SIST-TP CEN/TR 16797-1:2015

SLOVENSKI STANDARD
SIST-TP CEN/TR 16797-1:2015
01-oktober-2015
*UDGEHQLSURL]YRGL±2FHQMHYDQMHVSURãþDQMDQHYDUQLKVQRYL1DYRGLOR]D
VWDWLVWLþQRRFHQMHYDQMHGHNODULUDQLKYUHGQRVWLGHO1DþHODLQSUDYLODXSRUDEH
Construction products: Assessment of release of dangerous substances - Guidance on
the statistical assessment of declared values - Part 1: Principles and rules of application
Produits de construction - Evaluation de l'émission de substances dangereuses  Guide
pour l’évaluation de la performance et la vérification de sa constance - Partie 1 :
Principes et règles d’application
Ta slovenski standard je istoveten z: CEN/TR 16797-1:2015
ICS:
13.020.99 Drugi standardi v zvezi z Other standards related to
varstvom okolja environmental protection
91.100.01 Gradbeni materiali na Construction materials in
splošno general
SIST-TP CEN/TR 16797-1:2015 en
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.

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

SIST-TP CEN/TR 16797-1:2015

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

SIST-TP CEN/TR 16797-1:2015

TECHNICAL REPORT
CEN/TR 16797-1

RAPPORT TECHNIQUE

TECHNISCHER BERICHT
August 2015
ICS 03.120.20; 91.100.01
English Version
Construction products: Assessment of release of dangerous
substances - Guidance on the statistical assessment of declared
values - Part 1: Principles and rules of application
Produits de construction - Evaluation de l'émission de
substances dangereuses  Guide pour l'évaluation de la
performance et la vérification de sa constance - Partie 1 :
Principes et règles d'application


This Technical Report was approved by CEN on 16 January 2015. It has been drawn up by the Technical Committee CEN/TC 351.

CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia,
Finland, Former Yugoslav Republic of Macedonia, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania,
Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and United
Kingdom.





EUROPEAN COMMITTEE FOR STANDARDIZATION
COMITÉ EUROPÉEN DE NORMALISATION

EUROPÄISCHES KOMITEE FÜR NORMUNG

CEN-CENELEC Management Centre: Avenue Marnix 17, B-1000 Brussels
© 2015 CEN All rights of exploitation in any form and by any means reserved Ref. No. CEN/TR 16797-1:2015 E
worldwide for CEN national Members.

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
Contents Page
European foreword .3
Introduction .4
1 Scope .5
2 Declared values.5
3 Other principles .6
4 Uncertainty .8
5 Rules of application .8
5.1 Introduction to the rules of application .8
5.2 Rule of application based on assessment by variables . 11
5.2.1 Principles of assessment by variables . 11
5.2.2 Type-testing . 14
5.2.3 Further-testing . 14
5.2.4 Use of existing data and sharing data . 16
5.2.5 Outliers . 17
5.2.6 No-further-testing (NFT) . 17
5.3 Rule of application for products where the test values are significantly below the
declared value (gamma rule) . 18
5.4 Rule of application based on assessment by attributes . 20
5.4.1 Where assessment by attributes should be selected . 20
5.4.2 Type-testing . 21
5.4.3 Further-testing . 21
5.4.4 Use of existing data and sharing data . 22
5.4.5 Outliers . 22
5.4.6 No-further-testing . 22
Bibliography . 24

2

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
European foreword
This document (CEN/TR 16797-1:2015) has been prepared by Technical Committee CEN/TC 351
“Construction products: Assessment of release of dangerous substances”, the secretariat of which is held by
NEN.
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.
CEN/TR 16797, Construction products: Assessment of release of dangerous substances — Guidance on the
statistical assessment of declared values, comprises the following two parts:
— Part 1: Principles and rules of application [the present document];
— Part 2: Technical and statistical background.

