approach, the metrological follow-up corresponds to the traceability to the national references (the standards) and to the checking done within the firm. These latter checks make it possible to ensure that the equipment has an adequate calibration status. This makes it possible to create confidence in your exchanges with the clients, as the client’s and his supplier’s results are similar. It is necessary to make a periodic check of of the calibration status to be able to confirm that it is fit for use.
238
Metrology in Industry
Step 5 – availability This step comprises the work environment, the conditions of the implementation, the measurement procedures and the operator’s competence. Also included are the methods of protection while the material is used, stored or transported. The significant moments of the “life” of the equipment are to be recorded on the identification sheet mentioned in step 3. This makes it possible to complete the whole set of the measurement processes.
11.4. The ISO 9001 (2000) standard step-by-step This chapter addresses the different requirements of the ISO 9001 standard and provides point-by-point explanations and practical illustrations: Section 7 – Product realization 7.5 Production and service provision 7.5.1 Control of production and service provision d) the availability and use of monitoring and measuring devices
The requirement about measuring equipment is integrated into the chapter that is devoted to the “realization of the product”. It is about the availability and the implementation of the equipment, which are presented as one of the elements of the control of the realization of the products of the company. Available equipment means that the need has already been defined, both at the technical level and concerning the amount of equipment needed to carry out the measurements. The implementation implies that one knows and complies with the measurement procedures and/or the specific competence of the personnel. In addition, the environmental conditions of the measurement must be defined and the setting up of the appropriate means must be ensured. For example, this may mean premises where the temperature is controlled and where there are no vibrations, and where electromagnetic radiations, dust, dampness, etc. are kept away. This makes it possible to minimize the components of uncertainty or, when this is not convenient, to assess these components in order to take them into account when stating the result of the measurement. Section 7.6 – Control of monitoring and measuring devices
It is to be noticed that a specific paragraph of the ISO 9001 standard is devoted to the control of the measuring equipment, in the same way as it is for the other
Metrology within the Scope of the ISO 9001 Standard
239
requirements of the realization. This confirms the place of metrology in the management of quality and in the control of the product (or service). Section 7.6 (continued) – The organization shall determine the monitoring and measurement to be undertaken …
This requirement corresponds to the step where the need for measurement. It must be satisfied by the functions which are concerned with the result of the measurement is defined. The functions should assess their needs for measurement and have an objective knowledge of these needs; necessary competence has to be on hand to assess this. Section 7.6 (continued) … and the monitoring and measuring devices needed to provide evidence of conformity of product to determined requirements (see section 7.2.1) …
This step corresponds with the definition of the technical response that is to be set up, in relation to the equipment capable of meeting the need determined in the previous step. Into this notion of accuracy should be integrated the type of measurement, the fitting range for this measurement and the tolerance which goes with it. This last point is provided earlier, either by the ultimate client or by the person who has conceived the measured element (the measurand). The answer as regards equipment: – type of measurement; – available range; – uncertainty that goes with it; – periodicy of external calibration and/or verification useful to ensure SI traceability. The determination of the uncertainty about the measurement is one of the essential elements for the definition of the aptness of the measurement. This includes various parameters that associate the methods implemented and the personnel’s competence to the equipment which has been used. This point is amply developed elsewhere in this book. A real, specific competence is unquestionably necessary to see this step through successfully. It ought to be available inside the firm. Section 7.6 (continued) – The organization shall establish processes to ensure that monitoring and measurement can be carried out and are carried out in a manner that is consistent with the monitoring and measurement equipment.
240
Metrology in Industry
After the need for measurement and the relevant responses have been defined, this phase corresponds to the implementation of the tools in accordance with defined methods, which are the measurement processes themselves. This phase also broadens the notion of equipment to the notion of the process as a whole, which includes the measuring instrument as well as the personnel who operate it (and their competence), the methods, the environment, etc. We now return to the need to control measuring equipment and associated uncertainty. It is vital to say what methods are to be set up to realize the measurements, what competence the personnel who implement them must have and what environment conditions are required. Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall a) be calibrated or verified at specified intervals, or prior to use, against measurement standards traceable to international or national measurement standards; …
This requirement concerns the connection to the national traceability chains. In France, these connections are made under the aegis of the COFRAC2 whether the quantities concerned are physical or chemical. The setting up of the rules is described elsewhere in this book. It is the firm’s responsibility to make sure they are implemented and complied with. The evidence of the connection with the references (metrological traceability) has to be available at the level of the firm. Taking the uncertainties into account is a part of the fundamental elements of these connections, in accordance with whether it suits the need. More information can be found on the EA website: http://www.european– accrediation.org. Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall a) …; where no such standards exist, the basis used for calibration or verification shall be recorded.
This phase widens the notion of “standard” as it is generally used in the fields of physical measurement to the other fields of monitoring and measurement (for example, in chemistry). 2 Since 1989, there has been a multilateral agreement of recognition of the equivalence of the calibration certificates delivered by European calibration laboratories (www.europeanaccreditation.org). Since then, an identical agreement at global level has been reached (www.ilac.org).
