ISO 14385-2:2014

INTERNATIONAL ISO
STANDARD 14385-2
First edition
2014-08-01
Stationary source emissions —
Greenhouse gases —
Part 2:
Ongoing quality control of automated
measuring systems
Émissions de sources fixes — Gaz à effet de serre —
Partie 2: Contrôle qualité continu des systèmes de mesurage
automatiques
Reference number
©
ISO 2014
© ISO 2014
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested 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 2014 – All rights reserved

Contents Page
Foreword .iv
Introduction .v
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviations . 2
4.1 Symbols . 2
4.2 Abbreviations . 2
5 Principle . 3
5.1 General . 3
5.2 Limitations . 3
5.3 Measurement site and installation . 4
5.4 Testing laboratories performing SRM measurements . 4
6 Ongoing quality assurance during operation . 4
6.1 General . 4
6.2 Procedures to maintain ongoing quality . 5
6.3 Choosing control charts . 6
6.4 Setting parameters for control charts . 6
6.5 Zero and span measurements . 8
6.6 Documentation of control charts . 9
6.7 Check on validity of measured values .10
7 Annual surveillance test (AST) .10
7.1 Functional test .10
7.2 Parallel measurements with an SRM .10
7.3 Data evaluation .12
7.4 Calculation of variability .13
7.5 Test of variability and validity of the calibration function .13
7.6 AST report .14
8 Documentation .14
Annex A (normative) AST functional test of AMS .15
Annex B (normative) Test of linearity .19
Annex C (informative) Documentation .21
Annex D (informative) Shewhart control charts .23
Annex E (informative) Exponentially weighted moving average (EWMA) charts .26
Annex F (informative) Example of calculation of the standard deviation σ of the AMS at zero
AMS
and span level .29
Bibliography .32
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.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www.iso.org/directives).
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. Details of
any patent rights identified during the development of the document will be in the Introduction and/or
on the ISO list of patent declarations received (see www.iso.org/patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the meaning of ISO specific terms and expressions related to conformity
assessment, as well as information about ISO’s adherence to the WTO principles in the Technical Barriers
to Trade (TBT) see the following URL: Foreword - Supplementary information
The committee responsible for this document is ISO/TC 146, Air quality, Subcommittee SC 1, Stationary
source emissions.
ISO 14385 consists of the following parts, under the general title Stationary source emissions —
Greenhouse gases:
— Part 1: Calibration of automated measuring systems
— Part 2: Ongoing quality control of automated measuring systems
iv © ISO 2014 – All rights reserved

Introduction
The measurement of greenhouse gas emissions (carbon dioxide, nitrous oxide, methane) in a framework
of emission trading requires an equal and known quality of data.
This part of ISO 14385 describes the quality assurance procedures for calibration and ongoing quality
control needed to ensure that automated measuring systems (AMS) installed to measure emissions
of greenhouse gases to air are capable of meeting the uncertainty requirements on measured values
specified by legislation, competent authorities, or in an emission trade scheme.
INTERNATIONAL STANDARD ISO 14385-2:2014(E)
Stationary source emissions — Greenhouse gases —
Part 2:
Ongoing quality control of automated measuring systems
1 Scope
This part of ISO 14385 specifies procedures for establishing quality assurance for automated measuring
systems (AMS) installed on industrial plants for the determination of the concentration of greenhouse
gases in flue and waste gas and other flue gas parameters.
This part of ISO 14385 specifies the following:
— a procedure to maintain and demonstrate the required quality of the measurement results during
the normal operation of an AMS, by checking that the zero and span characteristics are consistent
with those determined using the relevant procedure in ISO 14956;
— a procedure for the annual surveillance tests (AST) of the AMS in order to evaluate a) that it functions
correctly and its performance remains valid and b) that its calibration function and variability
remain as previously determined.
This part of ISO 14385 is designed to be used after the AMS has been accepted according to the procedures
specified in ISO 14956.
This part of ISO 14385 is restricted to quality assurance (QA) of the AMS and does not include QA of the
data collection and recording system of the plant.
