SIST EN ISO 15796:2008

SLOVENSKI STANDARD
SIST EN ISO 15796:2008
01-oktober-2008
$QDOL]DSOLQRYâWXGLMDLQGRORþDQMHELDVD,62
Gas analysis - Investigation and treatment of analytical bias (ISO 15796:2005)
Gasanalyse - Untersuchung und Behandlung von systematischen Abweichungen (ISO
15796:2005)
Analyse des gaz - Investigation et traitement des biais analytiques (ISO 15796:2005)
Ta slovenski standard je istoveten z: EN ISO 15796:2008
ICS:
71.040.40 Kemijska analiza Chemical analysis
SIST EN ISO 15796:2008 en
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.

---------------------- Page: 1 ----------------------

SIST EN ISO 15796:2008

---------------------- Page: 2 ----------------------

SIST EN ISO 15796:2008
EUROPEAN STANDARD
EN ISO 15796
NORME EUROPÉENNE
EUROPÄISCHE NORM
August 2008
ICS 71.040.40

English Version
Gas analysis - Investigation and treatment of analytical bias
(ISO 15796:2005)
Analyse des gaz - Investigation et traitement des biais Gasanalyse - Untersuchung und Behandlung von
analytiques (ISO 15796:2005) systematischen Abweichungen (ISO 15796:2005)
This European Standard was approved by CEN on 30 July 2008.
CEN members are bound to comply with the CEN/CENELEC Internal Regulations which stipulate the conditions for giving this European
Standard the status of a national standard without any alteration. Up-to-date lists and bibliographical references concerning such national
standards may be obtained on application to the CEN Management Centre or to any CEN member.
This European Standard exists in three official versions (English, French, German). A version in any other language made by translation
under the responsibility of a CEN member into its own language and notified to the CEN Management Centre has the same status as the
official versions.
CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Finland,
France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal,
Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland and United Kingdom.
EUROPEAN COMMITTEE FOR STANDARDIZATION
COMITÉ EUROPÉEN DE NORMALISATION
EUROPÄISCHES KOMITEE FÜR NORMUNG
Management Centre: rue de Stassart, 36  B-1050 Brussels
© 2008 CEN All rights of exploitation in any form and by any means reserved Ref. No. EN ISO 15796:2008: E
worldwide for CEN national Members.

---------------------- Page: 3 ----------------------

SIST EN ISO 15796:2008
EN ISO 15796:2008 (E)
Contents Page
Foreword.3

2

---------------------- Page: 4 ----------------------

SIST EN ISO 15796:2008
EN ISO 15796:2008 (E)
Foreword
The text of ISO 15796:2005 has been prepared by Technical Committee ISO/TC 158 “Analysis of gases” of
the International Organization for Standardization (ISO) and has been taken over as EN ISO 15796:2008 by
Technical Committee CEN/SS N21 “Gaseous fuels and combustible gas” the secretariat of which is held by
CMC.
This European Standard shall be given the status of a national standard, either by publication of an identical
text or by endorsement, at the latest by February 2009, and conflicting national standards shall be withdrawn
at the latest by February 2009.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. CEN [and/or CENELEC] shall not be held responsible for identifying any or all such patent rights.
According to the CEN/CENELEC Internal Regulations, the national standards organizations of the following
countries are bound to implement this European Standard: Austria, Belgium, Bulgaria, Cyprus, Czech
Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia,
Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain,
Sweden, Switzerland and the United Kingdom.
Endorsement notice
The text of ISO 15796:2005 has been approved by CEN as a EN ISO 15796:2008 without any modification.