3

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
Introduction
The present document provides a brief introduction as to how to declare performance for the potential release,
emission and/or content of dangerous substances from or in construction products and gives the principles
which underpin the acceptance criteria of test results in relation to a declared value. The main rules of
application are introduced, all of which satisfy the given principles.
CEN/TR 16797-2 [1] provides more detailed background and technical explanation together with examples
and the statistical justification for the rules of application. The definitions and abbreviations listed in
CEN/TR 16797-2:2015, Clause 2 also apply to CEN/TR 16797-1:2015. CEN/TR 16797-2:2015, Annex D
contains a model clause and the rules of application introduced in this Part are drafted as normative text that
may be copied into or cited by product standards. A recommended solution is to copy the model clause into
the product standard and specify the rule of application given in CEN/TR 16797-2:2015, Annex D to be used.
This Technical Report was developed on the basis of experience with the control of release into soil and
water. As it is an assessment of data against a declared value regardless of the source of the data, it is the
technical view of CEN/TC 351 that these procedures are also valid for the assessment of emission from
construction products into indoor air and assessment of gamma radiation from construction products.
It is suggested that all product technical committees follow the principles set out in this CEN Technical Report
and it is hoped that all regulators will accept that these principles achieve their objectives with respect to an
acceptable AVCP procedure. The rules of application are examples of the ways in which these principles may
be applied. There is no obligation on a product technical committee to adopt these rules of application and
they are free to determine their own rules of application. The given rules of application may also be used as a
benchmark for assessing alternative rules of application.
If product technical committees and producers could streamline their approaches in a way that could be
accepted by all regulators, it might support a common understanding on the European market and it might
encourage regulators to harmonize their existing different approaches and requirements on reliability and
meaning of performance declarations in legislation and enforcement.
4

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
1 Scope
This Technical Report provides guidance on the statistical assessment of declared values with respect to the
release, emission and/or content of dangerous substances. This Technical Report provides statistically-based
criteria for type-testing (TT), further-testing (FT) and where a product has been shown to be consistent with
measured values for the release, emission or content that are significantly below the declared values, the
point where no-further-testing (NFT) is permitted.
A series of fundamental principles are defined in the present document and two statistical approaches are
defined. The first approach is to use assessment by variables and this approach requires the data to be
normally or log-normally distributed. This approach is recommended as the default option. The alternative
approach based on assessment by attributes is appropriate for data sets that are not normally or log-normally
distributed. The downside to this form of assessment is that more test data are needed for the same level of
reliability. The present document introduces these assessment procedures and CEN/TR 16797-2 provides
more detail and the statistical proof that they satisfy the principles defined in this document. With both of these
approaches the minimum frequency of testing is a function of the distance between the mean value and
declared value and the variability of the data set, i.e. the sample standard deviation.
To reduce the costs of testing, production plants producing a similar product may share data, e.g. be grouping
the product into clusters for statistical assessment of declared values. Rules for the use of clusters are given
in CEN/TR 16797-2.
CEN/TR 16797-2 also contains rules for identifying outliers within a data set and guidance on using tests other
than the reference method for FT.
A list of tasks for product technical committees is given in CEN/TR 16797-2 as is a model clause for including
in product standards and rules of applications that may be cited in the product standard or copied into product
standards.
2 Declared values
Any declared value with respect to the potential release, emission and/or content of dangerous substances
needs to be justified. This justification is based on either:
— the product conforming to the conditions given in the relevant product standard for a declared value/class
based on the without-further-testing concept; or,
— type-testing followed by further-testing at the determined frequency.
Where there is no requirement to carry out a determination, the producer may declare performance for this
characteristic as 'NPD' using the ‘no-performance determined’ option.
The Construction Products Regulation [2] defines the ways in which a declaration of performance may be
made by the producer. The declared value, or declared class, provides a level of release, emission and/or
content that has a low probability of being exceeded in the production. A producer is free to select the value to
be declared. The validity of the declared value is assessed using statistical techniques described in this
CEN Technical Report using a sufficient number of tests according to a standardized test procedure (the
reference test or a combination of tests with the reference test and adequate indirect tests). The declared
value applies on the scale of a batch as defined in the product standard. As it is a numerical value it can,
where required, be compared directly with a limit value in a regulation or specification. If, in those cases, the
declared value is equal to or less than the limit value, the product satisfies the requirement. A product
technical committee is also free to introduce classes (as technical classes), but the upper numerical value
defining these classes has the same technical meaning as a declared value. The lower numerical limit of the
class will be zero.
5

