Determination and use of straight-line calibration functions

ISO/TS 28037:2010 is concerned with linear, that is, straight-line, calibration functions that describe the relationship between two variables X and Y, namely, functions of the form Y = A + BX. Although many of the principles apply to more general types of calibration function, the approaches described exploit the simple form of the straight-line calibration function wherever possible. Values of the parameters A and B are determined on the basis of measured data points (xi, yi), i = 1, ... , m. Various cases are considered relating to the nature of the uncertainties associated with these data. No assumption is made that the errors relating to the yi are homoscedastic (having equal variance), and similarly for the xi when the errors are not negligible. Estimates of the parameters A and B are determined using least squares methods. The emphasis is on choosing the least squares method most appropriate for the type of measurement data, in particular methods that reflect the associated uncertainties. The most general type of covariance matrix associated with the measurement data is treated, but important special cases that lead to simpler calculations are described in detail. For all cases considered, methods for validating the use of the straight-line calibration functions and for evaluating the uncertainties and covariance associated with the parameter estimates are given. ISO/TS 28037:2010 also describes the use of the calibration function parameter estimates and their associated uncertainties and covariance to predict a value of X and its associated standard uncertainty given a measured value of Y and its associated standard uncertainty.

Détermination et utilisation des fonctions d'étalonnage linéaire

L'ISO/TS 28037:2010 traite des fonctions d'étalonnage linéaire, c'est-à-dire une droite, qui décrivent la relation entre deux variables X et Y, notamment les fonctions de la forme Y = A + BX. Bien que de nombreux principes s'appliquent à des types de fonction d'étalonnage plus généraux, les approches décrites utilisent dans la mesure du possible cette forme simple de la fonction d'étalonnage linéaire. Les valeurs des paramètres A et B sont déterminées sur la base de paires de données mesurées (xi,yi), i = 1,?, m. Selon la nature des incertitudes associées à ces données, différents cas sont considérés. Aucune hypothèse n'est faite sur les erreurs relatives à yi pour savoir si elles sont homoscédastiques (de variances égales), et de même pour xi lorsque les erreurs ne sont pas négligeables. Les estimations des paramètres A et B sont déterminées par les méthodes des moindres carrés. La présente Spécification Technique met l'accent sur le choix de la méthode des moindres carrés la plus appropriée au type de données de mesure, notamment les méthodes qui reflètent les incertitudes associées. Le cas le plus général de matrice de covariances associée aux données de mesure est traité, mais des cas particuliers importants, donnant lieu à des calculs plus simples, sont décrits en détail. Pour tous les cas considérés, les méthodes de validation du choix de fonctions d'étalonnage linéaire et d'évaluation des incertitudes et de la covariance associées aux estimations des paramètres sont données. L'ISO/TS 28037:2010 décrit aussi l'utilisation des estimations des paramètres de fonction d'étalonnage et de leurs incertitudes et covariance associées dans la prédiction d'une valeur de X et son incertitude-type associée pour une valeur mesurée de Y donnée avec son incertitude-type associée.

General Information

Status
Published
Publication Date
29-Aug-2010
Current Stage
9092 - International Standard to be revised
Start Date
15-Dec-2021
Completion Date
13-Dec-2025
Ref Project

Relations

Technical specification
ISO/TS 28037:2010 - Determination and use of straight-line calibration functions Released:8/30/2010
English language
63 pages
sale 15% off
Preview
sale 15% off
Preview
Technical specification
ISO/TS 28037:2010 - Détermination et utilisation des fonctions d'étalonnage linéaire Released:7/15/2013
French language
70 pages
sale 15% off
Preview
sale 15% off
Preview

Standards Content (Sample)


TECHNICAL ISO/TS
SPECIFICATION 28037
First edition
2010-09-01
Determination and use of straight-line
calibration functions
Détermination et utilisation des fonctions d'étalonnage linéaire

Reference number
©
ISO 2010
PDF disclaimer
This PDF file may contain embedded typefaces. In accordance with Adobe's licensing policy, this file may be printed or viewed but
shall not be edited unless the typefaces which are embedded are licensed to and installed on the computer performing the editing. In
downloading this file, parties accept therein the responsibility of not infringing Adobe's licensing policy. The ISO Central Secretariat
accepts no liability in this area.
Adobe is a trademark of Adobe Systems Incorporated.
Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation
parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In
the unlikely event that a problem relating to it is found, please inform the Central Secretariat at the address given below.

©  ISO 2010
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or
ISO's member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO 2010 – All rights reserved

Contents Page
Foreword:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: v
Introduction::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: vi
1 Scope :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 1
2 Normative references::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 1
3 Terms and de nitions :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 1
4 Conventions and notation ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 4
5 Principles of straight-line calibration:::::::::::::::::::::::::::::::::::::::::::::::::: 5
5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5.2 Inputs to determining the calibration function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5.2.1 Measurement data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5.2.2 Associated uncertainties and covariances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
5.3 Determining the calibration function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
5.4 Numerical treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5.5 Uncertainties and covariance associated with the calibration function parameters . . . . . . . . . . . . . . . . . 7
5.6 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.7 Use of the calibration function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.8 Determining the ordinary least squares best- t straight line to data . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
6 Model for uncertainties associated with the y :::::::::::::::::::::::::::::::::::::::::: 9
i
6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
6.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 10
6.3 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
6.4 Organization of the calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
7 Model for uncertainties associated with the x and the y :::::::::::::::::::::::::::::::: 17
i i
7.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
7.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 18
7.3 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
7.4 Organization of the calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
8 Model for uncertainties associated with the x and the y and covariances associated with the
i i
pairs (x ;y ):::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 24
i i
8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
8.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 24
9 Model for uncertainties and covariances associated with the y :::::::::::::::::::::::::::: 25
i
9.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 25
9.3 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
9.4 Organization of the calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
10 Model for uncertainties and covariances associated with the x and the y :::::::::::::::::: 31
i i
10.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
10.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 31
10.3 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
11 Use of the calibration function::::::::::::::::::::::::::::::::::::::::::::::::::::::: 37
11.1 Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
11.2 Forward evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

c ISO 2010 | All rights reserved iii

Annexes
A (informative) Matrix operations:::::::::::::::::::::::::::::::::::::::::::::::::::::: 41
A.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2 Elementary operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2.1 Matrix-vector multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2.2 Matrix-matrix multiplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2.3 Matrix transpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2.4 Identity matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.2.5 Inverse of a square matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.3 Elementary de nitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
A.3.1 Symmetric matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.3.2 Invertible matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.3.3 Lower-triangular and upper-triangular matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.3.4 Orthogonal matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.4 Cholesky factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.4.1 Cholesky factorization algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
A.4.2 Interpretation of the Cholesky factorization of a covariance matrix . . . . . . . . . . . . . . . . . . . . . . 43
A.4.3 Solution of a lower-triangular system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
A.4.4 Solution of an upper-triangular system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
A.5 Orthogonal factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
A.5.1 QR factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
A.5.2 RQ factorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
B (informative) Application of the Gauss-Newton algorithm to generalized distance regression :: 46
C (informative) Orthogonal factorization approach to solving the generalized Gauss-Markov prob-
lem :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 48
C.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
C.2 Calibration parameter estimates and associated standard uncertainties and covariance . . . . . . . . . . . . . 48
C.3 Validation of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
D (informative) Provision of uncertainties and covariances associated with the measured x- and
y-values ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 52
D.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
D.2 Response data 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
D.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
D.2.2 Measurement model for uncertainties and covariances associated with the y . . . . . . . . . . . . . . . 52
i
D.3 Response data 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
D.4 Stimulus data 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
D.5 Stimulus data 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
D.6 Stimulus and response data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
E (informative) Uncertainties known up to a scale factor :::::::::::::::::::::::::::::::::: 56
F (informative) Software implementation of described algorithms ::::::::::::::::::::::::::: 60
G (informative) Glossary of principal symbols :::::::::::::::::::::::::::::::::::::::::::: 61
Bibliography::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: 63
iv
c ISO 2010 | All rights reserved

Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO
member bodies). The work of preparing International Standards is normally carried out through ISO technical
committees. Each member body interested in a subject for which a technical committee has been established has the
right to be represented on that committee. International organizations, governmental and non-governmental, in liaison
with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission
(IEC) on all matters of electrotechnical standardization.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of technical committees is to prepare International Standards. Draft International Standards adopted
by the technical committees are circulated to the member bodies for voting. Publication as an International Standard
requires approval by at least 75 % of the member bodies casting a vote.
In other circumstances, particularly when there is an urgent market requirement for such documents, a technical
committee may decide to publish other types of normative document:
| an ISO Publicly Available Speci cation (ISO/PAS) represents an agreement between technical experts in an ISO
working group and is accepted for publication if it is approved by more than 50 % of the members of the parent
committee casting a vote;
| an ISO Technical Speci cation (ISO/TS) represents an agreement between the members of a technical committee
and is accepted for publication if it is approved by 2/3 of the members of the committee casting a vote.
An ISO/PAS or ISO/TS is reviewed after three years in order to decide whether it will be con rmed for a further
three years, revised to become an International Standard, or withdrawn. If the ISO/PAS or ISO/TS is con rmed, it is
reviewed again after a further three years, at which time it must either be transformed into an International Standard
or be withdrawn.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights.
ISO shall not be held responsible for identifying any or all such patent rights.
mittee SC 6, Measurement methods and results.