3

---------------------- Page: 5 ----------------------

SIST EN ISO 15796:2008

---------------------- Page: 6 ----------------------

SIST EN ISO 15796:2008


INTERNATIONAL ISO
STANDARD 15796
First edition
2005-03-01

Gas analysis — Investigation and
treatment of analytical bias
Analyse des gaz — Investigation et traitement des biais analytiques




Reference number
ISO 15796:2005(E)
©
ISO 2005

---------------------- Page: 7 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
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 2005
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 2005 – All rights reserved

---------------------- Page: 8 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
Contents Page
Foreword. iv
Introduction . v
1 Scope. 1
2 Terms and definitions. 1
3 Symbols. 3
4 Bias related to instrumental drift. 4
4.1 Principle. 4
4.2 Stability monitoring. 4
4.3 Drift correction. 6
5 Bias related to effects of sample composition . 10
5.1 Principles. 10
5.2 Local bias handling. 11
5.3 Bias handling for an extended measuring range . 23
6 Treatment of matrix interferences. 27
Annex A (normative)  Critical values for the trend test . 29
Annex B (informative) Uncertainty issues . 30
Bibliography . 33

© ISO 2005 – All rights reserved iii

---------------------- Page: 9 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
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.
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.
ISO 15796 was prepared by Technical Committee ISO/TC 158, Analysis of gases.

iv © ISO 2005 – All rights reserved

---------------------- Page: 10 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
Introduction
Traceability is considered as one of the key items of quality assurance in gas analysis. In general, it is defined
by the existence of unbroken chains of comparisons, relating the analytical result to acknowledged standards
of measurement. More specifically, an analytical result is considered traceable if, by way of these
comparisons, it has been demonstrated to be free of significant bias, significance referring to the specified
uncertainty of the result.
As a rule, traceability is not demonstrated individually for a single analytical result but for a defined analytical
procedure with specified ranges of analyte concentration and matrix composition. An analytical procedure is
considered traceable if it has been demonstrated to be free of significant bias, or if significant bias has been
corrected, by measurement on representative samples of known traceable composition. These may be
samples of appropriate reference gas mixtures. Alternatively, other representative samples may be analysed
in parallel using an accepted reference procedure.
This International Standard provides generic methods for demonstrating, or establishing, traceability of
analytical procedures using reference gas mixtures or reference analytical procedures, implementing
[1] [2]
principles laid out in ISO 14111 and ISO/TS 14167 , and respecting the principles of the Guide to the
[3]
Expression of Uncertainty in Measurement (GUM) .
In this International Standard, the term “concentration” is used for two different purposes:
 as a general term for quantities measured in gas composition analysis, replacing the term “content” (see
[4]
ISO 7504 );
 as a generic substitute for any of the specific quantities measured in gas composition analysis such as
[4]
the mass concentration or the mole fraction of a specified analyte (see ISO 7504 ).

© ISO 2005 – All rights reserved v

---------------------- Page: 11 ----------------------

SIST EN ISO 15796:2008

---------------------- Page: 12 ----------------------

SIST EN ISO 15796:2008
INTERNATIONAL STANDARD ISO 15796:2005(E)

Gas analysis — Investigation and treatment of analytical bias
1 Scope
This International Standard specifies generic methods for detecting and correcting bias (systematic errors) of
analytical procedures for the analysis of gases, using reference gas mixtures or reference analytical
procedures, as well as for estimating the correction uncertainty.
The main sources of (and parameters affecting) bias of analytical procedures are instrumental drift (time) and
matrix interferences (matrix composition). Moreover, bias normally varies with analyte concentration. This
International Standard therefore establishes protocols for
 detecting and correcting drift for an analytical system of limited stability,
 investigating and handling bias of a stable analytical system for a specified range of sample composition,
which are intended to be used in method development and method validation studies, either separately or
sequentially.
This International Standard specifies procedures for two options, applicable to systematic effects, as follows:
a) tracing the observed pattern of deviations and correcting for their effect,
b) averaging over their effects and increasing the uncertainty,
where normally the first option entails lower uncertainty at the expense of higher effort.
For the convenience of the user, the methods specified in this International Standard are described for
procedures of composition analysis, i.e. procedures for measuring the concentration of a specified analyte in a
gas mixture. However, they are equally applicable to measurements of physico-chemical properties of a gas
or gas mixture relevant to gas analysis, and translation into this subject field is straightforward.
2 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
2.1
bias
estimate of systematic error
NOTE Since the true value of a measurand cannot be known exactly, systematic errors cannot be determined exactly
but have to be estimated using reference values.
2.2
correction
procedure by which the uncorrected result of a measurement is adjusted to compensate for systematic error
NOTE 1 Since systematic errors cannot be determined exactly, a correction can never be complete.
[5]
NOTE 2 In the VIM , the term correction is used with a different meaning.
© ISO 2005 – All rights reserved 1