Metrology within the Scope of the ISO 9001 Standard
241
One speaks of “references” or of “reference materials”. These standard references have to be evaluated to give a reference value. The methods of assessment use the classical statistical tools which make it possible to get as close as possible to the true value and determine the uncertainty around the assessment. (Some elements are given by the ISO guide 35 on this point.) It is the recording of this analysis which should be retained. Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall b) be adjusted or re-adjusted as necessary;
This phase comes after a verification that has concluded that a piece of equipment is beyond permissible error limits. It makes it possible to restore conformity to this equipment by using its fitting devices, if it is equipped with any. Metrologists make a distinction between: – fitting: bring an equipment “as close to zero as possible”, and – adjusting: fitting it by using only the devices that are at the user’s disposal. It is to be noted that after any fitting (and therefore any adjustment), a new calibration and a new verification must take place which will make it possible to confirm that the equipment can be used (and is back within “maximum permissible errors”). Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall c) be identified to enable the calibration status to be determined; …
Identification consists of providing the user with information about the extent to which the equipment can be used in relation to its suitability or its possible restrictions of use. For example, a multimeter is limited to one type of quantity (“use only on ohmmeter function”), or some ranges of a measurement (“use only between 100 V and 500 V”), or the verification of some values of “product” tolerance. The method of identification must be adapted to the context (environment) and to the users. The solution can be anything from a mere label (with the date of the limit of validity) to the supplying of the calibration certificate (or its copy). When the calibration status is being considered, the point is to determine the appropriateness of the equipment to be used and the degree of criticity which is associated with it. “Best before …” says the inscription printed on the pot of yoghurt; likewise a calibration value may still be used beyond the date that ends its
242
Metrology in Industry
effectiveness, but there is a risk that only the user can accept. He can decide whether to take the risk from the follow-up of the corrections made between two successive calibrations, what metrologists generally call the “drift”. Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall d) be safeguarded from adjustments that would invalidate the measurement result; …
In order to avoid undue adjustment of the equipment, whether initially or after verification, access to the devices which make it possible to make these adjustments should be limited to competent persons. The users or handlers (transfer, storing, etc.) should not be able to make adjustments, even by mistake. The instruments may therefore be equipped with blocking devices: “locks” (physical or computer) or physical protection (shutters, hatches, etc.) to prevent access or adjustment, or which can detect these: varnish, seals, etc. Section 7.6 (continued) – Where necessary to ensure valid results, measuring equipment shall e) be protected from damage and deterioration during handling, maintenance and storage.
Measuring equipment is generally fragile or at the least needs to be handled (during use, transfer, cleaning, etc.) with care in order to guarantee the preservation of their metrological qualities. Likewise, the storing conditions must take into account the restraints relative to the materials, the components, etc., which go into them. Consequently, the measuring equipment ought to be protected from extreme variations of temperature, from dust, shocks, humidity, light (e.g., UV), etc. Frequently, the most fragile instruments are delivered in packaging which protect them during transport, and, of course, storage. The instruments should be kept in these containers when they are not in use. Section 7.6 (continued) – In addition, the organization shall assess and record the validity of the previous measuring results when the equipment is found not to conform to requirements. The organization shall take appropriate action on the equipment and any product affected …
This requirement concerns metrology, as well as a “quality assurance” approach. Investigation of the consequences of a doubtful measurement result concerns metrology function and quality assurance function through on the one hand the measuring equipment and on the other hand the measurement of the product.
Metrology within the Scope of the ISO 9001 Standard
243
The metrologist makes use of his knowledge of the equipment and of the consequences of the registered deviation through asking the following questions: – Is the deviation significant in relation to the measurement and the use to be made of it? – What is the relation between the level of the measured non-conformity and the uncertainty on the method of measurement? – Does the deviation have an influence on the process regarding the accepted tolerances? This technical information has been passed on to the firm which, thus informed, makes a decision about the product that has been measured with the faulty equipment. This technical information is passed on to the firm which, thus informed, makes a decision about the product that has been measured with the faulty equipment: – recall of the doubtful products; – dispensation, with or without informing the user (external or internal client); – accepting products as they are, the deviation on the instrument having had no impact on the quality of the product. The equipment itself is subjected to specific action so that the fault does not occur again: – small verification intervals, which limits the consequences of non-conformity; – modification of the permissible error limits set on the measurement if relevant; – change of measurement method and/or of equipment. Section 7.6 (continued) – Record of the results of calibration and verification shall be maintained (see section 4.2.4).
There are two aspects of this requirement: – all the calibration and verification results have to prove that the operation has been performed; the information that the operations has to be done has to be kept available. With this objective in view, the requirements of “control of the recordings relative to quality” of section 4.2.4 are applicable; – it is particularly important to be in possession of the information on the initial state of the equipment before a calibration or any other intervention (adjustment, user adjustment, repair, etc.), so that the previous point may be applied efficaciously. Let us point out that the materialization of an action is not the only aim of a calibration certificate; it is also (and is chiefly) intended to apply the corrections necessary for the use of the measuring equipment.
244
Metrology in Industry Section 7.6 (continued) – When used in the monitoring and measurement of specified requirements, the ability of computer software to satisfy the intended application shall be confirmed. This shall be undertaken prior to initial use and reconfirmed as necessary.
More and more frequently, measuring equipment is connected with software which directly intervenes in the process of measurement. It stands to reason that the software, too, should be subjected to the same principles of control. Two levels have been defined by the standardization body: 1. You have to ensure that the software does not bias the final result provided by the equipment. The classical methods of validation of software that apply here are: – measurement in parallel with other software that is certified to be fit for the purpose; – non-automatic verification that the software is working correctly. 2. Periodically, or before each new use, a test should make it possible to ensure the software has not wandered. Such a verification is integrated in some software; it is called the “check sum”, and is invisible to the user and it makes the sum of the “0” or “1” of the program in binary. There is only one result and it is characteristic of the program; it is its “genetic fingerprint”. Any modification of only one of these bits results in a different sum and the user is alerted. If this verification is not integrated into the software (find out from the supplier or manufacturer), it is possible to have it installed later. Section 7.6 (continued) – Note: see ISO 10012
This note allows the possibility of using the ISO 10012 standard ‘Measuring equipment – meteorological confirmation’. The detail of the technical answers to be implemented has been partly transferred to the ISO 10012 standard. As this point is given in a “note”, it is not compulsory to put the recommendations of these standards into practice. Nevertheless, they should be known and complied with. The ISO 10012 recommendations can be completed by reading the norms NF X 07010, 07011, 07015 and 07017, among other metrological norms which were drafted by French experts in the field. Further technical norms, particularly about the determination of measurement uncertainties, are to be found in the bibliography of this book. Section 8 – Measurement, analysis and improvement Section 8.2 – Monitoring and measurement Section 8.2.3 – Monitoring and measurement of processes
Metrology within the Scope of the ISO 9001 Standard
245
The organization shall apply suitable methods for monitoring and, where applicable, measurement of the quality management system processes. These methods shall demonstrate the ability of the processes to achieve planned results … Controlling the progress of the processes may require the implementation of measuring equipment. This is especially the case for production processes (the “proceedings”). After it has determined the critical points of the manufacturing process, the firm must define the corresponding checks and set them up. The associated equipment is then within the competence of the metrological control mentioned in section 7.6 and already analyzed. Section 8.2.4 – Monitoring and measurement of product The organization shall monitor and measure the characteristics of the product to verify that product requirements have been met.