c ISO 2010 | All rights reserved v

Introduction
Calibration is an essential part of many measurement procedures and often involves tting to measured data a cali-
bration function that best describes the relationship of one variable to another. This Technical Speci cation considers
straight-line calibration functions that describe a dependent variable Y as a function of an independent variable X.
The straight-line relationship depends on the intercept A and the slope B of the line. A and B are referred to as the
parameters of the line. The purpose of a calibration procedure is to determine estimates a and b of A and B for a
particular measuring system under consideration on the basis of measurement data (x ; y ), i = 1;:::;m, provided
i i
by the measuring system. The measurement data have associated uncertainty, which means there will be uncertainty
associated with a and b. This Technical Speci cation describes how a and b can be determined given the data and
the associated uncertainty information. It also provides a means for evaluating the uncertainties associated with these
estimates. The treatment of uncertainty in this Technical Speci cation is carried out in a manner consistent with
ISO/IEC Guide 98-3:2008, \Guide to the expression of uncertainty in measurement" (GUM).
Given the uncertainty information associated with the measurement data, an appropriate method can be speci ed
to determine estimates of the calibration function parameters. This uncertainty information may include quanti ed
covariance e ects, relating to dependencies among some or all of the quantities involved.
Once the straight-line model has been tted to the data, it is necessary to determine whether or not the model and
data are consistent with each other. In cases of consistency, the model so obtained can validly be used to predict a
value x of the variable X corresponding to a measured value y of the variable Y provided by the same measuring
system. It can also be used to evaluate the uncertainties associated with the calibration function parameters and the
uncertainty associated with the predicted value x.
The determination and use of a straight-line calibration function can therefore be considered to consist of ve steps:
1 Obtaining uncertainty and covariance information associated with the measurement data { although dependent
on the particular area of measurement, examples are provided within this Technical Speci cation;
2 Providing best estimates of the straight-line parameters;
3 Validating the model, both in terms of the functional form (does the data re
ect a straight-line relationship?)
and statistically (is the spread of the data consistent with their associated uncertainties?) using a chi-squared
test;
4 Obtaining the standard uncertainties and covariance associated with the estimate of the straight-line parameters.
5 Using the calibration function for prediction, that is, determining an estimatex of theX-variable and its associated
uncertainty corresponding to a measured value y of the Y -variable and its associated uncertainty.
The above steps are shown diagrammatically in Figure 1.
The main aim of this Technical Speci cation is to consider steps 2 to 5. Therefore, as part of step 1, before using this
Technical Speci cation, the user will need to provide standard uncertainties, and covariances if relevant, associated
with the measured Y -values and, as appropriate, those associated with the measured X-values. Account should be
taken of the principles of the GUM in evaluating these uncertainties on the basis of a measurement model that is
speci c to the area of concern.
ISO 11095:1996 [14] is concerned with linear calibration using reference materials. It di ers from this Technical
Speci cation in the ways given in Table 1.
The numerical methods given are based on reference [6].
vi
c ISO 2010 | All rights reserved

Inputs
measurement data (x ;y ), i = 1;:::;m,
i i
and associated covariance matrixU
model
Y =A +BX
Calibration
estimates a of A and b of B
Validation
model residuals and observed
chi-squared value 
obs
Uncertainty
standard uncertainties u(a) and
evaluation
u(b), and covariance cov(a;b)
Prediction
measured value y of Y and
associated standard uncertainty u(y)
predicted value x of X and
associated standard uncertainty u(x)
Figure 1 | Summary of the steps in the determination and use of straight-line calibration functions
Table 1 | Di erences between ISO 11095:1996 and ISO/TS 28037:2010
Feature ISO 11095:1996 ISO/TS 28037:2010
Speci cally addresses reference materials Yes More general
X-values assumed to be known exactly Yes More general uncertainty information
All measured values obtained independently Yes More general uncertainty information
Terminology aligned with GUM No Yes
Types of uncertainty structure treated Two Five, including the most general case
Only uncertainty associated with random errors Yes More general uncertainty information
Consistency test ANOVA Chi-squared
Uncertainty associated with predictions Ad hoc GUM compatible

c ISO 2010 | All rights reserved vii

TECHNICAL SPECIFICATION ISO/TS 28037:2010(E)

Determination and use of straight-line calibration functions
1 Scope
This Technical Speci cation is concerned with linear, that is, straight-line, calibration functions that describe the
relationship between two variables X and Y , namely, functions of the form Y = A +BX. Although many of the
principles apply to more general types of calibration function, the approaches described exploit the simple form of the
straight-line calibration function wherever possible.
Values of the parametersA andB, are determined on the basis of measured data points (x ; y ),i = 1;:::;m. Various
i i
cases are considered relating to the nature of the uncertainties associated with these data. No assumption is made
that the errors relating to the y are homoscedastic (having equal variance), and similarly for the x when the errors
i i
are not negligible.
Estimates of the parameters A and B are determined using least squares methods. The emphasis of this Technical
Speci cation is on choosing the least squares method most appropriate for the type of measurement data, in particular
methods that re
ect the associated uncertainties. The most general type of covariance matrix associated with the
measurement data is treated, but important special cases that lead to simpler calculations are described in detail.
For all cases considered, methods for validating the use of the straight-line calibration functions and for evaluating
the uncertainties and covariance associated with the parameter estimates are given.
The Technical Speci cation also describes the use of the calibration function parameter estimates and their associated
uncertainties and covariance to predict a value of X and its associated standard uncertainty given a measured value
of Y and its associated standard uncertainty.
NOTE 1 The Technical Speci cation does not give a general treatment of outliers in measurement data, although the validation
tests given can be used as a basis for identifying discrepant data.
NOTE 2 The Technical Speci cation describes a method to evaluate the uncertainties associated with the measurement data
in the case that those uncertainties are known only up to a scale factor (Annex E).
2 Normative references
The following referenced documents are indispensable for the application of this document. For dated references,
only the edition cited applies. For undated references, the latest edition of the referenced document (including any
amendments) applies.
ISO/IEC Guide 99:2007, International vocabulary of metrology | Basic and general concepts and associated terms
(VIM)
ISO/IEC Guide 98-3:2008, Uncertainty of measurement | Part 3: Guide to the expression of uncertainty in measure-
ment (GUM:1995)
3 Terms and de nitions
For the purposes of this document, the terms and de nitions given in ISO/IEC Guide 98-3:2008 and ISO/IEC Guide
99:2007 and the following apply.
A glossary of principal symbols is given in Annex G.

c ISO 2010 | All rights reserved 1

3.1
measured quantity value
quantity value representing a measurement result
[ISO/IEC Guide 99:2007 2.10]
3.2
measurement uncertainty
non-negative parameter characterizing the dispersion of the quantity values being attributed to a measurand, based
on the information used
[ISO/IEC Guide 99:2007 2.26]
3.3
standard measurement uncertainty
measurement uncertainty expressed as a standard deviation
[ISO/IEC Guide 99:2007 2.30]
3.4
covariance associated with two quantity values
parameter characterizing the interdependence of the quantity values being attributed to two measurands, based on
the information used
3.5
measurement covariance matrix
covariance matrix
matrix of dimension NN associated with a vector estimate of a vector quantity of dimension N 1, containing on
its diagonal the squares of the standard uncertainties associated with the respective components of the vector estimate
of the vector quantity, and, in its o -diagonal positions, the covariances associated with pairs of components of the
vector estimate of the vector quantity
NOTE 1 A covariance matrix U of dimension NN associated with the vector estimate x of a vector quantityX has the
x
representation
2 3
cov(x ;x )  cov(x ;x )
1 1 1 N
. .
.
4 5
U = . . . ;
x
.
. .
cov(x ;x )  cov(x ;x )
N 1 N N
where cov(x ;x ) = u (x ) is the variance (squared standard uncertainty) associated with x and cov(x ;x ) is the covariance
i i i i i j
associated with x and x . cov(x ;x ) = 0 if elements X and X ofX are uncorrelated.
i j i j i j
NOTE 2 Covariances are also known as mutual uncertainties.
NOTE 3 A covariance matrix is also known as a variance-covariance matrix.
NOTE 4 De nition adapted from ISO/IEC Guide 98-3:2008/Suppl. 1:2008, de nition 3.11 [13].
3.6
measurement model
mathematical relation among all quantities known to be involved in a measurement
[ISO/IEC Guide 99:2007 2.48]
3.7
functional model
statistical model involving errors associated with the dependent variable
c ISO 2010 | All rights reserved