2 Normative references
The following documents, in whole or in part, are normatively referenced in this document and are
indispensable for its application. For dated references, only the edition cited applies. For undated
references, the latest edition of the referenced document (including any amendments) applies.
ISO 14385-1, Stationary source emissions — Greenhouse gases — Part 1: Calibration of automated
measuring systems
ISO 14956, Air quality — Evaluation of the suitability of a measurement procedure by comparison with a
required measurement uncertainty
3 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO 14385-1 and the following
apply.
3.1
control chart
graphical presentation of the regular recording of the difference of the reading of an instrument or
measuring system, when measuring the pollutant concentration in a gas with known concentration, and
the nominal value of the pollutant concentration in that gas
4 Symbols and abbreviations
4.1 Symbols
D difference between SRM value y and calibrated AMS measured value ŷ
i i i
average of D
D i
k test value for variability (based on a χ -test, with a β-value of 50 %, for N numbers of paired meas-
v
urements)
N number of paired samples in parallel measurements
S standard deviation of the AMS used in ongoing quality control
AMS
S standard deviation of the differences D in parallel measurements
D i
t value of the t distribution for a significance level of 95 % and a number of degrees of freedom of
0,95; N–1
N – 1
u uncertainty due to instability (expressed as a standard deviation)
inst
u uncertainty due to influence of temperature (expressed as a standard deviation)
temp
u uncertainty due to influence of pressure (expressed as a standard deviation)
pres
u uncertainty due to influence of voltage (expressed as a standard deviation)
volt
u any other uncertainty that can influence the zero and span reading (expressed as a standard
others
deviation)
th
x i measured signal obtained with the AMS at AMS measuring conditions
i
average of AMS measured signals x
x i
th
y i measured value obtained with the SRM
i
average of the SRM measured values y
y
i
y SRM measured value y at standard conditions
i,s i
y lowest SRM measured value at standard conditions
s,min
y highest SRM measured value at standard conditions
s,max
ŷ best estimate for the “true value”, calculated from the AMS measured signal x by means of the
i i
calibration function
ŷ best estimate for the ”true value”, calculated from the AMS measured signal x at standard condi-
i,s i
tions
ŷ best estimate for the ”true value”, calculated from the maximum value of the AMS measured sig-
s,max
nals x at standard conditions
i
ԑ deviation between y and the expected value
i i
standard deviation associated with the uncertainty derived from requirements of legislation
σ0
4.2 Abbreviations
AMS automated measuring system
AST annual surveillance test
2 © ISO 2014 – All rights reserved

EWMA chart exponentially weighted moving average chart
QA quality assurance
SRM Standard Reference Method
5 Principle
5.1 General
An AMS to be used shall have been proven suitable for its measuring task (parameter and composition
of the flue gas) by use of the procedures, as specified by ISO 14956. Using this part of ISO 14385, it shall
be proven that the total uncertainty of the results obtained from the AMS meets the specification for
uncertainty stated in legislation or in requirements and specifications established in an international
trading program. In ISO 14956, the total uncertainty required by the relevant regulations is calculated
by summing all the relevant uncertainty components arising from the individual performance.
NOTE It is advisable that uncertainty figures are provided by independent testing bodies.
This part of ISO 14385 provides two procedures.
— A procedure which is used to check drift and precision in order to demonstrate that the AMS is in
control during its operation so that it continues to function within the required specifications for
uncertainty. This is achieved by conducting periodic zero and span checks of the AMS, based on those
used in the procedure for zero and span repeatability tests according to ISO 14956:2002, and then
evaluating the results obtained using control charts. Zero and span adjustments or maintenance of
the AMS can be necessary, depending on the results of this evaluation.
— A procedure which is used to evaluate whether the measured values obtained from the AMS still
meet the maximum permissible uncertainty criteria, as demonstrated in the calibration procedure
(ISO 14385-1). It also determines whether the calibration function obtained during the calibration
procedure is still valid. The validity of the measured values obtained with the AMS is checked by
means of a series of functional tests, as well as by the performance of a limited number of parallel
measurements using an appropriate SRM.
5.2 Limitations
Figure 2 illustrates the components of the AMS covered by this part of ISO 14385.