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
Where a product is to be placed on a regulated market, and where the mean value based on the reference
test method is low in relation to the regulatory limit, a producer may benefit from setting the declared value at
the regulatory limit. Doing so will tend to minimize the test frequency and may lead to satisfying the conditions
for no-further-testing (NFT) given in 5.2.6 and 5.4.6. Whereas, a declared value that is significantly lower than
the regulatory limit, and hence much nearer to the mean value, will probably result in a higher test frequency
or even batch testing under these rules of application. On the other hand, setting a high declared value in
order to minimize the test frequency might affect a product's competitiveness in relation to products with a
lower declared value. So the freedom for the producer to select the declared value introduces more flexibility.
It is, however, for the product technical committee to decide whether performance should be declared using
declared values or classes or whether both options are permitted in its product standards.
If the confidence at which the declared value is to be achieved were only to be defined (qualitatively) as ‘a low
probability of being exceeded in the production’, its meaning would be interpreted differently by different
product technical committees and different regulators. In existing legislation and formal enforcement
procedures in different Member States requirements are specified on the reliability of the declared values and
while these specifications are not harmonized, the intentions are usually similar. Therefore, a common, agreed
quantitative, i.e. statistical, definition is necessary and based on existing regulations and experience, the
following is proposed:
Principle 1: The rules of application verify with a confidence of 90 % that the 50th percentile of the production
is less than or equal to the declared value when the scale of declaration is a batch as defined in the product
standard.
Put another way, it should be expected that the average quality of a batch would be equal to or better than the
declared value after taking into account uncertainty (see Clause 4). The criterion of 50th percentile may seem
too relaxed, but in practice it means that products placed on the market will rarely exceed the declared values.
This is explained in detail in CEN/TR 16797-2.
The rules of application described in this document all satisfy this principle and the technical explanation as to
why is given in CEN/TR 16797-2:2015, Clause 8. In the following sections where the term ‘declared value’ is
used, it may be interchanged with the terms 'regulatory class limit' or ‘technical class limit’.
3 Other principles
It is assumed that the reader is familiar with the principles of the Construction Products Regulation and
therefore these principles are not repeated in this Technical Report. There are also a number of principles
associated with issues such as confidence in the test laboratory and rules for enforcement testing, but these
are outside the scope of this Technical Report.
The following principles all relate to the declared values with respect to the potential release, emission or
content of dangerous substances.
Principle 2: The declared value relates to the performance of the product in a reference test procedure.
The appropriate test method will be defined in the product standard.
Principle 3: The test frequency is permitted to vary. The test frequency reduces as the risk of exceeding the
declared value diminishes, e.g.:
— the distance between the mean value and the declared value increases;
— the standard deviation reduces.
Producers should benefit from lower rates of testing where test results show low variability and where:
— declared values are particularly conservative i.e. where the mean value is well below the declared value;
6