3.8
structural model
statistical model involving errors associated with the independent and dependent variables
3.9
calibration
operation that, under speci ed conditions, in a rst step, establishes a relation between the quantity values with
measurement uncertainties provided by measurement standards and corresponding indications with associated mea-
surement uncertainties and, in a second step, uses this information to establish a relation for obtaining a measurement
result from an indication
NOTE 1 A calibration may be expressed by a statement, calibration function, calibration diagram, calibration curve, or
calibration table. In some cases, it may consist of an additive or multiplicative correction of the indication with associated
measurement uncertainty.
NOTE 2 Calibration should not be confused with adjustment of a measuring system, often mistakenly called `self-calibration',
nor with veri cation of calibration.
NOTE 3 Often the rst step alone in the above de nition is perceived as being calibration.
[ISO/IEC Guide 99:2007 2.39]
3.10
probability distribution
hrandom variablei function giving the probability that a random variable takes any given value or belongs to a given
set of values
NOTE 1 The probability on the whole set of values of the random variable equals 1.
NOTE 2 A probability distribution is termed univariate when it relates to a single (scalar) random variable, and multivariate
when it relates to a vector of random variables. A multivariate probability distribution is also described as a joint distribution.
NOTE 3 A probability distribution can take the form of a distribution function or a probability density function.
NOTE 4 De nition and note 1 adapted from ISO 3534-1:1993, de nition 1.3 and ISO/IEC Guide 98-3:2008, de nition C.2.3;
notes 2 and 3 adapted from ISO/IEC Guide 98-3:2008/Suppl. 1:2008, de nition 3.1 [13].
3.11
normal distribution
probability distribution of a continuous random variable X having the probability density function
" #
 
1 1 
g () = p exp ;
X
2 
 2
for1<< +1
NOTE 1  is the expectation and  is the standard deviation of X.
NOTE 2 The normal distribution is also known as the Gaussian distribution.
NOTE 3 De nition and note 1 adapted from ISO 3534-1:1993, de nition 1.37; note 2 adapted from ISO/IEC Guide 98-3:2008,
de nition C.2.14.
3.12
t-distribution
probability distribution of a continuous random variable X having the probability density function
 
 + 1
 
(+1)=2
2 
g () =p 1 + ;
X
(=2) 
c ISO 2010 | All rights reserved 3

for1<< +1, with parameter , a positive integer, the degrees of freedom of the distribution, where
Z
z1 t
(z) = t e dt; z> 0;
is the gamma function
[ISO/IEC Guide 98-3:2008/Suppl. 1:2008 3.5]
3.13
chi-squared distribution
 distribution
probability distribution of a continuous random variable X having the probability density function
 
(=2)1
 
g () = exp ;
X
=2
2 (=2) 2
for 0<1, with parameter , a positive integer, where is the gamma function
NOTE The sum of the squares of independent standardized normal variables is a  random variable with parameter; is
then called the degrees of freedom.
3.14
positive de nite matrix
>
matrixM of dimension nn having the propertyz Mz> 0 for all non-zero vectorsz of dimension n 1
3.15
positive semi-de nite matrix
>
matrixM of dimension nn having the propertyz Mz 0 for all non-zero vectorsz of dimension n 1
4 Conventions and notation
For the purpose of this Technical Speci cation the following conventions and notations are adopted.
4.1 X is termed the independent variable and Y the dependent variable even when the knowledge of X and Y is
`interchangeable', as in Clause 7, for example.
4.2 The quantities A and B are termed the parameters of the straight-line calibration function Y =A +BX. A
and B are also used to denote (dummy) variables in expressions involving the calibration function parameters.
4.3 The quantities X and Y are used as (dummy) variables to denote the co-ordinates of the ith data point.
i i
 
4.4 The constants A and B are (unknown) values of A and B that specify the straight-line calibration function
 
Y =A +B X for a particular measuring system under consideration.
 
4.5 The constants X and Y are the (unknown) co-ordinates of the ith data point provided by the measuring
i i
   
system satisfying Y =A +B X .
i i
4.6 x and y are the measured values of the co-ordinates of the ith data point.
i i
4.7 a and b are estimates of the calibration function parameters for the measuring system.
   
4.8 x and y are estimates of the co-ordinates of the ith data point satisying y =a +bx .
i i i i
c ISO 2010 | All rights reserved

4.9 A vector of dimension m 1 is denoted thus:
2 3
x
 
6 7
. >
x = . ; x = x ::: x ;
4 5
1 m
.
x
m
and a matrix of dimension mn is denoted thus:
2 3 2 3
a ::: a a ::: a
11 1n 11 m1
6 . . 7 6 . . 7
. > .
. . . . . .
A = ; A = :
4 5 4 5
. .
. . . .
a ::: a a ::: a
m1 mn 1n mn
The dimension of the vector or matrix is always speci ed to avoid possible confusion.
4.10 > denotes transpose.
4.11 The zero matrix is denoted by 0 and the unit vector is denoted by 1.
4.12 Some symbols have more than one meaning. The context clari es the usage.
4.13 Numbers displayed in tables to a xed number of decimal places are correctly rounded representations of
numbers stored to higher precision, as would be the case in a spreadsheet, for example. Therefore, minor inconsistencies
may be perceived between displayed column sums and the column sums of the displayed numbers.
4.14 In some tables, a subclause number above a column or columns indicates where the formula is given for
determining the values below.
4.15 In the examples, while data values are provided to a given precision, the results of calculations are provided
to a higher precision to allow the user to compare results when undertaking the calculations.
5 Principles of straight-line calibration
5.1 General
5.1.1 This clause considers how a relationship Y = A +BX describing the dependent variable Y (also called
`response') as a function of the independent variable X (also called `stimulus') can be determined from measurement
data. In the context of calibration, the measurement data arise when a measuring instrument speci ed by (unknown)
 
valuesA andB of the calibration function parameters is `stimulated' by artefacts with calibrated values of X given
i
in standard units, of a property of the artefacts, and the corresponding `responses' or indications Y of the instrument
i
are recorded. The relationship provides the response Y of the system given an artefact with calibrated quantity X.
This process is termed `forward evaluation'. More useful in practice, the relationship allows a measured response y
of Y to be converted to an estimate x, in standard units, of the property X of an artefact. This process is termed
`inverse evaluation' or `prediction'.
5.1.2 The calibration of a measuring system should take into account measurement uncertainties, and, if present,
covariances associated with the measurement data. The output of a calibration procedure is a calibration function to
be used for prediction (and, if required, forward evaluation). The output also includes the standard uncertainties and
covariance associated with the estimatesa andb of the parameters describing the calibration function, which are used
to evaluate the standard uncertainties associated with prediction (and forward evaluation).
5.2 Inputs to determining the calibration function
5.2.1 Measurement data
The information required to determine the straight-line calibration function are the measurement data and their
associated standard uncertainties and covariances. In this Technical Speci cation, the measurement data are denoted

c ISO 2010 | All rights reserved 5

by (x ;y ), i = 1;:::;m, that is, m pairs of measured values of X and Y . It is assumed that m is at least two and
i i
that the values of x are not all equal to each other.
i
NOTE The uncertainties associated with the estimatesa andb generally decrease asm increases. Therefore, calibration should
aim to use as many measured data points as is economically viable.
5.2.2 Associated uncertainties and covariances
The standard uncertainties associated with x and y are denoted by u(x ) and u(y ) respectively. The covariance
i i i i
associated with x and x is denoted by cov(x ;x ). Similarly, those associated with y and y , and with x and y ,
i j i j i j i j
are denoted by cov(y ;y ) and cov(x ;y ), respectively. Annex D indicates how the uncertainties and covariances
i j i j
associated with the measured response and stimulus variables can be evaluated and gives an interpretation of that
uncertainty information. The complete uncertainty information is represented by an array of elements (matrix) U of
2 2
dimension 2m 2m holding the variances (squared standard uncertainties) u (x ) and u (y ) and the covariances:
i i
2 3
u (x ) ::: cov(x ;x ) cov(x ;y ) ::: cov(x ;y )
1 1 m 1 1 1 m
6 . . . . 7
. .
. . . . . .
6 7
. .
. . . .
6 7
6 7
cov(x ;x ) ::: u (x ) cov(x ;y ) ::: cov(x ;y )
m 1 m m 1 m m
6 7
U = :
6 7
cov(y ;x ) ::: cov(y ;x ) u (y ) ::: cov(y ;y )
1 1 1 m 1 1 m
6 7
6 7
. . . .
. .
. . . . . .
4 5
. .
. . . .
cov(y ;x ) ::: cov(y ;x ) cov(y ;y ) ::: u (y )
m 1 m m m 1 m
For many applications, some or all covariances are taken as zero (see 5.3).
NOTE This Technical Speci cation is concerned with problems in which the u(x ) or the u(y ) are generally di erent.
i i
5.3 Determining the calibration function
5.3.1 The inputs to determining the calibration function are the measurement data and their associated uncertain-
ties and possibly covariances. Given parametersA andB, the inputs can be used to provide a measure of the departure
of the ith data point (x ;y ) from the line Y =A +BX. The estimates a and b are determined by minimizing a sum
i i
of squares of these departures, or a more general measure when any covariances are non-zero. How this is achieved
depends on the `uncertainty structure' associated with the measurement data. This uncertainty structure relates to
the answers to the following questions:
i) Are the uncertainties associated with the measured values x negligible?
i
ii) Are the covariances associated with pairs of measured values negligible?
5.3.2 The following cases, given in increasing order of complexity and depending on the answers to the questions
in 5.3.1, are considered in this Technical Speci cation:
a) The only uncertainties are associated with the measured values y and all covariances associated with the data
i
are regarded as negligible (Clause 6);
b) Uncertainties are associated with the measured values x and y and all covariances associated with the data are
i i
regarded as negligible (Clause 7);
c) Uncertainties are associated with the measured values x andy and the only covariances are associated with the
i i
pairs (x ;y ) (Clause 8);
i i
d) The only uncertainties are associated with the measured values y and the only covariances are associated with
i
the y and the y (i6=j) (Clause 9);
i j
e) The most general case in which there are uncertainties associated with the measured values x and y and
i i
covariances associated with all pairs of values of the x , the x , the y and the y (Clause 10).
i j k `
c ISO 2010 | All rights reserved