Figure 2 — Limits for the QA of the AMS excluding the data acquisition and handling system
NOTE 1 The influence of the uncertainty of the measurement results, which arise from the data acquisition
recording and handling system of the AMS or of the plant system, and its determination, are excluded from this
part of ISO 14385.
NOTE 2 The performance of the data collection and recording system can be as influential as the AMS
performance in determining the quality of the results obtained from the whole measuring system/process. There
are different requirements for data collection recording and presentation in different countries.
When conducting parallel measurements, the measured signals from the AMS are taken directly from
the AMS (e.g. expressed as analogue or digital signal) during the AST procedures specified in this part
of ISO 14385, by using an independent data collection system provided by the organization(s) carrying
out the AST tests. All data shall be recorded in their uncorrected form (without corrections for, e.g.
temperature and oxygen). A plant data collection system with quality control can additionally be used
to collect the measured signal from the AMS.
5.3 Measurement site and installation
The AMS shall be installed in accordance with the requirements of the relevant national or international
standards, as specified by legislation, competent authorities, or in emission trade scheme. Special
attention shall be given to ensure that the AMS is readily accessible for regular maintenance and other
necessary activities.
NOTE The AMS is intended to be positioned as far as practical in a position where it measures a sample,
which is representative of the stack gas composition.
All measurements shall be carried out on a suitable AMS and peripheral AMS installed within an
appropriate working environment.
The working platform used to access the AMS shall readily allow parallel measurements to be performed
using an SRM. The sampling ports for measurements with the SRM shall be placed as close as possible,
but not more than three times the equivalent diameter up-stream or down-stream of the location of the
AMS, in order to achieve comparable measurements between AMS and SRM.
It is necessary to have good access to the AMS to enable inspections to take place and also to minimize
time taken to implement the quality assurance procedures of this part of ISO 14385. A clean, well-
ventilated, and well-lit working space around the AMS is required to enable the staff to perform this
work effectively. Suitable protection is required for the personnel and the equipment, if the working
platform is exposed to the weather.
5.4 Testing laboratories performing SRM measurements
The testing laboratories, which perform the measurements with the SRM, shall have an accredited
quality assurance system according to ISO/IEC 17025 or shall be approved directly by the relevant
competent authority. They shall also have sufficient experience in performing the measurements
using the appropriate SRM. The SRM used shall be an international or national standard to ensure the
provision of data of an equivalent scientific quality.
6 Ongoing quality assurance during operation
6.1 General
An AMS can drift or become less precise during routine operation. Drift or instability can be due to,
for example, changes in the AMS, such as contamination of an optical surface, a gradual failure of a
component, or a blockage in a filter. Such changes cause systematic errors in the data from the AMS.
On the other hand, AMS are also subject to short-term variations in stability and precision due to the
influences of factors such as changes in ambient temperature. These variations cause random errors.
The magnitude of the random errors is assessed during the certification process of the AMS.
4 © ISO 2014 – All rights reserved

After the acceptance and calibration of the AMS, further quality assurance and quality control procedures
shall be followed so as to ensure that the measured values obtained with the AMS meet the stated or
maximum permissible uncertainty on a continuous basis (also described as ongoing quality control). The
implementation and performance of the procedures given in this part of ISO 14385 are the responsibility
of the plant owner (i.e. the owner of the AMS). It is also the responsibility of the plant owner to ensure that
the AMS is operating inside the valid calibration range (see 6.5). The procedures shall be implemented
and be in place at the same time that the collection of emission data by means of the AMS is mandatory
for reporting to the authorities. It is recommended, however, that these procedures commence as soon
as possible after the installation of the AMS in order to gain as much information on the performance of
the AMS as possible. This can begin before the AMS has to be calibrated with the SRM in order to fulfil
the procedure requirements according to ISO 14385-1.
The instrument reading shall reflect the actual drifts in both zero and span readings. Negative instrument
readings at zero level shall be recorded.
For some monitors, it is difficult to achieve a zero and span readings. In those situations, the supplier
shall give instructions on how to achieve readings that reflect the actual drift in zero and span readings,
as demonstrated in the procedures according to ISO 14956, and conforming to the definition of the zero
reading.