This section corresponds to section 7.5.1d), previously discussed. It goes back, with a greater precision with more details, to the need for measurement, and it replaces the need into its context of surveillance of the products. It is the section that connects the control of the process of measurement to the need for measurement itself.
11.5. Conclusion Putting the answers which have been proposed in this chapter into concrete form makes it possible to satisfy the requirements of an audit of certification which relate to the control of the processes of measurement. But beyond strict answers to the questions of an auditor, controlling measuring equipment is, first of all, a means of progress for a firm, which is then certain of optimizing its measurements and the cost of its metrology; at the same time it generates greater dependence on and a greater trust in the relationship with partners, either clients or suppliers.
This page intentionally left blank
Chapter 12
Training for the Metrology Professions in France
12.1. The metrology function in a firm’s strategy Metrology training at education’s higher level is provided by a few organizations in France1. If firms are short of specialized metrologists, it is obvious that metrology still remains a mystery in higher education curricula. It is often difficult to know what comprises the metrological activity of a firm. It is generally limiting and metrology is often understood as management of the measuring equipment or laboratory activities. It has, in fact, a vast field of applications. Metrology, the science of measuring, as the dictionary defines it, is therefore an activity which should enable the user to give meaning and reliance to the stated measurement results. It does supply the necessary tools to claim the conformity of a product while controlling the risks.
Chapter written by Bernard LARQUIER – BEA Métrologie. 1 Secondary education students in France have to pass an exam called baccalauréat -Bac- at the end of the cycle to acceed to higher education. Bac+ 2, + 3, etc., indicates the level reached in higher education (how many years after the Bac are normally required to reach that level). Bac+ 2 is the level required to enter a school for engineers; it is also the level necessary (and sufficient) to be a higher-level technician. CAP (Certificat d’Aptitude), BEP (Brevet d’Enseignement Professionnel) and Bac Pro are lower-level exams opening straight into professional life.
248
Metrology in Industry
It is clear then that controlling measuring amounts to controlling processes which may be complex. Managing a firm’s metrology function requires a competence which reaches far beyond merely managing measuring instruments or knowing about calibration techniques. All the industrial sectors are concerned: mechanics, electronics, agribusiness, chemistry, pharmacy, medicine, environment, biology, aeronautics, space, nuclear power, agriculture, etc. The evolution of the norms relating to the control of quality systems in firms leads one to ponder over the growing influence of the metrology function. How, indeed, can one give meaning to a survey of clients’ satisfaction, or to an investigation of performance, without looking into the influence factors which affect the results and, therefore, into the uncertainties of measurement; they are an important aspect of the competence expected from the person in charge of the metrology function of a firm. It is logical to think that the position of the metrology function, in the fullest sense of the word, will be strategic for the management of firms in the years to come. To reach this objective the metrologists will have to have much broadened competences, far beyond the mere technical aspect.
12.2. Metrology profession Long-lasting specialized training courses in the field of metrology are most often provided at the higher education level and they generally lead to management jobs. The different professional categories (engineers, technicians, operators) consequently get very different training. If engineers and technicians have been able to benefit by specialized training courses, operators have entered the metrology function thanks only to brief training courses within the framework of continuing education. The synthesis tables show which long-lasting courses are currently available in France. The set of organizations given do not provide an exhaustive list of the establishments likely to offer long-lasting training courses in metrology. The general-education universities are beginning to offer supplementary training in the field of measurement. It is therefore probable that the list to be found here will be greatly extended in the years to come.
Training for the Metrology Professions in France
249
12.2.1. Metrological engineer Having received a higher scientific education, the metrological engineer is in charge of the metrology department; his mission is to implement all the actions that are needed to optimize the metrology function. He usually manages a team of technicians and operators. He may also intervene in the phase of conception of methods and manufacture. The metrological engineer is seldom employed by a small- or medium-sized business in which the metrology department is often limited to one or two persons from whom a broad polyvalency is generally expected. So it is in large companies, or in organizations specializing in measurement, that he holds his position. In laboratories or technical centers, his role is to control the measurement techniques and their traceability. He is responsible for the laboratory or the accreditation of the organization. He can go on studying to obtain a doctorate in metrology. He then becomes an expert in some field and generally works as a researcher in a top-level laboratory. He may, among other missions, have to see to the improvement of the national standards, or the development or settlement of calibration methods.