5.3.3 For each case in 5.3.2 are given
a) the prescribed measurement data and uncertainty structure,
b) the corresponding statistical model,
c) the least squares problem addressed,
d) the calculation steps,
e) properties of the statistical model,
f) validation of the model,
g) organization of the calculations for the computer, where appropriate,
h) a numerical algorithm, where appropriate, and
i) one or more worked examples.
5.4 Numerical treatment
In Annex C, a general approach to the most general case e) in 5.3.2 is given. It can be used to treat all the other
cases and uses sophisticated, numerically stable methods. The cases a) to c) in 5.3.2 can, however, be treated using
elementary operations, which can be implemented in a spreadsheet, for example. The cases d) and e) in 5.3.2 require
some matrix operations, which are straightforward to implement in a computer language supporting matrix arithmetic,
but are not well suited to spreadsheet calculations.
5.5 Uncertainties and covariance associated with the calibration function parameters
5.5.1 For all cases considered, estimates of the calibration function parameters can be expressed (explicitly or
implicitly) as functions of the measurement data. The principles of the GUM [ISO/IEC Guide 98-3:2008] can be
applied to propagate the uncertainties and covariances associated with the measurement data through these functions
to obtain those associated with these parameter estimates. In this way, the measurement data are used to provide
estimates a and b of the calibration function parameters, and to evaluate standard uncertainties u(a) and u(b) and
the covariance cov(a;b) associated with these estimates. For the cases a) and d) in 5.3.2, the propagation is exact
since the parameter estimates can be expressed as linear combinations of the inputs y . For the other cases, in which
i
the parameter estimates cannot be so expressed, the propagation is approximate, based on a linearization about the
parameter estimates. For many purposes, the approximation incurred by the linearization will be suciently accurate.
NOTE When the propagation of uncertainty is approximate, and particularly if the uncertainties involved are large (for example,
in some areas of biological measurement), an approach based on the propagation of distributions can be employed. This approach
[ISO/IEC Guide 98-3:2008/Suppl. 1:2008] uses a Monte Carlo method (not treated in this Technical Speci cation).
5.5.2 The primary outputs in describing the straight-line calibration function are the parameter estimate vector a
of dimension 2 1 and the covariance matrixU of dimension 2 2 given by
a
   
a u (a) cov(a;b)
a = ; U = ; (1)
a
b cov(b;a) u (b)
where u(a) and u(b) are the standard uncertainties associated with a and b, respectively, and cov(a;b) = cov(b;a) is
the covariance associated with a and b.

c ISO 2010 | All rights reserved 7

5.6 Validation of the model
5.6.1 In determining the estimates a and b of the straight-line calibration function parameters, it is assumed
that the modelY =A +BX is valid and that the uncertainties associated with the measurement data give a credible
measure of the departure of the measurement data from a straight line. Oncea andb have been determined, the actual
departure of the data points from the best- t calibration function can be assessed against a predicted departure. This
comparison involves an aggregate measure of departure expressed in terms of the sum of squares  of m weighted
obs
residuals, theith weighted residual being a measure of the departure of the ith data point from the line, or, when the
covariance associated with the ith data point (x ;y ) is non-zero, a more general form. If  is much bigger than
i i
obs
expected, on statistical grounds there is reason to call into question the validity of the model assumptions.
5.6.2 From a statistical viewpoint, the measurement data can be regarded as realizations of random variables. If
the probability distributions characterizing these random variables were known, it would be possible in principle to
determine the probability distribution for the aggregate measure of departure in 5.6.1. Then the probability could
be calculated that  , regarded as a draw from this aggregate distribution, exceeded any particular quantile of the
obs
distribution. However, as the information about these quantities is often limited to the measured values themselves
and their associated variances (taken to be the expectations and variances, respectively, of the random variables
characterized by these distributions), there is insucient information to determine the distribution for this measure.
Instead, the assessment of validity is performed assuming that the distributions for these quantities are normal. With
this assumption, which is henceforth taken to hold, at least for validation purposes, the distribution for this measure
2 2 2
is with =m 2 degrees of freedom. Accordingly, the probability that  exceeds any particular quantile of 
 obs 
can be determined (see 6.3, 7.3, 9.3, 10.3). The 95 % quantile is used.
2 2
NOTE 1 If  exceeds the 95 % quantile of  , the straight-line calibration function can be regarded as not explaining the
obs 
data suciently well for practical purposes. In such a case, the data and associated uncertainties should be checked for possible
mistakes. A calibration function consisting of a polynomial in X of degree 2 or higher or some other mathematical form can be
entertained; such a consideration is beyond the scope of this Technical Speci cation.
NOTE 2 There is a possibility that the model is `too good' in that the observed value  is signi cantly smaller than the
obs
expected value. This possibility typically corresponds to the uncertainties associated with the measurement data being quoted
as too large, and is not considered further in this Technical Speci cation.
5.6.3 In order to obtain as much value as possible from a calibration, it is desirable that input uncertainties are
derived prior to determining the calibration function parameters, rather than being evaluated once a t to the data
has been determined, with associated uncertainties estimated from the data or known up to a scale factor. The latter
case is considered in Annex E.
2 2
5.6.4 If, in any particular case, the validation of the model fails, that is,  exceeds the 95 % quantile of 
obs 
(see 5.6.2), the calculated standard uncertaintiesu(a) andu(b) and covariance cov(a;b) (see 5.5.2) should be regarded
as unreliable, as should the uncertainty associated with a predicted value (see 5.7).
5.7 Use of the calibration function
5.7.1 The calibration function is typically used for prediction (inverse evaluation) where, given an estimate ofY and
its associated standard uncertainty, the corresponding value ofX is estimated and its associated standard uncertainty
is evaluated. Evaluation of the latter uncertainty makes use of the standard uncertainties associated with the estimates
a and b as well as their associated covariance. See 11.1.
5.7.2 Forward evaluation where, given an estimate of X and its associated uncertainty, the corresponding value
ofY is obtained together with its associated standard uncertainty, is sometimes required, for example, when comparing
the calibrations of a set of similar instruments. See 11.2.
NOTE It is assumed that the conditions of measurement that held dur
...


SPÉCIFICATION ISO/TS
TECHNIQUE 28037
Première édition
2010-09-01
Détermination et utilisation des
fonctions d’étalonnage linéaire
Determination and use of straight-line calibration functions
Numéro de référence
©
ISO 2010
DOCUMENT PROTÉGÉ PAR COPYRIGHT
© ISO 2010
Droits de reproduction réservés. Sauf indication contraire, aucune partie de cette publication ne peut être reproduite ni utilisée
sous quelque forme que ce soit et par aucun procédé, électronique ou mécanique, y compris la photocopie, l’affichage sur
l’internet ou sur un Intranet, sans autorisation écrite préalable. Les demandes d’autorisation peuvent être adressées à l’ISO à
l’adresse ci-après ou au comité membre de l’ISO dans le pays du demandeur.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Version française parue en 2013
Publié en Suisse
ii © ISO 2010 – Tous droits réservés