6.2 Procedures to maintain ongoing quality
The aim of the procedure is to maintain and demonstrate the quality of the AMS, so that the requirements
for the stated zero and span repeatability and drift values are met during ongoing operation and the AMS
is maintained in the same operational condition as when installed. This shall be achieved by confirming
that the drift and precision determined during the procedures according to ISO 14956 remain under
control. A suitable methodology shall determine the combined drift and precision of the AMS.
The methodology shall identify whether an extra-maintenance (e.g. by the manufacturer) is necessary in
order to adjust the AMS. The procedure uses control charts which plot the drifts (zero and span) against
the time. In this procedure, reference materials are needed. The value of the reference material shall be
known. The drift and precision components obtained from the procedure described in ISO 14956 and
the uncertainty shall be combined and compared against the combined drift and precision obtained in
the field.
Control charts require regular and ideally frequent measurements. The needed frequency of the ongoing
quality control is at least the period of the maintenance interval. In order to extend a maintenance
interval, some AMS suppliers developed automatic checks and adjustments which guarantee very
limited drifts over time. Regular measurements at zero and reference points are the foundations of the
procedure. Using control charts to show trends in the zero and reference point measurements show
each measurement in context and can help prevent the operator from making adjustments to the AMS
only when required.
A frequency of the ongoing quality of at least once every 2 weeks is recommended. Depending on the
results of the zero and span checks, this frequency can be changed.
Therefore, ongoing quality control requires plant operators to have a procedure which describes the
requirements for
— measuring zero and span values,
— plotting these values on control charts, and
— using the control charts to determine whether there are systematic errors, whether the random
errors exceed the acceptable limits established by the implementation requirements in an
international trading scheme.
The following sub-sections describe the following:
— choosing control charts;
— setting parameters for control charts;
— zero and span measurements;
— documentation and interpretation of the control charts.
6.3 Choosing control charts
6.3.1 General
Any type of control chart, manual or automated, can be used. Different charts have different advantages
and can be more or less complicated to use, depending on the type of chart chosen. This part of ISO 14385
describes two types of chart: the Shewhart chart and the EWMA chart.
6.3.2 Shewhart chart
Shewhart charts simply plot the readings and test them against multiples of S . Its advantage is its
AMS
simplicity; its disadvantage is that the approach is not as sensitive as other approaches such as EWMA
charts. Furthermore, Shewhart charts cannot distinguish between systematic errors and random
errors. Shewhart charts only indicate if the AMS has drifted or whether the precision has worsened.
However, the Shewhart chart method is simple to set up and understand, and it is well suited for manual
procedures.
Annex D describes in detail the procedure for Shewhart chart.
6.3.3 EWMA chart
Compared with the Shewhart chart, the exponentially weighted moving average (EWMA) chart is more
appropriate for early detection of small- or medium-sized maladjustments. It keeps the graphical format
of the Shewhart chart. This approach also implements only one decision rule. The approach also reduces
the risks of unnecessary intervention due to the natural variability of the process.
Annex E describes in detail the procedure for EWMA chart.
6.3.4 Built-in methods
An alternative to an external control chart is to use an instrument built-in method. Many instruments
have a built-in check of zero and span points, and give alarm, if set limits are surpassed.
Some AMS equipped with automatic systems for zero and span checks do not ordinarily output the data
for zero and span drift for plotting on control charts, even though the automatic systems are designed to
achieve the same result as control charts, i.e. measuring drift and alerting the plant operator if the AMS
has drifted out of control. Some systems also automatically adjust the zero and/or the span point in order.
If a plant operator has such a system, it can be accepted as a method for ongoing quality control provided
that an assessment of the total drifts and adjustments are possible during the AMS maintenance by the
AMS supplier and that the information is also accessible to the operator and for third party auditing.
6.4 Setting parameters for control charts
6.4.1 Calculation of the standard deviation S using performance data
AMS
The standard deviation S shall be derived from the information obtained for the calculations
AMS
according to ISO 14956. S shall be calculated considering actual plant conditions and not the test
AMS
conditions during the procedures according to ISO 14956.