12.2.2. Metrological technician Initially, the metrological technician is trained as a higher-level technician, meaning that two years after completing secondary education – Bac+ 2 – he has passed a DUT (university diploma of technology) or a BTS (higher level technician diploma). Wishing to specialize, he has added to this qualification by spending an extra year in one of the organizations that provide specific training courses. These specific courses enable the technician to have a broad knowledge in metrology, as well as a good basic understanding of the domain of quality. He is then in the position to be in charge of the metrology function in a small- or mediumsized firm. He may also be called upon to manage the quality section and it is not unusual for him to have to manage both quality and metrology. He is able to head a team of operators, but he can also implement specific measurement processes. He can, after a few years’ experience, aspire to take charge of an accredited laboratory. He has the necessary competence to determine the uncertainties of the measuring processes and initiate actions to optimize the metrology function. If the organization chart of the firm includes a metrological engineer, the technician assists him.
250
Metrology in Industry
12.2.3. Metrological operator In general, a metrological operator has not had any specific training in metrology. Frequently, he becomes a metrological operator through advancement inside his firm. His initial training at the vocational-training certificate level (CAP, BEP or Bac) is supplemented by short, specific training courses in metrology. These courses, which are not discussed in this chapter, are dispensed by numerous organizations and by most accredited laboratories. Their curricula are general to prepare to the metrological trades, or specific to the command of a particular quantity. The metrological operator works along procedures and measurement methods established by an engineer or a technician. On his first job a technician can serve as an operator.
12.3. Initial training Metrology is very seldom taught in level IV and V (CAP, BEP, Bac) of initial education. It is found mostly at a post-secondary education level and it delivers engineer, technician or specialized operator diplomas. It is a pity that there is no specific training for metrological operators at the level of secondary education because firms frequently bemoan the lack of training of their operators. The firms find it necessary to resort to short (less than three weeks) continuing education sessions, which are generally too brief to master the different aspects of measurement.
12.3.1. Schools for engineers Schools for engineers turn out metrological engineers after five or six years of training (Bac+ 5 or Bac+ 6). In France, the most comprehensive courses at this level are provided by the Ecole Supérieure de Métrologie at Ecole des Mines of Douai, the Conservatoire National des Arts et Métiers (CNAM) and the Ecole Nationale Supérieure des Ingénieurs du Mans. Depending on the school, the training is provided either the traditional way, or through continuing education. In the latter case it is a supplementary or an alternate course.
Training for the Metrology Professions in France
251
The characteristics of these different schools are presented in the tables below. The particularities of the CNAM’s course are worth noticing: it offers working people the possibility of upgrading their training by attending evening classes; so do those of the Ecole Supérieure de Métrologie which has an international vocation and attracts many foreign students.
12.3.2. Courses for higher level technicians Higher level education in two years, Bac+ 2 (DUT or BTS), does not have any specific module for metrology. It is dealt with, more in a way to make students sensitive to it than as a specialized field of study. This has induced some lycées or university institutes to open supplementary courses (one extra year, Bac+ 3), equivalent to a professional degree or to a metrological technician diploma. These courses are open at the Lycée Jules Richard, the IUT of Aix en Provence, the University of Provence, the University of Toulon and the Var. They are chiefly intended for young holders of diplomas who wish to go on with their initial training, but they can be open in some cases to people who already have professional experience. Except for the Lycée Jules Richard, mechanical topics holds first place in these training units.
12.3.3. Vocational high schools There are no high schools that specifically train metrologists at the end of secondary-education level. Metrology is on the syllabus of some of them, but in a very limited way. In general, the main concern of the course is the control of dimensional checking and metrology.
12.4. Continuing education It has been said in Chapter 3 that some organizations that offer training courses in the context of initial training also give some candidates the opportunity to enroll for continuing education, personal training time-off and qualification contract.
252
Metrology in Industry
Applicants have to be under 26 years old to benefit from a qualification contract that makes it possible to receive remuneration and which subsidizes the firm in relation to the training costs. The personal training time-off can be used by employees who have been working for their firm for several years. It makes it possible, if the organization that gives the personal training time-off money agrees, to remunerate the candidate, completely or partially, while he is away from his firm and to pay, completely or partially, for the training costs. The continuing-education courses offered by the training organizations are of two types: they are either long-lasting (over 8 months) or short, from one day to a few weeks. The long-lasting courses train technicians, either as metrological operators (CETIM - AFPI de la Vallée de l’Oise), or as higher-level technicians at Bac+ 3 level (Bordeaux ENSAM which trains quality metrologists). New courses are being established: the CNAM has set up a program called “metrologist for the year 2000” and Bordeaux’s ENSAM offers a flexible course intended for the heads of metrology functions. The training course of the CETIM – AFPI Vallée de l’Oise is meant for candidates at Bac level. Its position is such that it complements the different diplomas and qualifications identified in France. It enables small- and medium-sized businesses to depend on personnel that are versatile in dimensional metrology and metrology function. Larger firms can rely on candidates with good basic knowledge to specialize, if necessary, in production control or laboratory metrology. The “Quality Metrologist” course at the ENSAM trains versatile technicians who are capable of setting up a metrology function and managing it in a small- or medium-sized firm, but also of assuming the care of the quality section in a smallor medium-sized firm. Short courses are provided by a number of organizations, among which are: – the laboratories accredited by the COFRAC; – the technical centers; – the training centers in large companies; – the schools for engineers;
Training for the Metrology Professions in France
253
– the adult training organizations; – private specialized companies. Very diverse subjects, general or specialized, are taught. The different organizations come up with a catalogue of inter-enterprise training courses. However, they are also in a position to organize specific courses according to the specific needs of a company. In this case, only one company is responsible for the training. The choice of the organization is made along several criteria: its reputation in the selected subject, how long the course lasts, where it will be, how much it will cost, what teaching methods are used, how much theory and how much practice (it is important that there should be a practical side as it helps the students to grasp the theoretical concepts), the level of knowledge required to attend the course, and the coursework to be submitted. Specific training courses are experiencing a boom; they make it possible to aim at precise objectives. A large enough number of trainees are necessary to enable a company to amortize the cost of the course more easily. The very small firms find it difficult to have their personnel trained because the size of their staff is not large enough to make up for the absence of those people who are away training. The development of training via the internet may become one solution.