Sommaire Page
Avant-propos .v
Introduction .vi
1 Domaine d’application . 1
2 Références normatives . 1
3 Termes et définitions . 1
4 Conventions et notation . 4
5 Principes de l’étalonnage linéaire . 5
5.1 Généralités . 5
5.2 Éléments d’entrée pour la détermination de la fonction d’étalonnage . 5
5.3 Détermination de la fonction d’étalonnage . 6
5.4 Traitement numérique . 7
5.5 Incertitudes et covariances associées aux paramètres de la fonction d’étalonnage . 7
5.6 Validation du modèle . 8
5.7 Utilisation de la fonction d’étalonnage . 8
5.8 Détermination de la droite de meilleur ajustement des données par la méthode des
moindres carrés ordinaires . 9
6 Modèle applicable aux incertitudes associées à y . 9
i
6.1 Généralités . 9
6.2 Estimations des paramètres d’étalonnage, des incertitudes-types et de la
covariance associées .10
6.3 Validation du modèle .11
6.4 Organisation des calculs .12
7 Modèle applicable aux incertitudes associées à x et y .16
i i
7.1 Généralités .16
7.2 Estimations des paramètres d’étalonnage, des incertitudes-types et de la
covariance associées .18
7.3 Validation du modèle .19
7.4 Organisation des calculs .20
8 Modèle applicable aux incertitudes associées à x et y et aux covariances associées aux
i i
paires (x , y ) .24
i i
8.1 Généralités .24
8.2 Estimations des paramètres d’étalonnage et incertitudes-types et covariance associées .25
9 Modèle applicable aux incertitudes et aux covariances associées à y .25
i
9.1 Généralités .25
9.2 Estimations des paramètres d’étalonnage, incertitudes-types et covariance associées .26
9.3 Validation du modèle .28
9.4 Organisation des calculs .28
10 Modèle applicable aux incertitudes et aux covariances associées à x et y .32
i i
10.1 Généralités .32
10.2 Estimations des paramètres d’étalonnage, des incertitudes-types et
covariances associées .33
10.3 Validation du modèle .35
11 Utilisation de la fonction d’étalonnage .39
11.1 Prédiction directe .39
11.2 Prédiction inverse .40
Annexe A (informative) Opérations matricielles .42
Annexe B (informative) Application de l’algorithme de Gauss-Newton à la régression selon le
critère de distance généralisée .48
Annexe C (informative) Approche de factorisation orthogonale pour résoudre le problème de
Gauss-Markov généralisé .50
Annexe D (informative) Disposition relative aux incertitudes et covariances associées aux valeurs
mesurées x et y.56
Annexe E (informative) Incertitudes connues avec un facteur d’échelle donné .61
Annexe F (informative) Application logicielle des algorithmes décrits.66
Annexe G (informative) Glossaire des principaux symboles .68
Bibliographie .70
iv © ISO 2010 – Tous droits réservés

Avant-propos
L’ISO (Organisation internationale de normalisation) est une fédération mondiale d’organismes
nationaux de normalisation (comités membres de l’ISO). L’élaboration des Normes internationales est
en général confiée aux comités techniques de l’ISO. Chaque comité membre intéressé par une étude
a le droit de faire partie du comité technique créé à cet effet. Les organisations internationales,
gouvernementales et non gouvernementales, en liaison avec l’ISO participent également aux travaux.
L’ISO collabore étroitement avec la Commission électrotechnique internationale (CEI) en ce qui concerne
la normalisation électrotechnique.
Les Normes internationales sont rédigées conformément aux règles données dans les Directives
ISO/CEI, Partie 2.
La tâche principale des comités techniques est d’élaborer les Normes internationales. Les projets de
Normes internationales adoptés par les comités techniques sont soumis aux comités membres pour vote.
Leur publication comme Normes internationales requiert l’approbation de 75 % au moins des comités
membres votants.
Dans d’autres circonstances, en particulier lorsqu’il existe une demande urgente du marché, un comité
technique peut décider de publier d’autres types de documents normatifs:
— une Spécification publiquement disponible ISO (ISO/PAS) représente un accord entre les experts
dans un groupe de travail ISO et est acceptée pour publication si elle est approuvée par plus de 50 %
des membres votants du comité dont relève le groupe de travail;
— une Spécification technique ISO (ISO/TS) représente un accord entre les membres d’un comité technique
et est acceptée pour publication si elle est approuvée par 2/3 des membres votants du comité.
Une ISO/PAS ou ISO/TS fait l’objet d’un examen après trois ans afin de décider si elle est confirmée pour
trois nouvelles années, révisée pour devenir une Norme internationale, ou annulée. Lorsqu’une ISO/PAS
ou ISO/TS a été confirmée, elle fait l’objet d’un nouvel examen après trois ans qui décidera soit de sa
transformation en Norme internationale soit de son annulation.
L’attention est appelée sur le fait que certains des éléments du présent document peuvent faire l’objet de
droits de propriété intellectuelle ou de droits analogues. L’ISO ne saurait être tenue pour responsable de
ne pas avoir identifié de tels droits de propriété et averti de leur existence.
L’ISO/TS 28037:2010 a été élaborée par le comité technique ISO/TC 69, Applications des méthodes
statistiques, sous-comité SC 6, Méthodes et résultats de mesure.
Introduction
L’étalonnage est un élément essentiel de la plupart des méthodes de mesure qui consiste souvent à
ajuster des données mesurées au moyen d’une fonction d’étalonnage décrivant au mieux la relation d’une
variable à une autre. La présente Spécification Technique considère les fonctions d’étalonnage linéaire
qui décrivent une variable dépendante Y comme une fonction d’une variable indépendante X. La relation
linéaire dépend de l’ordonnée à l’origine A et de la pente B de la droite. A et B sont désignés sous le nom
de paramètres de la droite. Une procédure d’étalonnage a pour objet de déterminer des estimations
a et b de A et B pour un système de mesure particulier soumis à étude, sur la base des données de
mesure (x ,y ), i = 1,…, m, fournies par le système de mesure. Une incertitude est associée aux données
i i
de mesure, ce qui signifie qu’une incertitude sera associée à a et b. La présente Spécification Technique
décrit la manière dont a et b peuvent être déterminés sur la base des données et des informations
relatives à l’incertitude associée. Elle fournit aussi un moyen d’évaluation des incertitudes associées
à ces estimations. Le traitement de l’incertitude dans la présente Spécification Technique est effectué
conformément au Guide ISO/CEI 98-3:2008, « Guide pour l’expression de l’incertitude de mesure » (GUM).
Les informations sur l’incertitude associée aux données de mesure permettent de spécifier une méthode
appropriée pour déterminer les estimations des paramètres de la fonction d’étalonnage. Ces informations
sur l’incertitude peuvent inclure les effets de covariance quantifiés, concernant les dépendances entre
toutes ou certaines des grandeurs considérées.
Une fois le modèle linéaire ajusté aux données, il est nécessaire de déterminer si ce modèle et les données
sont cohérents entre eux ou non. En cas de cohérence, le modèle ainsi obtenu peut valablement être
utilisé pour prévoir une valeur x de la variable X correspondant à une valeur mesurée y de la variable Y
fournie par le même système de mesure. Il peut aussi être utilisé pour évaluer les incertitudes associées
aux paramètres de la fonction d’étalonnage et l’incertitude associée à la valeur prédite x.
Il est par conséquent possible de considérer que la détermination et l’utilisation d’une fonction
d’étalonnage linéaire se composent de cinq étapes:
1) Obtenir les informations relatives à l’incertitude et à la covariance associées aux données de
mesure – bien que dépendant d’un domaine spécifique de mesure, des exemples sont donnés dans la
présente Spécification Technique;
2) Fournir les meilleures estimations des paramètres de la droite;
3) Valider le modèle, aussi bien sur le plan fonctionnel (les données reflètent-elles bien une relation
linéaire ?) que statistique (l’étendue des données est-elle cohérente avec leurs incertitudes
associées ?) en utilisant un test de chi deux;
4) Obtenir les incertitudes-types et la covariance associées à l’estimation des paramètres de la droite;
5) Utiliser la fonction d’étalonnage pour la prédiction, c’est-à-dire déterminer une estimation x de la
variable X et son incertitude associée qui corresponde à une valeur mesurée y de la variable Y et à
son incertitude associée.
Les étapes ci-dessus sont illustrées schématiquement sur la Figure 1.
L’objectif principal de la présente Spécification Technique est de traiter les étapes 2 à 5. Par conséquent,
en ce qui concerne l’étape 1, l’utilisateur, avant d’utiliser la présente Spécification Technique, devra
fournir les incertitudes-types et les covariances le cas échéant, associées aux valeurs mesurées Y et,
selon le cas, celles associées aux valeurs mesurées X. Il convient de tenir compte des principes du GUM
pour l’évaluation de ces incertitudes sur la base d’un modèle de mesure spécifique au domaine d’étude.
[14]
L’ISO 11095:1996 traite de l’étalonnage linéaire utilisant des matériaux de référence. Il se différencie
de la présente Spécification technique par les points mentionnés dans le Tableau 1.
[6]
Les méthodes numériques fournies sont basées sur la référence.
vi © ISO 2010 – Tous droits réservés