---------------------- Page: 13 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
2.3
uncertainty
parameter, associated with the result of a measurement, that characterizes the dispersion of the values that
could reasonably be attributed to the measurand
[3]
[GUM:1993 , definition 2.3.2]
2.4
traceability
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
[5]
[VIM:1993 , definition 6.10]
[1]
NOTE In ISO 14111 , the term traceability is defined as the ability to provide evidence of the overall accuracy
attributed to measurement results through documented calibrations, using measurement standards of known accuracy
and comparison measurements of known performance.
2.5
reference value
estimate of a quantity, with sufficiently well established traceability and specified uncertainty, used as a
reference for a specified purpose
NOTE In gas analysis, reference values of composition or physico-chemical properties are most often provided by
reference gas mixtures and reference analytical procedures.
2.6
reference gas mixture
calibration gas mixture whose composition is sufficiently well established and stable to be used as a reference
standard of composition from which other composition data are derived
[4]
[ISO 7504:2001 , definition 4.1.1]
2.7
reference analytical procedure
analytical procedure which is capable of providing traceable results with sufficiently well established
uncertainty for use as reference values
2.8
drift
slow change of output, at constant input, of a measuring system
2.9
stability
〈of a measuring system〉 absence of significant drift
2.10
matrix interference
change in analytical response for a specified analyte, caused by variations in matrix composition
2 © ISO 2005 – All rights reserved

---------------------- Page: 14 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
3 Symbols
A, B specified drift-control mixtures
b parameters of a bias correction model
i
δ deviation (from a reference value)
I interferent under consideration
M gas mixture under consideration
m, n, N number of data in a data series
p number of correction parameters
Q recovery (with respect to a reference value)
r number of reference gas mixtures used for bias investigation
R reference gas mixtures used for bias investigation
i
s standard deviation of a data series
s relative standard deviation of a data series (coefficient of variation)
r
2
s variance of a data series
t time
u(x) standard uncertainty of an estimated quantity x
2
u (x) variance of an estimated quantity x
u (x) relative standard uncertainty of an estimated quantity x
r
u(x, y) covariance between two estimated quantities x and y
x, y quantity under consideration
〈x〉 mean value of several quantities x
i
X analyte under consideration
2
∆ mean-square successive difference of a data series
© ISO 2005 – All rights reserved 3

