Version HTML de base
“Uncertainty-noise” Le Mans
19
Acoustique
&
Techniques n° 40
Treatment of Measurement Uncertainties in International and European Standards on Acoustics
3- The expanded measurement uncertainty for a coverage
probability of 95% is then to be taken as two times the
standard deviation of reproducibility, unless more specific
knowledge is available.
In the informative annex to each standard in the series, a
functional relationship is stated between the measurand
(the sound power level or the sound energy level of the
noise source) and all the known parameters, including the
physical quantities and a number of input quantities which
allow for variability in the test results (that is, the sources of
measurement uncertainty). While most of these input quantities
are known relevant to the various test procedures described
in the series of standards, it is admitted that research could
show that there are others. Moreover, the magnitudes and
functional relationships of the input quantities are not known
at present. The annex goes on to list in tabular form the kind
of information which is required to evaluate the magnitudes
of the sources of uncertainty, but because of the present lack
of knowledge, there are no data in the table.
The aim of the ISO Working Group is to conduct extensive
research in order to obtain the missing information, and then
in a further future revision of all the standards in the series
to include the information and so to comply fully with the
requirements of the GUM at that time.
Observations from the Committee Secretariat
It is obvious that it was not until the establishment of the
general policy paper, doc ISO/TC 43 N 1023, that most
Working Groups and project leaders started taking sincere
notice of the requirement for uncertainty statements. And
because standards development is a relatively slow process,
experience with the implications is still limited. Nevertheless,
several reactions have been registered. Some first reactions
were very opposing, arguing that, e.g. in relation to external
noise measurement there are so many unknown influences that
the establishment of an uncertainty budget is impossible. Other
groups have - consciously or not - ignored the requirement,
and, finally, a number of groups have accepted the requirement
and tried to include appropriate uncertainty statements in their
drafts. In all three cases, however, it is obvious that the new
uncertainty requirements have caused delays in the finalization
of documents, some even to the extent that they have not
yet been finalized after the introduction of the uncertainty
policy document.
A great number of documents are still under development
in Working Groups and have not yet been seen by a broader
audience, and it is going to be interesting to see whether
they will contain appropriate uncertainty statements when
they reach the stage of the first circulation to the member
countries.
Despite these problems it is acknowledged that, basically, at
least attempts to implement the uncertainty requirements in
the existing drafts from ISO/TC 43 and SC 1 - and thus also
CEN/TC 211 (see chapter 4) are now being made relatively
readily. It is obvious, however, that this would never have been
accomplished without the personal efforts of the chairman
of ISO/TC 43 and SC 1, first in getting the general policy
paper established, and subsequently in making constructive
comments on all relevant drafts which have been circulated,
and even having direct discussions with some project
leaders.
Formally, the basic criterion for approval of international
standards is the approval of the member countries. It is
therefore worth noting that national comments on drafts also
- slowly - seem to start requesting uncertainty statements
where they are missing. On the other hand it is surprising that
there seems to be no pressure or assistance from the top
level or the Central Secretariat of ISO to include appropriate
uncertainty statements so at the moment the general success
of the need for uncertainty statements in international
standards depends fully on the initiatives and efforts in the
individual committees.
Recent results as seen by the Committee Chairman
Since the adoption of the TC 43 policy paper in September
2003, 35 documents specifying various methods for sound
measurement and prediction were prepared by different
Working Groups and circulated to ISO Member Bodies for
comments. Nearly 90% of them were at a rather early stage
of development (Committee Drafts). The others have already
reached the stage of Draft International Standards. Among the
Committee Drafts, nearly 30% of them were issued twice in
the given period, the second time following the consideration
of the comments received on the first version.
With one exception, all documents issued contain at least
some information on measurement uncertainty. In nearly 70%
of the documents, considerable efforts were made to adapt
the concluded policy to the specified measurement procedure.
For the remaining 30%, different approaches were followed.
To some extent, the latter is surprising since the policy paper
gives specific advice on how to draft the respective clauses
and provides a valuable example of the structure and partly
even of the wording of which could be copied.
A major problem still appears to be the expression of a
functional relationship between the quantity to be measured
(measurand) and those input quantities which have to
be considered when determining the measurand and its
corresponding uncertainty. Such an equation is unavoidable
when the uncertainty is to be evaluated according to the rules
of the GUM. Sometimes, the reluctance of Working Groups to
formulate such a relationship is hardly understandable since all
the needed information is included in the document at least in
a verbal form. Moreover, at least one example how to present
the information in a mathematical form is available and was
offered to all Working Groups for consideration.
Another observation of a technical nature is that apparently
only little solid data on the expected contributions from the
various uncertainty sources mentioned in the documents
exist. To some extent, this is astonishing since by far the
largest percentage of the presently developed documents
constitute revisions of existing standards having been in
use for at least a couple of years. In cases where data are
included, they sometimes appear to be rather optimistic
estimates. Here is much room for further investigations. From
the view of the ongoing standardization process, however,
this lack of knowledge does not really hinder the progress
of developing documents since the adopted policy merely
requires specification of the concept of uncertainty evaluation