12.5. Long-lasting training courses The information that appears in the following tables has been obtained from well-known organizations. A “training” group of the French College of Metrology has played a large part in the collection of the information. It is likely that, as this chapter is being written, general or specialized training courses about well-defined metrological aspects are being established. It is also likely that some organizations that provide long-lasting courses have not been identified. The information to be found in the tables in this chapter does not pretend to be exhaustive it needs expansion. The below-mentioned courses last more than eight months. They are opened to very different education levels (from the Bac level to that of engineer).
254
Metrology in Industry CNAM (PARIS) 292 rue Saint Martin - 75003 PARIS
Title
Engineer in measuring instruments
Year of setting up
unknown
Number of trainees
Training available in the whole of France
Level at admission
In 1st year: Bac+ 2 In 2nd year: after probationary cycle of the measuring instruments course
Duration of course
3 years 1st year: 360 hours (part-time) 2nd year: 280 hours (part-time) 3rd year: 2,028 hours (full-time) Economic and social management and communication: 240 hours (evening classes) The training is done in theoretical and practical modules outside working hours
Level at end Bac+ 5 of course Financing
Firm training scheme, or personal training time-off
Nature
Engineer
Main items of the program
Measurement and instruments: physical principles of sensors, properties of the instruments and acquisition of the signals Electrical and optical measurements Options in the 2nd year: industrial checking (ground networks, supervision, sensors and operators), quality metrology (signal, noise, quality, experiment plans, metrology), optics (images, optical measurements) Measurements and traceability Laser measurements, dimensional measurements Measurements of temperature and radiation Control of discrete event systems Management and economy of the firm Human and social management Communication, culture, expression Knowledge of professional English
Contact
Mr Himbert (33 1 40 27 27 73)
Notes
There are two stages in the training course: the probationary cycle (1st year) and the deepening cycle, outside working hours. To defend the thesis and obtain the diploma of engineer in instrumentsmeasurement, you have to be at least 24; you have to obtain all the scientific and technical modules and the management and communication modules; you have to take the (BULAT) test (technical English) and also meet the required conditions of professional experience (three years’ experience, two of which are in the specialty)
Training for the Metrology Professions in France
255
CNAM (PARIS) 292 rue Saint Martin - 75003 PARIS Title
Training of metrologist for the year 2000
Year of setting up
2002
Number of trainees
Maximum 12
Level at admission
Bac+ 2, or having worked for 3 to 5 years in a laboratory
Duration of course
10 months: one 30-hour-week per month, hence a total of 300 hours
Level at end Bac+ 3 of course Financing
Firm training scheme, or personal training time-off
Nature
Training of head of metrology; should award a certificate in the short-term
Main items of the program
Metrology function: organizational responsibilities and securing conformity with the quality systems of reference applicable to the firm (ISO 9001, ISO CEI 17025, ISO 10 012, NF X 07 010) Securing conformity and keeping the firm in conformity Metrology measurement expert: calculation of uncertainty, capability, method of supervision Training the firm’s personnel in metrology and setting up training programs Being able to organize normative watchfulness and watchfulness over the techniques of measurement Analysis of the value Management of the measuring means and of the personnel of a laboratory; qualification of the personnel: responsibilities linked to the internal management of the sets of measuring instruments, to the management of a set in the case of total or partial subcontracting; technical responsibilities connected with the operations of calibration and verification Definition of the methods and procedures of calibration, connection of standards Choices, technical negotiations, audit and follow up of subcontractors
Contact
Mr Himbert (33 1 40 27 27 73)
Notes
The training should lead to appointments as heads of the metrology function. In a small-or medium-sized firm the job may require its holder to head both quality and metrology
256
Metrology in Industry
CETIM AFPI Vallée de l’Oise Title
Controller in dimensional metrology
Year of setting up
1997
Number of trainees
8 to 16
Level at admission
Bac Pro. Employed person with recognized level of Bac Pro
Duration of course
10 months (4 months (450 hours) in training centre, 6 months in a firm). The trainees spend1 week at the centre and 2 weeks in the firm alternately
Level at end Bac+ 1 of course Financing
Continuing education Qualification contract Training time capital Personal training time-off
Nature
Training of controller in dimensional metrology with attribution of diploma MQ 97 04 60 0158
Main items of the program
Training centered on dimensional checking and metrology Metrology: vocabulary and generalities Concepts of quality and checking for quality Definition and setting up of procedures Measurement instruments and techniques Verification of the tolerances of products Applications of statistics, processing of the results, establishing uncertainties National calibration chain National standardization and ISO texts Analysis of needs in metrology Management of a set of measuring instruments Practice of measurements: influence quantities Rules about the setting up and the operation of a metrology laboratory
Contact
Mr Gabriel – CETIM (33 3 44 67 33 59) Mr Jacquemain – AFPI Vallée de l’Oise (33 3 44 63 81 63)
Notes
Great demand from industry Finding a job is easy after the course. Over 50% of the trainees are hired by the firm where they have been trained This qualification should be widened in 2002, by the establishment of a less demanding course, for firms’ personnel with an experience in dimensional checking and metrology
Training for the Metrology Professions in France
257
Higher School of Metrology (Ecole des Mines de Douai) 941 rue Charles Bourseul BP 838 - 59508 DOUAI Cedex Title
Metrological engineer or specialized master degree
Year of setting up
1929
Number of trainees
15 to 20
Level at admission
Hold a scientific diploma of Bac+ 5 or Bac+ 4 level and have professional experience International recruiting
Duration of course
1 year 7 months at the school (700 hours of lectures, practical work and supervised practical work) 4 months or more of training in a firm Modulated over several years for firm executives
Level at end Bac+ 6 of course The training is finalized by an engineer diploma or by a specialized masters degree accredited by the Conference of Higher Schools Nature