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
— and/or the mean value is well below a market's regulatory limit.
This variable test frequency therefore acts as an incentive to producers to control their products and reduce
the environmental impact.
Principle 4: The production is split into batches in order to facilitate the variable test frequency. For
continually produced products, the batch size associated with Principle 1 is not more than one tenth (10 %) of
the production over one year and the maximum batch size needs to be defined by the product technical
committee.
For continuous production, it may be simpler to split the year into 12 batches; each month of production
representing one batch. The maximum test frequency is one test per batch but in most cases the test
frequency is significantly less (down to one batch per three years). While the maximum batch size needs to be
specified in the product standard, the producer may benefit from using this maximum size during random
testing and using a smaller batch size during batch testing. During batch testing a smaller batch size helps
speed the end of batch testing and reduces the quantity of product at risk if it is declared as non-conforming.
Products conform to the declared value until non-conformity is shown by the test results in combination with a
statistical evaluation detailed in this report.
Principle 5: For one product and intended use there is a single reference test method. In the case of dispute
the reference method has precedence.
This principle does not stop the use of alternative test methods (‘indirect tests’) for further-testing, but they
need to be correlated to, or a safe relationship established with, the reference method, see
CEN/TR 16797-2:2015, Clause 10. From the point at which the reference methods are introduced, type-
testing of 'new' products will require their use; however, see 5.2.4 and 5.4.4 for the discretionary use of
existing test data.
Principle 6: The assessment approach is allowed for products in a production that have a normal, more or
less known, variation of release, emission or content. If the factory production control expects that a change in
production or materials might lead to products outside the normal variations, a separate assessment
procedure should be started.
The assessment approach for declared values needs to include factory production control systems on the
input materials and processing. In some cases the dangerous substances are part of the raw materials used
in production and the producer has little control over the content of dangerous substances. In other cases,
selection and processing can significantly control the level of dangerous substances. The assumption on
which the assessment of conformity is based is that the product is homogeneous, i.e. its variability is
controlled within limits, and the factory production control system needs to ensure that this assumption is
valid. If, for example, raw materials are used with much higher content of dangerous substances that might
influence release from the final product, this should be assessed as a different product.
Principle 7: For products from specific sources where the mean value is well below the producers declared
value, a point may be reached where no-further-testing (NFT) is needed to fulfil Principle 1. Assessment of
NFT verifies with a confidence of 99 % that the 90th percentile of the production is less than or equal to the
declared value when the scale of declaration is a batch as defined in the product standard.
Statistically it makes no sense to continue to test a product that has been shown to consistently meet the
declared value by a very large margin; however, NFT is only valid if the product does not change and the
factory production control needs to ensure that the product has not changed.
The NFT procedure is different to the ‘without-further-testing’ (WFT) procedure. The WFT procedure is a
generic way for declaring release, emission or content. If a product conforms to certain rules defined in the
product standard, a given declared value/class may be stated without testing by the producer for release,
7

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
emission or content. The use of the WFT procedure to declaring a value/class is based on a dossier of
historical test data for a generic product that has been approved by the European Commission. On the other
hand declaring a value based on the NFT procedure is specific to a manufacturer’s particular product and it is
a technical process based on the assessment of previous test results, either using the reference method or a
correlated alternative method, and conditional on the product not changing.
4 Uncertainty
There is always some uncertainty in the relationship between a test result and the true mean value of the
product it is representing. There are two main sources of uncertainty and neither can be eliminated entirely.
The first is the uncertainty associated with the sample being the true average quality of the product being
represented. With certain products, e.g. an aggregate, it is practical to take a series of spot samples over the
period of production, combine the samples and take a representative sample from this combined sample.
Such a procedure will minimize sampling uncertainty. With other products it is not always possible to combine
samples sufficiently, e.g. windows or doors, and in these cases the uncertainty associated with sampling is
likely to be greater. This sampling uncertainty can be reduced by testing more than one sample in separate
tests, or include several items in one test procedure. (e.g. 5 or more bricks in one tank leaching test, or
several wooden panels in an emission test chamber.) The product technical committee is expected to define
the minimum sample size and number of increments. If the product is variable and the release close to the
declared value, testing more than the minimum number of increments may reduce the variability of successive
test results and the frequency of testing.
The other source of uncertainty is associated with the test procedure. When a test is repeated using identical
material, a range of results is obtained. This range is greater when there is more than one operator using
different sets of equipment of the same type. It is normal for a European test standard to include the
uncertainty associated with the test procedure. This is usually given as the repeatability and reproducibility of
the test procedure. There is a 95 % probability that the true mean of the sample will be between ± R of the test
result where R is the reproducibility limit for the test. The lower the values of repeatability and reproducibility
the more precise is the test.
When testing a specific batch for conformity:
— where the test result is just below the declared value, not taking account of uncertainty will lead to this
batch being accepted whereas it might truly be non-conforming (this is called the consumer’s risk); or,
— where the test result is just above the declared value, not taking account of uncertainty will lead to this
batch being rejected whereas it might truly be conforming (this is called the producer’s risk).
As these risks to the consumer and producer balance out, no allowance for the uncertainty of sampling and
measurement is given when assessing a single batch for conformity.
NOTE The uncertainty of measurement is taken into account for enforcement purposes in the Netherlands and
Finland. In these cases all the product under investigation is regarded as a single ‘batch’.
Uncertainty will be reflected in successive test results and lead to the spread of the test results being larger
with higher uncertainty. It is therefore taken into account when random testing, i.e. not testing every batch.
5 Rules of application
5.1 Introduction to the rules of application
The flowchart in Figure 1 provides a guide to the choice of the appropriate rule of application.
The relevant product technical committee needs to select the appropriate rule(s) of application. All the rules of
application given in this Technical Report fulfil the principles given in Clauses 2 and 3.
8