For example, during establishing performance characteristics of an instrument testing the influence of
ambient temperature on the AMS could be defined in a range such as 5 °C to 40 °C. However, if the AMS
6 © ISO 2014 – All rights reserved

is kept in a climate controlled enclosure where the temperature varies from 18 °C to 23 °C, then the
operator uses a temperature variation of 5 °C in the calculation for S .
AMS
S shall be calculated by
AMS
22 22 2
Su=+uu++uu+ (1)
AMSinsttempvoltpresothers
where
u is the uncertainty from instability;
inst
u is the uncertainty relating from variations in ambient temperature;
temp
u is the uncertainty relating from variations in voltage;
volt
u is the uncertainty relating from variations in ambient pressure;
pres
u is any other uncertainty that can influence the reading on zero and span reference
others
material (e.g. dilution).
NOTE 1 S is expressed as a standard deviation; therefore, all above uncertainties are expressed as standard
AMS
deviations. E.g. if the uncertainties are given as values at 95 % confidence, it is divided by the coverage factor
(k = 2) for the correct calculation of S .
s AMS
NOTE 2 It is advisable that values of uncertainties are provided by independent testing bodies
.
If any of the above uncertainties are time dependent, this shall be taken into account. For example, if the
uncertainty for instability is given as an upper and lower boundary such as a percentage value ± p per q
days, then q equals the time between two readings for the control charts.
Examples of calculation of the standard deviation of the AMS at zero and span level are given in Annex F.
6.4.2 Setting limits for control charts
The control charts required for ongoing quality control are a means of determining whether any zero
and span readings are true outliers, rather than acceptable random variations. As the standard deviation
S determines the positions of the warning and alarm limits, the value of S is a critical part of the
AMS AMS
control chart. In statistical terms, the purpose of S is to determine if there is a significant probability
AMS
that a zero or span measurement is different from the target value. Therefore, S is usually chosen
AMS
to represent one standard deviation of the acceptable variations in zero and span readings. Multiples
of S can be chosen to represent statistical confidence intervals for the variations in zero and span
AMS
readings.
6.4.3 Alternative approach for setting the control chart limits
Site data are necessary to calculate S . They are also needed to check whether the AMS is suitable for
AMS
the monitoring purpose, i.e. to check that the uncertainty of the AMS after start-up at that particular site
is lower than the maximum permissible uncertainty.
However, the collection of site data is sometimes complicated and several assumptions have to be made.
Therefore, instead of using a S value including assumptions, some operators prefer a pragmatic and
AMS
simpler approach consisting in using fixed limits for their control chats. The limits are then equal to the
maximum permissible uncertainty.
Another pragmatic solution is to calculate S based on the specifications from performance testing.
AMS
For example, an AMS can have easily met the requirements. If so, then the value of S based on data
AMS
from test reports can result in relatively low warning and alarm limits. If the performance of the AMS
falls slightly, then the control charts can direct the operator to perform more frequent maintenance than
required, as some variations in performance could mean that the AMS still easily meets the uncertainty
allowances specified. A minimum value of 3 % of the measuring range shall be used for the standard
deviation S . This avoids that the control charts wrongly indicate an out-of-range operation in case of
AMS
very small standard deviations.
6.5 Zero and span measurements
6.5.1 General
Ongoing quality control requires the AMS to have a means to perform zero and span measurements.
For some AMS, the use of test gases is not possible. In this case, surrogate material and/or a procedure
developed by the AMS manufacturer must be used provided they are validated during a certification
process.
To carry out zero and span checks internally in the AMS or in the data recording system, the AMS or
the data recording systems have to be able to record both positive and negative values and record zero
and span data results for a time period longer than one year to enable auditing of the data during the
periodic check of the AST or during a new calibration procedure as described in ISO 14385-1.
Some AMS have been designed to perform automatic zero and span measurements. In order to fulfil
the ongoing quality control requirements, the data from the zero and span measurements needs to be
available to the operator.
The operator of an AMS should be aware that internal standards used to perform automatic zero and
span measurements could fail.