Engineer or masters degree This training is based on 4 main concepts: innovation (contribution from research laboratories), excellence, practice, international (teaching in French and in English) General metrology Sensors and signals, software engineering Data processing Working safety and legislation Quality and project management
Main items of the program
Metrology of different physical quantities, particularly: Electricity, magnetism Dimensional metrology Mass, force, pressure Volume, flowmetry Time/frequency Acoustics Ionizing radiations Physio-chemical tests
Contact
Mr Senelaer (33 3 27 71 23 24) or Mrs Cordelle (33 3 27 71 22 22)
Notes
Open to all holders of positions involving responsibility in metrology and who are able to: Understand metrology as a full-blown discipline Integrate the metrological component into the conception of products Conceive and implement measuring systems International character of the course, only training course of Bac+ 6 level
258
Metrology in Industry
ENSAM Bordeaux Esplanade des Arts et Métiers 33400 TALENCE Title
Training of personnel in charge of metrology
Year of setting up
2002
Number of trainees
6 to 12
Level at admission
Member of personnel in charge of metrology with Bac+ 2 level or with 10 years experience
Duration of course
248 hours over 9 months at the rate of 3 to 4 days every 3 weeks Assistance for a firm’s project possible (10 half-days)
Level at end of course
Bac+ 3
Financing
Provided by firms. The course consists of 3 modules which can be separated: Metrology function (84 hours) Uncertainties and optimization (80 hours) Quality-audit training (84 hours)
Nature
There is no provision at present for a diploma to be delivered at the end of the course
Main items of the program
Organization of the metrology function Expression of the metrological requirement and drafting of specifications Management of the measuring equipment Organization of a checking and calibration laboratory Determination of uncertainties and optimization of the metrology function Statistical Process Control (CMM) Quality process Setting up of self-checking and its management Training of personnel Audit of the metrology function
Contact
Mr Le Roux (33 5 56 84 53 21) Mr Larquier (33 5 56 34 20 63)
Notes
This course, based on the principle of alternation, enables some people in charge of metrology to increase their knowledge with a possibility of choosing modules. It also offers the trainee the opportunity to be assisted in the accomplishment of a specific mission in his firm
Training for the Metrology Professions in France
259
ENSAM Bordeaux Esplanade des Arts et Métiers 33400 TALENCE Title
Training of metrologists in charge of quality
Year of setting up
1997
Number of trainees
11 to 20
Level at admission
Bac+ 2 post-diploma, or job-seeker or working person with acknowledged Bac+ 2 level
Duration of course
10 months (4 months (470 hours) in a laboratory, 6 months in a firm) The trainees do the 4 months in a laboratory, then the mission in a firm Significant assistance in the firm is provided (4 to 6 visits of about half a day) Moreover, the trainee can get in touch with his professional tutor at any time to obtain advice about accomplishing his mission
Level at end of course
Bac+ 3
Financing
Contribution of the Ministry of Industry to help make the small- and medium-sized firms responsive to metrology A contribution is requested from firms
Nature
Diploma at the end of the course
Main items of the program
Setting up of the metrology function Stimulation of awareness of different quantities: dimensional metrology, electricity, mass, accelerometry, pressure, temperature, chemical metrology, etc. Dimensional and three-dimensional checking, geometrical permissibility Checking of machine tools and other checkings connected with mechanical manufacturing Drafting of procedures, realization of audits Determination of uncertainties of measurement and use of the Statistic Process Control Production management and self-checking in production Communication
Contact
Mr Le Roux (33 5 56 84 53 21) Mr Larquier (33 5 56 34 20 63)
Notes
There are many prospects In charge of the metrology function in small- or medium-sized firms In charge of quality: client or supplier In charge of quality in production In charge of a laboratory
260
Metrology in Industry
ENSIM (LE MANS) University of Maine Title
ENSIM engineer in industrial measurements
Year of setting up
1995
Number of trainees
60 to 70
Level at admission
2-year post-Bac classes, plus success at competitive exam, or Bac+ 2, plus school records. Masters degree-holders in 2nd year
Duration of course
3 years, including: 800 hours of practical work 300 hours of lectures and industrial projects 6 to 10 months of training
Level at end of course
Bac+ 5
Nature
Engineer diploma authorized since 1995 by the commission of engineer titles
Main items of the program
General education in industrial instruments and measurements. General education in physics, chemistry, electronics, data-processing, management; technological training in engineering, electronics, signal processing and automatics Measurements and sensors: organization of metrology, measurements of temperature, pressure, flow, velocity, viscosity, volumic mass, force, weight, acceleration, length, hygrometry, optics, colorimetry, polarimetry Non-destructive control Calculation of the uncertainties when using the different types of sensors Use of experiment plans, management of quality Organization of firms, techniques of job seeking
Contact
Mr Breteau (33 2 43 83 39 51)
Notes
The fields open to the trainees are those of research (integration of sensors, vibratory analysis, digital modelization, etc.), production (manufacturing processes, security systems, etc.), quality control (metrology, nondestructive control, acoustic and vibratory control, etc.), and environment At the end of the course, the trainees can prepare one of the DEAs which are on the curriculum at the University of Maine (acoustics, engineering, materials, user-machine interaction) There are many opportunities for jobs and all the engineers find a job within months of leaving the school
Training for the Metrology Professions in France
261
University of Provence – University of Aix-Marseille Title
Instrument metrology professional degree
Year of setting up
unknown
Number of trainees
unknown
Level at admission
Technical Bac+ 2, or equivalent Open to working people as part of continuing education
Duration of course
600 hours, followed by 12 weeks’ practical training Spread out over one school year
Level at end of course
Bac+ 3
Financing
Public financing
Nature
Professional degree
Main items of the program
Adaptation modules Theoretical metrology Applied metrology Methods English 120-hour tutored project 12-week training course in industry: its purpose is to materialize the acquired knowledge in the context of professional practice
