A proposed EU directive to set up
an efficiency rating system for gas fired boilers alarmed the industry,
which realised that different test centres could not test boilers
with the high level of reproducibility required by the rating system.
Testing centres and boiler manufacturers therefore launched a Measurements
And Testing (MAT) project to develop and improve testing methods so
that the maximum difference between laboratories measuring the same
boiler would be less than 2.5% (reproducibility). The rating system
of the directive is based on 3% (absolute) steps in efficiency.
The partners successfully harmonised procedures in different test
laboratories, allowing laboratories to streamline testing and cut
costs. Some of the procedures are also being incorporated in European
The partners have continued their collaboration to further improve
A sense of panic descended on producers
of gas and fuel oil fired boilers when the European Commission proposed
to introduce a rating system for such appliances based on measurements
of 3% steps in efficiency. The reason was that a 1991 comparison
of efficiency measurements by different European laboratories, on
the same boiler circulated among them, showed that measurements
could not be made with a reproducibility of better than 4.5%.
This held out the prospect of incorrect labelling of boilers assessed
under the rating system.
To avoid being unable to meet the requirements of the rating system,
gas suppliers, testing laboratories and boiler manufacturer companies
began a project with the support of the European Commission's Measurements
And Testing (MAT) programme to establish test methods that could
be used to assess boiler efficiencies with a reproducibility of
less than 2.5%.
The partners, coordinated by the Danish Gas Technology Centre,
investigated a number of aspects of boiler operation and testing.
These included the problem of determining when a boiler has reached
stable operation, the effects of ambient conditions, the way that
the waste heat from a boiler's water pump affects the efficiency
readings and the influence of burning gases with different calorific
values. For example, a laboratory can sometimes find it difficult
to determine whether a boiler has reached operating stability so
that measurements can be made. The MAT partners therefore developed
a way of establishing that stability had been reached by statistically
comparing efficiency measurements of a boiler made during four successive
Because boilers are tested under different ambient conditions, the
partners also developed correction formulae that take into account
the influence of ambient humidity, temperature and pressure. Traditionally,
only few laboratories can conduct tests at the standard specified
atmospheric conditions, although in some cases, such as when boilers
are operated at part load, the correction can be more than 1%. This
is a significant level in terms of overall test accuracy.
Adding up influences
Tests have traditionally not taken into account the heat from pumps
and other electrical components in boilers, with the result that
efficiency figures have been artificially high. In a typical 20
kW boiler, the pump might contribute about 100 W to the total heat
production, and the relative effect is greater when a boiler runs
at part-load. As a practical solution to the problem, the project
partners developed a universal model that gives reasonable results
based on 10 common pumps. A laboratory can therefore easily take
into account the effect of a pump without having to carry out expensive
testing of the pump. Theoretical and practical studies of the influence
of burning gases of different calorific values in test boilers showed
that the effects were too low to worry about.
The partners also realised that they used different formulae to
calculate heat output and flue gas losses of boilers under test.
Yet when they developed harmonised formulae the results were little
different from the results obtained using the laboratories' own
approaches. This showed that the different calculations were not
the reason for the poor reproducibility of test results.
Common test results
The project has led to the development of correction formulae that
can be conveniently applied by all laboratories to produce common
test results throughout Europe.
The partners accept that implementing all of their findings will
cost laboratories some money, typically to write correction algorithms.
However, they stress that overall costs will fall because thanks
to stability criteria, a boiler will now have to be tested only
once, rather than the several times previously needed to overcome
instability problems. Any investment is therefore likely to be recovered
in six to 12 months, say the partners. The new approach gives a
reproducibility of less than 2.5% at full boiler load. This is much
better than the 4.5% previously achieved by the laboratories.
The work has also identified new sources of systematic uncertainty,
including the drift found in instruments, particularly water and
gas flow meters. Although drift can be minimised by regular calibration,
laboratories tend to have different frequency of calibrating instruments.
The harmonisation of calibration methods across Europe is hampered
by the influence of national standards organisations, but the partners
believe that greater cooperation between them and the exchange of
information about instrumentation and calibration will lead to more
common approaches and therefore even more consistent test results.
For example, meters of a certain type might be calibrated once a
Good practice documents have already been produced, and the results
of the MAT project have been passed to the European standardisation
organisation, CEN. Some features, such as correction formulae for
condensing boilers, have already been adopted in standards. The
partners also consider that an important result of the MAT project
has been their decision to continue cooperating. In particular,
they intend to work together to solve the problems when boilers
are operating at part load. In this regard, the project has opened
up an interesting area for further study, through boosting the contact
between organisations that would not otherwise have worked together.