---------------------- Page: 15 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
4 Bias related to instrumental drift
4.1 Principle
This clause specifies procedures for investigating potential drift of an analytical system and taking corrective
actions, if significant drift is encountered.
If the analytical system is expected to be stable, a “drift-control mixture” is measured on a regular basis. For
each specified analyte, the results are recorded on a control chart, and the time series of control data is
examined continuously. As long as these data vary at random within established control limits, the analytical
system has been demonstrated to be stable. Monotonic decrease or increase in control data indicates drift. As
an alternative to visual inspection, a statistical test based on successive differences may be used to detect a
significant trend. As soon as drift becomes significant, e.g. when data exceed control limits, or when a
significant trend is observed, the analytical system is removed from service. After adjustment and re-
calibration, the analytical system is returned to service.
The “drift-control mixture” should contain all analytes currently being measured. Given sufficient information
on the response behaviour of the analytical system, a representative subset of analytes may be used.
If the analytical system is known to exhibit significant drift, it can at least be expected that the drift behaviour of
the system does not significantly depend upon sample composition, since this would allow for drift correction
based on measurements of appropriate “drift-correction mixtures”. This expectation is tested by analysing two
mixtures of distinct composition (different concentrations of the analytes under consideration, different matrix
compositions) on a regular basis. The results are recorded, and for each analyte the two time-series of drift
measurement data are compared in order to investigate whether drift behaviour can be expressed in a
concentration-invariant manner. Given this for every analyte under consideration, the two time-series for each
analyte are pooled to derive a drift correction. In fortunate cases, a joint drift correction may be used for
several or even all analytes.
If the system exhibits significant drift, and if moreover drift characteristics depend upon concentration, drift
correction needs to be integrated with calibration. This topic is beyond the scope of this International Standard.
4.2 Stability monitoring
4.2.1 General considerations
A drift-control mixture is required which is typical of the gases for which the analytical procedure is used. The
composition of the mixture shall be stable, but the concentration of the specified analytes used for stability
monitoring does not have to be established in advance with high accuracy.
An analysis of the drift-control mixture should be carried out with each batch of samples. Its composition is
unvarying so the results of this analysis can be used as an indication as to whether the procedure is no longer
working satisfactorily or re-calibration of the instrument is necessary, or both.
Stability can be monitored using either concentration data or the corresponding response data.
4.2.2 Use of Shewhart control charts
Before first use, the drift-control mixture is analysed at least 10 times in order for precision data to be
calculated. For each specified analyte in the drift-control mixture, the mean concentration (or response) and its
standard deviation are calculated. If the within-day standard deviation is less than the between-day standard
deviation, then these precision data need to be collected one a day for 10 days.
For each specified analyte in the drift-control mixture, a control chart is constructed with points marked on the
Y-axis representing
a) the mean concentration (or response),
4 © ISO 2005 – All rights reserved

---------------------- Page: 16 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
b) the mean ± 1 standard deviation,
c) the mean ± 2 standard deviations (warning limits), and
d) the mean ± 3 standard deviations (action limits).
Lines parallel to the X-axis are drawn from these points. Each time the drift-control mixture is analysed, the
value is plotted using the X-axis as a time scale. As more information becomes available the means and
standard deviations can be updated. This assumes that the analytical system has remained stable. Data
which clearly indicate some fault shall not be used to revise the control limits.
The plotted values from the analysis of the control gas are compared with the mean value and the ± 1, ± 2 and
± 3 standard deviation lines. It is assumed that the composition of the drift-control mixture is stable and that
the analytical results for this follow the normal distribution. If this is true, then while the system is behaving
normally, any individual results for the components of the drift-control mixture may fall outside the warning
limits on 1 occasion in 20. This means that if individual results fall outside the warning limits more than just
occasionally, this can indicate that either there is a systematic tendency for the results to be too high (or too
low) or the random error for measurement of that component has increased. Likewise, individual results may
fall outside the action limits on 3 occasions out of 1 000.
[6]
ISO 8258 contains the following tests which can be used for indicating the presence of variation:
 one point exceeding ± 3 standard deviations;
 nine points in a row on one side of the mean;
 six points in a row steadily increasing or decreasing;
 fourteen points in a row alternating up and down;
 two out of three points in a row exceeding +2 standard deviations or −2 standard deviations;
 four out of five points in a row exceeding +1 standard deviation or −1 standard deviation;
 fifteen points in a row above and below the mean, but not exceeding ± 1 standard deviation;
 eight points in a row above and below the mean, but all exceeding ± 1 standard deviation.
If any of these tests indicates the presence of variation, then that shall be diagnosed and corrected. If this
investigation indicates that there is no fault with the measuring procedure, then re-calibration of the instrument
is required.
4.2.3 Statistical trend test
As an alternative to monitoring monotonic increase or decrease, drift-control data may be investigated for
trends using a statistical test based on successive differences. Given a time-series of drift-control data x , x ,
1 2
2
…, x with mean value 〈x〉, the mean-square successive difference ∆ is determined according to
N
22 2 2