Contact
Mr Bois ( 33 4 91 10 62 05)
Notes
The aim of the “instrument metrology” professional degree is to train some foremen and higher-level technicians for the metrology function of firms to be capable of implementing, in a statutory and lawful industrial setting, technical and methodological abilities about instruments, measurements, calculation of uncertainties, detection of sources of uncertainty, automatisms and tests The intended prospects are: Being in charge of research or business in checking, measurement and instruments Being in charge of the metrology/quality services Designer of measuring equipment Being in charge of a quality metrology mission Being in charge of maintaining process instruments
262
Metrology in Industry
IUT of Aix en Provence – University of the Mediterranean Title
Professional degree in industrial production sciences with optional industrial checking, dimensional metrology, production quality Other option: simultaneous engineering
Year of setting up
2000
Number of trainees
8
Level at admission
Bac+ 2, or equivalent Open to working people as part of continuing education
Duration of course
600 hours followed by 12 weeks’ practical training (150-hour foundation course, 300 hours of profession-oriented options, 150 hours for synthesis project and application) Spread over one school year
Level at end of course
Bac+ 3
Financing
Public financing
Nature
Professional degree
Main items of the program
Completing a project: management, control of the project, methodological tools (AMDEC, experience plans, etc.) General training: communication, technical English, economy and growth of the firm, labor laws Computer and mathematical tools and methods, fundamental functions of industrial CAD systems Statistics Metrology: Qualification of a measurement, setting up a checking at the surface plate based inspection, metrology of surfaces, metrology of great lengths ISO permissibility, setting up a checking on a coordinate measuring machine (CMM), non-destructive checkings, non-dimensional industrial measurement Tutored project: it is the materialization, by one individual or a team, of an industrial subject, or a technology transfer subject 12 weeks’ practical training during which trainees must assume responsibilities
Contact
33 4 42 93 90 82
Notes
The intended openings are: Being in charge of the metrology department Designer of checking, measuring and testing equipment Being in charge of a quality metrology mission Coordinator of research, methods, checking unit
Training for the Metrology Professions in France
263
Lycée Jules Richard (PARIS) Title
Training of metrological technician (qualification accepted by the National Joint Commission for Employment in the Metallurgical Industry)
Year of setting up
1995
Number of trainees
12 to 20
Level at admission
To have passed a DUT or a BTS
Duration of course
1 year; alternately 600 hours’ training in Paris/the rest in the firm. The trainee is paid by the firm which employs him under a qualification contract
Level at end of course
Bac+ 3
Financing
By the firm and an approved collecting joint organization
Nature
Qualification of metrological technicians
Main items of the program
Mathematics: calculation of uncertainties, statistics, matrix calculus, parameter curves, integral calculus, probabilities, complex numbers Electrical measurements: definition and calculation of the mean values which are effective For variable currents, ammeter, voltmeter, measurement of power, of the resistance of a resistor, of impedance, of time and frequency, magnetic and electronic measurements Dimensional measurements: measurements of lengths, linear and angular measurements; measurement of flow, acceleration, masses, force, material resistance, volumic mass, pressure Other physical measurements: temperatures, principles used in measuring temperatures, measurements of humidity, acoustics, vibrations, light Scientific and legal metrology: national and international official organizations, certifying organizations, definition of physical quantities, vocabulary of metrology, measurements, results of measurements and connected uncertainties Quality assurance and communication: standards, metrology function, quality system, audit, tools and methods for total quality, written and oral communication, French and English Technical vocabulary of the metrological technician (French, English)
Contact
Mr Desbordes (33 1 53 72 83 60)
Notes
Openings: In charge of the metrological service Assistant of person in charge of quality assurance Laboratory technician The trainees do not have any problem finding jobs; what is a problem is finding 20 candidates for the course
264
Metrology in Industry
University of Toulon and Var BP 132 83957 LA GARDE Cedex Title
Training of metrologist in charge of quality
Year of setting up
1992
Number of trainees
10 to 14
Level at admission Duration of course
Bac+ 2 (DUT GMP, OGP, Physical Measurements, Production engineering BTS) or people of a like level recognized by validation of professional experience 520 hours + project and training in a firm, all spread over one year
Level at end of Bac+ 3 course Financing
By the trainee
Nature
University degree of metrologist in charge of quality
Main items of the program
Applied mathematics and physics Characterization of materials Non-destructive checking Scientific and legal metrology Surface plate based inspection Calibration of measuring instruments Measuring machines Dimensions, reading of plan and CAD (design and drawing) English Communication Office automation Applied statistics Quality Reliability
Contact
33 4 94 14 21 77
Notes
The jobs offered come from all the types of firms, large companies, small- or medium-sized businesses or industries, laboratories. They are generally posts that involve responsibilities such as: Person in charge of the metrology service Designer of checking, measuring and testing equipment Person in charge of a quality metrology mission
Training for the Metrology Professions in France
265
12.6. The teaching of metrology in secondary schools The training courses specific to metrology are justified by the deficiencies in the traditional school system. The teaching of basic notions of measurement control has practically disappeared from the initial school years. The curriculum does not draw enough attention to the importance of measure in daily life and to the problems which arise when measurement should be controlled correctly. The user-friendliness of data-processing means has dimmed the notions of observation, of meaning of significant numbers, of doubting which goes with any measurement result. Initiatives from the French College of Metrology and the METRODIFF association to arouse awareness at different levels, particularly in secondary schools, have revealed pupils’ interest in metrology. It seems important to promote such initiatives until metrology is integrated into school programs.