6.5.2 Frequency of zero and span measurements
Operators have to plot zero and span data using control charts. The application of control charts requires
regular and ideally frequent zero and span measurements. The maintenance interval defined during the
performance testing of AMS is used as the minimum frequency for zero and span checks. However, the
operator of the industrial plant can perform more frequent zero and span checks.
Manual zero and span adjustments are only performed if the control chart used indicates a need for
manual adjustment.
The maximum allowable interval between zero and span measurements is known as the maintenance
interval. The maintenance interval is specified by the manufacturer or determined during performance
testing for approval to the requirements set by legislation. In most AMS, the maintenance interval is
typically between 8 d and 1 month. Some AMS have much longer maintenance intervals; for example,
from 3 to 6 months. The benefit of such AMS is that they have a proven long-term stability. Furthermore,
as they do not require frequent span measurements, this means that the AMS have a higher availability
for monitoring, as span measurements can be time consuming. This is most of the time achieved thanks
to appropriate internal checks and adjustments during normal operation. When generating appropriate
alarms, it can avoid the risk related to infrequent zero and span measurements of not detecting a
systematic error in the AMS or an increase in random errors.
6.5.3 Extractive gas analysis systems
In simple terms, there are two ways to perform zero and span measurements on AMS with extractive
sampling systems.
— Use of test gases: The exact concentration is not as important as the stability of the test gas.
Nitrogen or ambient air without measurement components can be used as zero gas. If the sampling
line serving the extractive AMS is relatively long, then the zero and span procedures can be time
consuming and can consume relatively large amounts of gas and reduce the availability of the AMS.
In this case, the test gas could be injected directly in front of the sampling conditioning system.
— Use of other reference materials, e.g. gas-filled cuvettes or filter devices within the AMS: The
drawback of these reference materials is that they do not allow a check of the complete AMS, i.e. they
8 © ISO 2014 – All rights reserved

allow zero and span readings in the analyser alone and not through the complete sampling system.
However, periodic manual checks by injecting gases through the sampling system can be used to
verify the effectiveness of these reference materials for zero and span measurements.
6.5.4 In situ and cross-stack gas-monitoring AMS
There are two options for performing zero and span checks on in situ and cross-stack AMS.
— Use of test-gases: Cross-stack AMS can include a sintered tube which encloses the optical path of the
AMS. Such tubes are similar to the sintered tubes which enclose the optical components of in situ
AMS. In both cases, test gases can be used to perform zero and span checks. However, the sintered
tubes have a relatively large volume when compared to the optical benches of most extractive AMS.
This means that they have two drawbacks, a) the tubes require a large volume of gas for each test,
and b) the time required can be longer than that required to perform zero and span checks on an
extractive AMS.
— Use of gas-filled cuvettes and filters: If possible, it is good a practice to perform periodic manual
checks by injecting gases through the sampling system, which can be used to verify the effectiveness
of cuvettes for zero and span measurements.
6.5.5 Automatic zero and span checks
Many types of AMS now have built-in systems which automatically perform zero and span checks.
Such systems can include extra functionality to meet all the requirements for ongoing quality control.
However, some systems only warn the operator if the AMS drifts out of control and hence requires
maintenance. If an AMS has an automatic system for zero and span checks, then these automatic systems
are tested during performance testing.
6.5.6 Replacing gas bottles or other surrogates
When replacing gas bottles, differences in the concentrations of bottles can mislead an operative into
believing that the AMS has drifted. This is because two gas bottles with seemingly identical contents
can produce different readings in an AMS because of the uncertainty of the concentrations. This results
in a step change in the AMS readings when one bottle is changed for another. Hence it is important to
differentiate an account for such step changes, instead of mistaking such changes for drift. Therefore
when changing gas bottles, the following procedure can be used:
a) take at least three span readings with the current gas bottle, and then take an average of the
readings;
b) if the span readings using the current gas bottle show that the AMS has not drifted beyond the
action limits since the last span readings, then go to d);
c) if the AMS has drifted, then carry out any necessary actions to remedy the drift and proceed to d);
d) take at least three span measurements using the replacement span-gas bottle, and then set a new
baseline for the control-chart span-level using an average of the three measurements.