∆=−()xx +(x−x)+.+(x −x)/(N−1) (1)
12 2 3 NN−1

2
This quantity is compared with the variance s given by
22 2 2

sx=()−〈〉x +(x −〈〉x) +.+(x −〈x〉)/(N−1) (2)
12 N

© ISO 2005 – All rights reserved 5

---------------------- Page: 17 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
2 2
If successive values in the series are independent (and moreover from a normal distribution), then ∆ ≈ 2 s . In
2 2
case of a trend ∆ < 2 s because successive values are closer than to be expected for values drawn at
random from a normal distribution.
2 2
For a significance test, the test statistic ∆ /s is compared with the critical value for the specified length N of
the series under investigation and the specified significance level. Values of the test statistic below the critical
value indicate a significant trend. In this International Standard a significance level of 95 % or 99 % is
recommended. Critical values for these significance levels are given in Table A.1, Annex A.
EXAMPLE Consider a series of drift-control data (carbon monoxide in nitrogen, expressed in mmol/mol): 1,28; 1,30;
2 −4
1,30; 1,28; 1,26; 1,24; 1,27; 1,27; 1,24; 1,26. For these data the mean-square successive difference ∆ is 38 × 10 /9,
2 −4 2 2
while the variance s is 40 × 10 /9. Hence the test statistic ∆ /s takes a value of 0,95. For N = 10, the critical value is
0,751 8 for a significance level of 99 % and 1,062 3 for a significance level of 95 %. Therefore, under the assumptions of
independence and normality, the drift-control data exhibit a significant trend on the 95 % level, while the trend is not
significant at the 99 % level.
Consider now the modified data series generated by interchange of the 3rd and 9th datum: 1,28; 1,30; 1,24; 1,28; 1,26;
2 −4
1,24; 1,27; 1,27; 1,30; 1,26. For these data, the mean square successive difference ∆ is now 98 × 10 /9, while the
2 −4 2 2
variance s still is 40 × 10 /9. Hence the test statistic ∆ /s takes a value of 2,45 which means that the modified data
series does not exhibit any indication of drift, under the above assumptions.
For stability monitoring based on regular drift-control measurements, a moving window comprising 10 to 20
data is recommended.
If any of these tests indicates the presence of variation, then that shall be diagnosed and corrected. If this
investigation indicates that there is no fault with the measuring procedure, then re-calibration of the instrument
is required.
4.3 Drift correction
4.3.1 General considerations
This clause specifies a general method for post-processing analytical data to correct for instrumental drift. For
this purpose an analytical system is treated as a “black box”. Here the input is the value of the measurand, i.e.
the (true) concentration of the analyte under consideration in the analysed sample, and the output is the
measured value of this analyte concentration.
This clause is applicable for absolute methods, i.e. analytical methods where analyte concentration is
determined directly, or relative methods where the relationship between measured response and analyte
concentration is known. It is also applicable for comparison methods, i.e. analytical methods where the
relationship between measured response and analyte concentration is determined empirically by calibration.
Two drift-correction mixtures are required which are typical of the gases for which the analytical procedure is
used. Each analyte to be determined by the procedure shall be present in both mixtures, at different levels
bracketing an appropriate concentration range. The composition of the mixtures shall be stable, and the
concentration of the analytes used for drift investigation shall be known with specified uncertainty.
4.3.2 and 4.3.3 specify two complementary approaches, based on additive and multiplicative drift modelling
respectively. Usually only one or neither of these procedures will work. If in a particular case both approaches
work, the one with the better performance, i.e. with lower correction uncertainty, should be used.
NOTE 1 For an analytical comparison method, drift correction is sometimes better performed using measured
responses instead of analyte concentrations.
NOTE 2 Using reference gas mixtures for drift control has the advantage of providing an “absolute” drift correction, i.e.
relative to reference values of analyte concentration. As an alternative, less well-characterized gas mixtures could be used
for drift correction relative to analyte concentrations measured at a specified time t . The latter approach would, in addition,
0
require a proof of stability of the drift-control mixtures. Secondly, the concentrations measured at t would have to be
0
investigated for bias in a later stage.
6 © ISO 2005 – All rights reserved

---------------------- Page: 18 ----------------------

SIST EN ISO 15796:2008
ISO 15796:2005(E)
4.3.2 Additive drift correction
In this subclause, instrumental drift is presented as an additive bias according to Equation (3):
xt =+x δ x,t (3)
...