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
All procedures for the analysis of test data use statistical techniques, but it is not necessary at the production
level to understand the statistical detail or background; the basics that are needed is the conscientious
application of the ‘rules’.
The model clause given in CEN/TR 16797-2:2015, Annex D requires the expected production to be split into
batches. For continual production, a batch needs to be no more than 1/10 of a year’s production. It is simpler
to split the year into 12 batches; each month of production representing one batch.
9

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)

Figure 1 —Flowchart on the selection of the rule of application
The term ‘assessment’ is used as the procedure described herein is more than an analysis of data. The
purpose of the assessment is to verify that the declared value is technically justified.
10

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

SIST-TP CEN/TR 16797-1:2015
CEN/TR 16797-1:2015 (E)
5.2 describes the main rule of application and it is based on assessment by variables. Assessment by
variables is a technique that uses the mean value of the last n consecutive results and the spread of these
results around the mean value. This is the main and recommended option, as it will require the minimum
amount of testing to fulfil Principle 1. This rule of application applies when the distribution of test results is
known and is normally or log-normally distributed.
Where a product technical committee knows that the release or emission is low in relation to what will be the
declared value and the variability in test results is dominated by the test procedure, they may specify type-
testing using the rule of application given in 5.3 (only low values – ‘Gamma rule’) in place of the procedure
given in 5.2. Once there are five test results, it is necessary to revert back to the rule of application given in
5.2. At present the rule of application given in 5.3 only applies to data from leaching tests. Potentially the
technique could also be used for emission into indoor air and ionizing radiations, but at the time of drafting this
Technical Report the essential statistical parameter, the coefficient of variation of the test data, is unknown.
This needs to be known before the correct criteria can be determined.
5.4 describes a rule of application based on the assessment by attributes. Assessment by attributes permits a
low proportion of test results to exceed the declared value within a defined number of consecutive test results.
The disadvantage of assessment by attributes is that a higher number of test results are needed to fulfil
Principle 1.
The rule of application given in 5.4 should be adopted where one of the following applies:
— the production process may substantially affect the test result; or,
— the distribution of tests results is unknown.
These two conditions may lead to the situation where the data are not normally or log-normally distributed and
so it is not technically sound to apply the rules of application given in 5.2 and 5.3.
The approach and calculations need to be executed separately for each of the substances for which the
performance will be declared. This may lead to different required test frequencies for different substances for
the same product.
To fulfil the conditions for testing a product under the rules of application given in 5.2 to 5.4 (or any other rule
of application), the product needs to be well defined a
...