12.7. Prospects for the development of long-lasting training courses It seems obvious that firms have a need for specialists in the sectors of measurement at a time when they are determined to reach absolute faultlessness, uppermost satisfaction from clients and the highest profitability. The hardest part for training organizations is to find candidates for these jobs, as students are poorly informed about them and the image of metrology professions is still austere. Probably the appeal can be emphasized today; the big companies have to act as catalysts to make the authorities, the Education Secretary, the agencies for the employment of managerial and non-managerial staff conscious of the risks that can be generated by badly-controlled measurements. Now, when the principle of precaution is called to mind, and when environmental, food, chemical, medical measuring grow more and more extensive, it is important not to make measurements any way and to remember that the measurement is not imputable to the instrument, but it is the outcome of a whole process in which the leading parts are played by the operator, the methodist, and the expert in metrology. The need for this collective awareness is essential so as not to run the risk of making irreparable errors.
266
Metrology in Industry
12.8. Bibliography Documentation from the different organizations referred to P. Souquet, S Gabriel and D Jacquemain, Qualification des opérateurs en métrologie dimensionnelle, un moyen pour intégrer la fonction métrologique dans les entreprises – Qualification of operators in dimensional metrology, a way to integrate the metrological function in small- or medium-sized firms, International Congress of Metrology (2001) B. Larquier, Le paysage de la formation longue durée en métrologie française – Background of long-lasting training in French metrology, International Congress of Metrology (2001) French College of Metrology, Metrology in the Firm: The Tool of Quality (1996 edition) P. Desbordes, Besoins des entreprises: compétences des métrologues en Europe – Requirements of firms: abilities of metrologists in Europe, International Congress of Metrology (1999) M. Fritz, L'école supérieure de métrologie: une nouvelle formation d’ingénieur – Higher School of metrology: a new training for engineers, International Congress of Metrology (1999)
The Authors
This book has been written by a working group of the Collège Français de Métrologie. The following writters have taken part in the compilation of the book: • •
Jean-Yves Arriat Luc Erard
• • • •
Claude Koch Philippe Lanneau Bernard Larquier Jean-François Magana
• • •
José Montes Roberto Perissi Marc Priel
• • • •
Patrick Reposeur Klaus-Dieter Schitthelm Patrizia Tavella Jean-Michel Virieux
Ascent Consulting Laboratoire National de Métrologie et d’Essais (LNE) Retired Management Services BEA Métrologie Organisation Internationale de Métrologie Légale (OIML) Solvay/Belgium ENIQ/Italy Laboratoire National de Métrologie et d’Essais (LNE) Comité Français d’Accréditation (COFRAC) Metrology Expert/Germany IENGF/Italy METAS/Switzerland
Pierre Barbier has led the working group and coordinated the compilation of the book.
Collège Français de Métrologie 1 rue Gaston Boissier 75724 Paris Cedex 15 – France www.cfmetrologie.com
This page intentionally left blank
Index
A, B Accreditation 54-55, 59, 75 Accuracy 36, 41 Adjusting 241 Air conditioning 198, 199 Bank of measuring instruments 113, 115-118
C Calibration 22-24, 28, 31, 34, 35-42, 97-100 interval 149, 150, 158, 159, 160, 161 label 231 results 133, 134 Capability of measuring instruments 29 Check standard 154, 159, 160 Continuing education 251-253 Continuous improvement 234 Control chart 152, 153, 154-157, 160 Covariance 180, 181, 182
D, E Differential measurement 214-215 Direct measurement 214-215 Distribution of the measurements 228 Error 164-165, 168-169
European cooperation 45, 53, 54-55, 70
F, G Field of measurement 210 Fitting 239, 241 Follow-up 123-125 Freedom from bias 96, 100 Freedom of bias 183, 186, 188 Graphs 220
I Identification 228-232 Identification sheet 112, 115-118 Indirect measurement 214-215 Influencing quantities 212-213, 215 Initial training 250-251 International system of units 129
L Label 111-112, 116 Legal metrology 67-77 Long-lasting training courses 253265
M, N Maintenance 116, 118, 120, 122 Maximum permissible error 79, 96, 101
270
Metrology in Industry
Measurement process 79, 86-89, 149, 150, 152-157 control process 236 uncertainty 163, 165, 174, 183 Measuring principle 213 Metrological confirmation 95, 98 engineer 248-250 function 20, 21, 26, 30, 31, 36, 42, 80-86 operator 250, 252 technician 249, 251, 263 Metrologist 247, 248, 251, 252, 255259, 264 Metrology profession 248 Mode of operation 207 Monitoring the measurement process 149 National calibration system 63, 65 metrology institute 130
S
P
V, W
Periodicity 111, 117, 119, 123-124 Procedure 195, 198, 200, 202, 205207 Process approach 235
Variance 166, 174-177, 180, 186, 191-192 Verification 81, 90, 91, 92, 94, 97100, 133 Verification results 133, 134 Work instruction 116-117
R Radioelectric disturbances 199 Random error 165, 169-170 Range of measurement 210 Receipt 119, 122 Recognition agreements 50, 54, 55 Reference materials 131, 137, 140, 142, 141-145 Regional organization 51-59 Relative humidity 198 Repeatability 87-88, 90, 96, 100, 101, 183, 184, 186, 187-193 Reproducibility 184, 187-188
Scientific metrology 74 Stability 100-101 Standard deviation 166, 175, 179, 181, 187, 188, 192 Standards 39-42 Storing 121 Subcontracting 232 Systematic error 165, 169-171
T Temperature variations 196 Traceability 127 chain 126, 129-131, 134, 135, 139, 141, 143 of the measurements 22, 36 to national standards 127, 135, 137, 145 Training 247-266 True value 169, 170, 186