Éléments d’entrée
Données de mesure (x , y ), i = 1 …, m,
i i
et matrice de covariances associée U
Modèle
Y = A + BX
Étalonnage
Estimations a de A et b de B
Validation
Modèles des résidus et
χ
valeur chi deux
obs
Évaluation
de l’incertitude
incertitudes-types u(a) et u(b),
et covariance cov(a,b)
Prédiction
valeur mesurée y de Y et
incertitude-type associée u(y)
valeur prédite x de X et
incertitude-type associée u(x)
Figure 1 — Résumé des étapes pour la détermination et l’utilisation des fonctions
d’étalonnage linéaire
Tableau 1 — Différences entre l’ISO 11095:1996 et l’ISO/TS 28037:2010
Caractéristiques ISO 11095:1996 ISO/TS 28037:2010
Traite spécialement des matériaux de référence Oui Plus général
Valeurs X supposées connues avec exactitude Oui Informations d’incertitude plus géné-
rales
Toutes les valeurs mesurées obtenues séparément Oui Informations d’incertitude plus géné-
rales
Terminologie en accord avec celle du GUM Non Oui
Types de structure d’incertitude traités Deux Cinq, y compris le cas le plus général
Incertitude seulement associée à des erreurs aléa- Oui Informations d’incertitude plus géné-
toires rales
Essai de cohérence ANOVA Chi deux
Incertitude associée à des prédictions Ad hoc Compatible avec le GUM
viii © ISO 2010 – Tous droits réservés

SPÉCIFICATION TECHNIQUE ISO/TS 28037:2010(F)
Détermination et utilisation des fonctions d’étalonnage
linéaire
1 Domaine d’application
La présente Spécification Technique traite des fonctions d’étalonnage linéaire, c’est-à-dire une droite, qui
décrivent la relation entre deux variables X et Y, notamment les fonctions de la forme Y = A + BX. Bien que
de nombreux principes s’appliquent à des types de fonction d’étalonnage plus généraux, les approches
décrites utilisent dans la mesure du possible cette forme simple de la fonction d’étalonnage linéaire.
Les valeurs des paramètres A et B sont déterminées sur la base de paires de données mesurées (x ,y ),
i i
i = 1,…, m. Selon la nature des incertitudes associées à ces données, différents cas sont considérés.
Aucune hypothèse n’est faite sur les erreurs relatives à y pour savoir si elles sont homoscédastiques (de
i
variances égales), et de même pour x lorsque les erreurs ne sont pas négligeables.
i
Les estimations des paramètres A et B sont déterminées par les méthodes des moindres carrés. La
présente Spécification Technique met l’accent sur le choix de la méthode des moindres carrés la plus
appropriée au type de données de mesure, notamment les méthodes qui reflètent les incertitudes
associées. Le cas le plus général de matrice de covariances associée aux données de mesure est traité,
mais des cas particuliers importants, donnant lieu à des calculs plus simples, sont décrits en détail.
Pour tous les cas considérés, les méthodes de validation du choix de fonctions d’étalonnage linéaire et
d’évaluation des incertitudes et de la covariance associées aux estimations des paramètres sont données.
La présente Spécification Technique décrit aussi l’utilisation des estimations des paramètres de fonction
d’étalonnage et de leurs incertitudes et covariance associées dans la prédiction d’une valeur de X et son
incertitude-type associée pour une valeur mesurée de Y donnée avec son incertitude-type associée.
NOTE 1 La Spécification Technique ne spécifie pas un traitement général des valeurs aberrantes parmi les
données de mesure, bien que les essais de validation puissent être utilisés comme base d’identification des
données divergentes.
NOTE 2 La Spécification Technique décrit une méthode d’évaluation des incertitudes associées aux données de
mesure lorsque ces incertitudes ne sont connues qu’avec un facteur d’échelle (Annexe E).
2 Références normatives
Les documents de référence suivants sont indispensables pour l’application du présent document. Pour
les références datées, seule l’édition citée s’applique. Pour les références non datées, la dernière édition
du document de référence s’applique (y compris les éventuels amendements).
Guide ISO/CEI 99:2007, Vocabulaire international de métrologie — Concepts fondamentaux et généraux et
termes associés (VIM)
Guide ISO/CEI 98-3:2008, Incertitude de mesure — Partie 3: Guide pour l’expression de l’incertitude de
mesure (GUM:1995)
3 Termes et définitions
Pour les besoins du présent document, les termes et définitions donnés dans le Guide ISO/CEI 98-3:2008
et le Guide ISO/CEI 99:2007 ainsi que les suivants s’appliquent.
Un glossaire des principaux symboles est donné dans l’Annexe G.
3.1
valeur mesurée
valeur d’une grandeur représentant un résultat de mesure
[SOURCE: Guide ISO/CEI 99:2007, 2.10]
3.2
incertitude de mesure
paramètre non négatif qui caractérise la dispersion des valeurs attribuées à un mesurande, à partir des
informations utilisées
[SOURCE: Guide ISO/CEI 99:2007, 2.26]
3.3
incertitude-type
incertitude de mesure exprimée sous la forme d’un écart-type
[SOURCE: Guide ISO/CEI 99:2007, 2.30]
3.4
covariance associée à deux valeurs
paramètre qui caractérise l’interdépendance des valeurs attribuées à deux mesurandes, à partir des
informations utilisées
3.5
matrice de covariances de mesure
matrice de covariances
matrice de dimension N × N, associée à une estimation vectorielle d’un vecteur de grandeurs de dimension
N × 1, qui comporte sur sa diagonale, les carrés des incertitudes-types associées aux composantes
respectives de l’estimation vectorielle du vecteur des grandeurs, et dans ses positions hors diagonale, les
covariances associées à des paires de composantes de l’estimation vectorielle du vecteur des grandeurs
Note 1 à l’article: Une matrice de covariances U de dimension N × N associée à l’estimation vectorielle x d’un
x
vecteur de grandeurs X se présente sous la forme
cov,xx cov,xx
 () ()
11  1 N
U =    
x
cov,xx …cov,xx
 () ()
NN1 N
 
où cov(x , x ) = u (x ) est la variance (incertitude-type au carré) associée à x et cov(x ,x ) est la covariance
i i i i i j
associée à x et x . cov(x ,x ) = 0 si les éléments X et X de X n’ont pas de corrélation.
i j i j i j
Note 2 à l’article: Les covariances sont aussi appelées incertitudes mutuelles.
Note 3 à l’article: Une matrice de covariances est aussi appelée matrice de variances-covariances.
[13]
Note 4 à l’article: Définition adaptée du Guide ISO/CEI 98-3:2008, définition 3.11/Suppl.1:2008.
3.6
modèle de mesure
relation mathématique entre toutes les grandeurs qui interviennent dans un mesurage
[SOURCE: Guide ISO/CEI 99:2007, 2.48]
3.7
modèle fonctionnel
modèle statistique prenant en compte des erreurs associées à la variable dépendante
3.8
modèle structurel
modèle statistique prenant en compte des erreurs associées aux variables indépendantes et dépendantes
2 © ISO 2010 – Tous droits réservés

3.9
étalonnage
opération qui, dans des conditions spécifiées, établit en une première étape une relation entre les
valeurs et les incertitudes de mesure associées qui sont fournies par des étalons et les indications
correspondantes avec les incertitudes associées, puis utilise en une seconde étape cette information
pour établir une relation permettant d’obtenir un résultat de mesure à partir d’une indication
Note 1 à l’article: Un étalonnage peut être exprimé sous la forme d’un énoncé, d’une fonction d’étalonnage, d’un
diagramme d’étalonnage, d’une courbe d’étalonnage ou d’une table d’étalonnage. Dans certains cas, il peut
consister en une correction additive ou multiplicative de l’indication avec une incertitude de mesure associée.
Note 2 à l’article: Il convient de ne pas confondre l’étalonnage avec l’ajustage d’un système de mesure, souvent
appelé improprement «auto-étalonnage», ni avec la vérification de l’étalonnage.
Note 3 à l’article: La seule première étape dans la définition est souvent perçue comme étant l’étalonnage.
[SOURCE: Guide ISO/CEI 99:2007, 2.39]
3.10
loi de probabilité
〈variable aléatoire〉 fonction donnant la probabilité qu’une variable aléatoire prenne toute valeur donnée
ou appartienne à un ensemble donné de valeurs
Note 1 à l’article: La probabilité pour un ensemble de valeurs de la variable aléatoire est égale à 1.
Note 2 à l’article: Une loi de probabilité est dite à une variable lorsqu’elle est relative à une seule variable aléatoire
(scalaire), et à plusieurs variables lorsqu’elle est relative à un vecteur de variables aléatoires. Une loi de probabilité
à plusieurs variables est aussi décrite comme distribution combinée.
Note 3 à l’article: Une loi de probabilité peut prendre la forme d’une fonction de répartition ou d’une fonction de
densité de probabilité.
Note 4 à l’article: Définition et note 1 adaptées de l’ISO 3534-1:1993, définition 1.3 et du Guide ISO/CEI 98-3:2008,
[13]
définition C.2.3; notes 2 et 3 adaptées du Guide ISO/CEI 98-3:2008, définition 3.1/Suppl.1:2008.
3.11
loi normale
loi de probabilité d’une variable aléatoire continue X avec pour fonction de densité de probabilité
 