Sets of readings with an existing bottle, followed by an equal number of readings with a second bottle,
establish the magnitude of any step-change.
6.6 Documentation of control charts
The control chart calculations shall be performed according to the requirements of this part of ISO 14385
and fully documented. If the zero and span checks are performed automatically with or without
automatic adjustments, the check values shall be recorded in the plant computer or in a separate media
(e.g. a spread sheet) in an auditable manner.
NOTE The use of a spread sheet is very suitable for the calculation procedure given above. At the same time,
[3]
the sheet can be used for providing traceable documentation on the performance of the AMS.
6.7 Check on validity of measured values
The requirement that measured values are within the valid calibration range shall be evaluated by
the plant owner on a weekly basis (Monday to Sunday). A full new calibration (ISO 14385-1) shall be
performed, reported, and implemented within 6 months, if any of the following conditions occur:
— >5 % of the number of AMS measured values calculated over this weekly period (based on
normalized calibrated values) are outside the valid calibration range for more than 5 weeks in the
period between two ASTs.
— >40 % of the number of AMS measured values calculated over this weekly period (based on
normalized calibrated values) are outside the valid calibration range for one or more weeks.
NOTE ISO 14385-1 allows the extension of the valid calibration range using reference materials within
specified constraints.
7 Annual surveillance test (AST)
7.1 Functional test
The first part of the AST is the functional test, which shall be performed according to Annex A. The
functional test shall be performed by an experienced testing laboratory, which has been recognized by
the competent authority.
7.2 Parallel measurements with an SRM
During AST, at least five parallel measurements with an SRM shall be performed. This shall be carried
out according to the procedure described in ISO 14385-1:2014, 6.4. The purpose of comparison
measurements is to verify if the calibration function of the AMS is still valid and if the precision of the
AMS is still within the required limits. If this is the case, and if these measurements include results
outside the valid calibration range, the valid calibration range can be increased with use of these results.
The sequence of the test of the validity of the calibration function and the variability test is described in
Figure 3.
10 © ISO 2014 – All rights reserved

See 7.2
See 7.3
See 7.4
See 7.5
See 7.6
Figure 3 — Flow diagram for the calibration and variability test
Examples of calculation of the calibration function and performance of the variability test in the AST are
given in Annex F.
The evaluation shall be based on a minimum of five valid measurements within the calibration range.
These measurements shall be uniformly spread over the whole measuring day (as described in 6.4).
A set of measurements is valid when all the requirements below are fulfilled:
— the SRM measurements are performed according to the appropriate standard;
— the SRM measurements fulfil all the requirements given in the appropriate standard;
— the time period of each AMS measured signal is larger than 90 % of the averaging time [excluding
the signals which are above 100 % or below 0 % of the measuring range of the AMS, signals obtained
during internal checks (auto calibration), and signals obtained during any other malfunctioning of
the AMS].
The sampling time per measurement shall be the same as used during the initial calibration as described
in ISO 14385-1:2014, 6.4.
The sampling time for the parallel measurements shall be at least 30 min or at least four times the
response time of the AMS, including the sampling system (as determined in the procedures according
to ISO 14956), whichever is the greater. In general, it is recommended that the sampling time used be
the shortest averaging time, which is related to the legislation or to the requirements and specifications
established in an international trading program.
If the sampling time is shorter than 1 h, the time interval between the start of each sample shall be
longer than 1 h.
During the performance of the parallel measurements with the AMS and SRM, each result is a
measurement pair (one AMS measured signal and one SRM measured value) and these shall cover the
same time period.
The results obtained from the SRM shall be expressed in the same conditions as the uncorrected results
obtained from the AMS.
7.3 Data evaluation
The steps for providing data required for performing the variability test and to test the calibration
function are illustrated in Figure 4.
The data sets obtained in the parallel measurements shall be checked for possible outliers. The method
used to assess outliers and reasons for excluding outliers shall be given in the calibration report. Outliers
shall be reported and identified in the calibration diagrams.
This part of ISO 14385 requires at least five valid
...