1 1 ξμ−
 
 
g×=()ξ exp −
 
2 σ
σπ2
   
 
pour −∞ < ξ < +∞
Note 1 à l’article: μ est l’espérance mathématique et σ est l’écart-type de X.
Note 2 à l’article: La loi normale est aussi appelée loi de Gauss.
Note 3 à l’article: Définition et NOTE 1 adaptées de l’ISO 3534-1:1993, définition 1.37; NOTE 2 adaptée du
Guide ISO/CEI 98-3:2008, définition C.2.14.
3.12
distribution t
loi de probabilité d’une variable aléatoire continue X avec pour fonction de densité de probabilité
v+1
−+(/v 12)
Γ
  2
 
2 ξ
 
g×=()ξ 1+
 
 
v
πvvΓ(/2)
 
pour −∞ < ξ < +∞, avec le paramètre ν, un entier positif, les degrés de liberté de la loi, où

zt−−1
Γ ()zt=>etd, z 0,

est la fonction gamma
[SOURCE: Guide ISO/CEI 98-3:2008, 3.5/Suppl.1:2008]
3.13
loi de chi deux
distribution χ
loi de probabilité d’une variable aléatoire continue X avec pour fonction de densité de probabilité
(/v 21)−
ξξ
g×=()ξ exp(− )
v/2
22Γ(/v )
pour 0 ≤ ξ < ∞, avec le paramètre ν, un entier positif, où Γ est la fonction gamma
Note 1 à l’article: La somme des carrés des variables normales centrées réduites indépendantes ν est une variable
aléatoire χ ayant pour paramètre ν; ν est alors appelé degrés de liberté.
3.14
matrice définie positive
T
matrice M de dimension n × n ayant la propriété z M z > 0 pour tous les vecteurs z non nuls de
dimension n × 1
3.15
matrice semi-définie positive
T
matrice M de dimension n × n ayant la propriété z M z ≥ 0 pour tous les vecteurs z non nuls de
dimension n × 1
4 Conventions et notation
Pour les besoins de la présente Spécification technique, les conventions et notations suivantes sont adoptées.
4.1 X est appelé variable indépendante et Y, variable dépendante même lorsque le contenu de X et Y
est « interchangeable », comme décrit dans l’Article 7, par exemple.
4.2 Les grandeurs A et B sont appelées les paramètres de la fonction d’étalonnage linéaire Y = A + BX.
A et B sont aussi utilisées pour désigner des variables (virtuelles) dans les expressions impliquant les
paramètres de la fonction d’étalonnage.
4.3 Les grandeurs X et Y sont utilisées comme variables (virtuelles) pour désigner les coordonnées
i i
ème
du i point de données.
* *
4.4 Les constantes A et B sont des valeurs (inconnues) de A et B qui caractérisent la fonction
* *
d’étalonnage linéaire Y = A + B X pour un système de mesure spécifique considéré.
**
4.5 Les constantes XY et sont les coordonnées (inconnues) du ième point de données fournies
ii
**
par le système de mesure satisfaisant à YA=+**BX .
ii
4.6 x et y sont les valeurs mesurées des coordonnées du ième point de données.
i i
4.7 a et b sont les estimations des paramètres de la fonction d’étalonnage relatives au système de mesure.
** * *
4.8 xy et sont les estimations des coordonnées du ième point de données satisfaisant à ya=+bx .
ii ii
4.9 Un vecteur de dimension m × 1 est ainsi noté:
4 © ISO 2010 – Tous droits réservés

x
 1
Τ
xx= …,,= xx
  []
1 m
x
 m
 
et une matrice de dimension m × n est ainsi notée:
a a a a
 
 11 1n  11 m1
Τ
AA= , = 
    
a …a a …a
 mm1 n  1nmn
   
Les dimensions du vecteur ou de la matrice sont toujours spécifiées pour éviter toute possibilité de confusion.
4.10 T désigne une transposée.
4.11 La matrice nulle est désignée par 0 et le vecteur unitaire est désigné par 1.
4.12 Certains symboles ont plusieurs significations. Le contexte en clarifie l’usage.
4.13 Les nombres affichés dans les tableaux avec un nombre fixe de décimales sont des représentations
correctement arrondies de nombres enregistrés avec une plus grande précision, comme c’est le cas
dans un tableur, par exemple. Par conséquent, des incohérences mineures peuvent apparaître entre les
affichages des sommes pour une colonne et l’affichage nombres sommés présents dans cette colonne.
4.14 Dans certains tableaux, un numéro de paragraphe au-dessus d’une ou de plusieurs colonnes
indique où se situe la formule de détermination des valeurs en dessous.
4.15 Dans les exemples, alors que les valeurs de données sont fournies avec une précision donnée,
les résultats des calculs sont fournis avec une plus grande précision afin de permettre à l’utilisateur de
comparer les résultats lorsqu’il effectue les calculs.
5 Principes de l’étalonnage linéaire
5.1 Généralités
5.1.1 Le présent article traite de la manière dont une relation Y = A + BX qui explique la variable
dépendante Y (également appelée « réponse ») en fonction de la variable indépendante X (également
appelée « stimulus ») peut être déterminée à partir de données de mesure. Lors de l’étalonnage, des
données de mesure sont obtenues lorsqu’un instrument de mesure caractérisé par une fonction
* *
d’étalonnage ayant des valeurs (inconnues) A et B est « stimulé » par des objets de valeurs étalonnées
X , exprimées en unités normalisées et les «réponses» correspondantes ou indications Y de l’instrument
i i
sont enregistrées. La relation fournit la réponse Y du système pour un objet de la grandeur étalon X. Ce
processus est appelé « prédiction directe ». D’une façon plus pratique, cette relation permet aussi de
convertir une réponse mesurée y de Y en une estimation x, en unités normalisées de la propriété X d’un
objet. Ce processus est appelé « prédiction inverse » ou « prédiction ».
5.1.2 Il conviendra que l’étalonnage d’un système de mesure tienne compte des incertitudes de mesure,
et, s’il en existe, des covariances entre les données de mesure. L’élément de sortie d’une procédure
d’étalonnage est une fonction d’étalonnage à utiliser pour la prédiction (et si nécessaire, pour une
évaluation de prédiction). Cet élément de sortie comprend également les incertitudes-types et la covariance
associées aux estimations de a et b des paramètres décrivant la fonction d’étalonnage, qui sont utilisées
pour évaluer les incertitudes-types associées à la prédiction inverse (et à la prédiction directe).
5.2 Éléments d’entrée pour la détermination de la fonction d’étalonnage
5.2.1 Données de mesure
Les informations nécessaires à la détermination de la fonction d’étalonnage linéaire sont les données de
mesure, leurs incertitudes-types et les covariances associées. Dans la présente Spécification technique, les
données de mesure sont désignées par (x , y ), i = 1,…, m, c’est-à-dire, les m paires de valeurs mesurées de X
i i
et Y. On considère que m est au moins égal à deux et que les valeurs de x ne sont pas toutes égales entre elles.
i
NOTE Les incertitudes associées aux estimations de a et b diminuent généralement en fonction de
l’augmentation de m. Par conséquent, il convient que l’étalonnage ait pour objet d’utiliser autant de points de
données mesurés qu’il est possible d’exploiter.
5.2.2 Incertitudes et covariances associées
Les incertitudes-types associées à x et y sont désignées respectivement par u(x ) et u(y ). La covariance
i i i i
entre x et x est désignée par cov(x , x ). De même, celles entre y et y , et entre x et y , sont désignées
i j i j i j i j
respectivement par cov(y , y ) et cov(x , y ). L’Annexe D indique la manière dont les incertitudes et les
i j i j
covariances associées à la réponse mesurée et aux variables de stimulus peuvent être évaluées et donne
une interprétation de ces informations d’incertitude. L’ensemble des informations d’incertitude sont
organisées sous la forme des éléments d’une matrice U de dimension 2m × 2m comportant les variances
2 2
(incertitudes-types au carré) u (x ) et u (y ) et les covariances:
i i
 
cov,xx cov xxy,
ux () () cov,xy
() ()
1 m 11
1 1 m
 
 
 
 
 
 
cov,xy
cov,xx () cov,xy
() ux  ()
 () m 1 
m 1 mm
m
U =
 
… …
cov,yy cov,yy
() cov,y xx uy ()
() ()
11 1 m
 1 m 1 
 
   
 
… … 2
 
cov,yx
() cov,yx cov,yy
() () uy
m 1 ()
mm m 1
m
 
Pour de nombreuses applications, certaines ou toutes les covariances sont prises égales à zéro (voir 5.3).
NOTE La présente Spécification technique traite des problèmes où u(x ) ou u( y ) sont généralement différentes.
i i
5.3 Détermination de la fonction d’étalonnage
5.3.1 Les éléments d’entrée qui permettent de déterminer la fonction d’étalonnage sont les données de
mesure, leurs incertitudes et éventuellement leurs covariances associées. Compte tenu des paramètres
ème
A et B, les éléments d’entrée peuvent être utilisés pour fournir une mesure de l’écart du i point de
données (x , y ) à la droite Y = A + BX. Les estimations de a et b sont déterminées en minimisant la
i i
somme des carrés de ces écarts, ou par une mesure plus générale si une covariance n’est pas nulle. La
réalisation de ce qui précède dépend de la «structure d’incertitude» associée aux données de mesure.
Cette structure d’incertitude est liée aux réponses aux questions suivantes:
i) Les incertitudes associées aux valeurs mesurées x sont-elles négligeables ?
i
ii) Les covariances associées aux paires de valeurs mesurées sont-elles négligeables ?
5.3.2 Les cas suivants, classés dans un ordre de complexité croissante et selon les réponses aux
questions de 5.3.1, sont pris en compte dans la présente Spécification technique:
a) Seules les incertitudes sont associées aux valeurs mesurées y et toutes les covariances associées
i
aux données sont considérées comme négligeables (Article 6);
b) Les incertitudes associées aux valeurs mesurées x et y et toutes les covariances associées aux
i i
données sont considérées comme négligeables (Article 7);
c) Les incertitudes associées aux valeurs mesurées x et y et seules les covariances sont associées aux
i i
paires (x , y ) (Article 8);
i i
d) Seules les incertitudes sont associées aux valeurs mesurées y et seules les covariances sont associées
i
à y et y (i ≠ j) (Article 9);
i j
e) Le cas le plus général où il existe des incertitudes associées aux valeurs mesurées x et y et des
i i
covariances associées à toutes les paires de valeurs de x , x , y et y (Article 10).
i j k l
6 © ISO 2010 – Tous droits réservés

5.3.3 Pour chacun des cas de 5.3.2, on a à disposition:
a) les données de mesure prescrites et la structure d’incertitude;
b) le modèle statistique correspondant;
c) le type de moindres carrés traité;
d) les étapes de calcul;
e) les propriétés du modèle statistique;
f) la validation du modèle;
g) l’organisation des calculs pour l’ordinateur, le cas échéant;
h) un algorithme numérique, le cas échant; et
i) un ou plusieurs exemples.
5.4 Traitement numérique
L’Annexe C donne une approche générale du cas le plus général e) de 5.3.2. Cette approche peut être
utilisée pour traiter tous les autres cas, et utilise des méthodes numériques stables et complexes. Les cas
a) à c) de 5.3.2 peuvent cependant être traités avec des opérations élémentaires qui peuvent être mises en
œuvre dans un tableur par exemple. Les cas d) et e) de 5.3.2 nécessitent certaines opérations matricielles,
qui sont simples et directes à mettre en œuvre avec un langage informatique prenant en charge une
arithmétique matricielle; ces opérations ne sont cependant pas très appropriées aux calculs sur tableur.
5.5 Incertitudes et covariances associées aux paramètres de la fonction d’étalonnage
5.5.1 Dans tous les cas considérés, les estimations des paramètres de la fonction d’étalonnage peuvent
être exprimées (explicitement ou implicitement) comme des fonctions des données de mesure. Les
principes du GUM [Guide ISO/CEI 98-3:2008] peuvent être appliqués pour dériver les incertitudes et
les covariances associées aux données mesurées à l’aide de ces fonctions pour obtenir celles qui sont
associées aux estimations des paramètres. Ainsi, les données mesurées sont utilisées pour fournir
les estimations de a et b des paramètres de la fonction d’étalonnage, et pour évaluer les incertitudes-
types u(a) et u(b) et la covariance cov(a, b) associée à ces estimations. Dans les cas a) et d) de 5.3.2, la
propagation des incertitudes est exacte puisque les estimations des paramètres peuvent être exprimées
comme des combinaisons linéaires des entrées y . Dans les autres cas où les estimations des paramètres
i
ne peuvent pas être exprimées de cette manière, la propagation est approximative, sur la base de la
linéarisation relative aux estimations des paramètres. Pour bon nombre de besoins, l’approximation due
à la linéarisation sera suffisamment exacte.
NOTE Lorsque la propagation des incertitudes est approximative, notamment si les incertitudes impliquées
sont grandes (par exemple, dans certains domaines de mesures biologiques), une approche basée sur la propagation
des distributions peut être utilisée. Cette approche [Guide 98-3:2008/Suppl.1:2008] utilise la méthode de Monte
Carlo (ne faisant pas l’objet de la présente Spécification technique).
5.5.2 Les principaux éléments de sortie dans le cadre de la description de la fonction d’étalonnage
linéaire sont le vecteur a des estimations des paramètres de dimension 2 x 1 et la matrice des covariances
U de dimension 2 x 2 donnés par
a
 
ua cov,ab
() ()
a
aU= , =   (1)
a
  2
b
  cov,ba ub 
() ()
 
où u(a) et u(b) sont les incertitudes-types associées à a et b, respectivement, et cov(a, b) = cov(b, a) est la
covariance associée à a et b.
5.6 Validation du modèle
5.6.1 Pour la détermination des estimations des paramètres de la fonction d’étalonnage linéaire a et b,
on considère que le modèle Y = A + BX est valide et que les incertitudes associées aux données de mesure
donnent une mesure réaliste de l’écart des données à la droite. Une fois, a et b déterminés, l’écart réel des
points de données à la fonction d’étalonnage de meilleur ajustement peut être comparé à l’écart prévu.
Cette comparaison implique une mesure agrégée des écarts, exprimée sous la forme d’une somme de
ème
carrés χ des m résidus pondérés, le i résidu pondéré étant une mesure de l’écart du ième point de
obs
données à la droite, ou, lorsque la covariance associée au ième point de données (x , y ) est différente de
i i
zéro, une forme plus générale. Si le χ est largement supérieure aux prévisions, sur le plan statistique,
obs
il existe des raisons de remettre en cause la validité des hypothèses du modèle.
5.6.2 D’un point de vue statistique, les données mesurées peuvent être considérées comme des
réalisations de variables aléatoires. Si les lois de probabilité qui caractérisent ces variables aléatoires
étaient connues, il serait en principe possible de déterminer la loi de probabilité de la mesure agrégée
des écarts de 5.6.1. La probabilité pourrait ensuite être calculée selon que le χ , considéré comme
obs
issu de cette loi d’agrégation, a dépassé tout quantile particulier de la loi. Cependant, les informations
relatives à ces grandeurs étant souvent limitées aux valeurs mesurées elles-mêmes et à leurs variances
associées (respectivement considérées comme espérances mathématiques et variances des variables
aléatoires caractérisées par ces lois), les informations nécessaires à la détermination de la loi pour cette
mesure sont insuffisantes. En revanche, la validité est vérifiée en supposant que les lois pour ces
grandeurs sont normales. Sur la base de cette hypothèse, qui est désormais utilisée, au moins pour les
besoins de la validation, la loi pour cette quantité est un χ avec v = m – 2 degrés de liberté. Par
v
2 2
conséquent, la probabilité que le χ dépasse un quantile particulier de χ peut être déterminée (voir
obs v
6.3, 7.3, 9.3, 10.3). Le quantile de 95 % est utilisé.
2 2
NOTE 1 Si le χ dépasse le quantile de 95 % du χ , la fonction d’étalonnage linéaire peut être considérée
obs v
comme ne fournissant pas de données suffisamment explicatives pour les besoins pratiques. Dans ce cas, il convient
de vérifier d’éventuelles erreurs dans les données et les incertitudes associées. Une autre fonction d’étalonnage
peut être utilisée comprenant un polynôme de degré 2 ou supérieur par rapport à X ou une autre forme mathématique;
cette remarque ne relève pas du domaine d’application de la présente Spécification Technique.
NOTE 2 Il est possible que le modèle soit «trop bon» en ce sens que la valeur observée du χ soit sensiblement
obs
inférieure à la valeur attendue. Cette possibilité correspond généralement aux incertitudes associées aux
données de mesure, considérées comme trop grandes, elle n’est donc pas traitée plus en détail dans la présente
Spécification Technique.
5.6.3 Afin d’obtenir autant de valeurs possibles d’un étalonnage, il est souhaitable que les incertitudes
d’entrée soient estimées avant la détermination des paramètres de la fonction d’étalonnage, au lieu de
l’être après l’ajustement des données, avec les incertitudes associées estimées à partir des données ou
déterminées par rapport à un facteur d’échelle. Ce dernier cas est traité dans l’Annexe E.
5.6.4 Si, dans tout cas particulier, la validation du modèle échoue, c’est-à-dire, si le χ dépasse le
obs
quantile de 95 % du χ (voir 5.6.2), il convient de considérer les incertitudes-types u(a) et u(b) et la
v
covariance cov(a, b) calculées (voir 5.5.2) comme n’étant pas fiables, de même que l’incertitude associée
à la valeur prédite (voir 5.7).
5.7 Utilisation de la fonction d’étalonnage
5.7.1 La fonction d’étalonnage est généralement utilisée pour la prédiction (prédiction inverse)
qui permet, pour une estimation de Y et son incertit
...

Questions, Comments and Discussion

Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.

Loading comments...