publishable result oriented report result oriented report.pdfatlete project - final publishable...
TRANSCRIPT
Intelligent Energy – Europe (IEE): Energy Efficiency (SAVE) -
September 2011
ATLETE Appliance Testing for Energy Label Evaluation
Publishable result-oriented
report
Author: Stefano Faberi (Editing)
In collaboration with Milena Presutto
Therese Kreitz
Atlete Project - Final Publishable Report
Acknowledgment
We would like to express our gratitude for the engagements and support which was offered by all
involved.
The project development and performance was only possible through the close cooperation of the
project partners and the help of the project officer.
Particular thanks go to the laboratories, the manufacturers, and the other stakeholders which,
through their professionalism and their proactive participation made the success of ATLETE
possible.
Atlete Project - Final Publishable Report | i
Summary
Objectives and Outcomes
The energy label tool has, since 1995, continued to provide easy-to-read information on
the characteristics of a household product placed on the European market.
Accuracy and confidence in how the European energy label is used when explaining the
characteristics of energy-using products is crucial. Information on the label is based on a
declaration made by the manufacturer about its product.
Consumers need to be able to purchase products, confident in the knowledge that the
information on the label is accurate. Meanwhile manufacturers need to be confident
that they are operating in a market where all competitors play by the rules on a level
playing field.
The ATLETE project aims to assist this process by highlighting how, through greater and
transparent market surveillance, European citizens can tap into the best European
technology when choosing labelled energy-using products.
In more detail the purpose of the ATLETE Project was to increase European-wide
implementation and control of energy labelling and eco-design implementing measures
for appliances. The developed methodology, once validated, will be applicable with very
minor adaptations for any Energy-using Products (EuP).
ATLETE is designed to demonstrate, and has actually demonstrated, that market
surveillance and testing can be done in a systematic, effective and cost-efficient way. It
helps transform the market and ensures the highest benefit for consumers,
manufacturers and the environment.
In this framework the major outcomes of the project have been:
• to set out and validate an overall procedure for the compliance monitoring of the
declared values for the EU policies;
• to carry out European-wide market surveillance on an EU policy measure by
testing 80 randomly selected refrigerating appliances;
• to review and compare the EU and international legislation related to energy
labelling of household appliances especially regarding the identification of the
effective enforcement of existing legislation through national market
surveillance. This last analysis has been furthermore compared with the ATLETE
methodology showing the aspects that must be reinforced to both improve the
commercial fairness and better support the EC sustainable objectives.
Verification procedure for the compliance monitoring of the
declared values
The methodology, tested and validated through the field work carried out by ATLETE on
refrigerating appliances, is applicable for any Energy Related Product as such or with
some minor adaptations. The overall verification assessment includes the steps shown in
the following figure:
ii | Atlete Project - Final Publishable Report
The geographical and technical scope concern the definition about the level of the
action (regional, national or EU-wide) and the types of the products to be monitored.
ATLETE has addressed its monitoring activity to the following four categories of
refrigerating appliances sold in the EU 27 Member States:
• bottom-mounted refrigerator-freezers (i.e. with freezer at the bottom) also
known as “combi” refrigerator-freezers;
• top-mounted refrigerator-freezers (i.e. with freezer on top of the appliance);
• freezers, upright and chest together;
• all the other refrigerating appliances (mainly simple refrigerator, but also side-by-
side refrigerator-freezers).
The models pertaining to these categories were selected through a “semi-random
selection” procedure, focused on best selleer models. This modality of selection was
actually considered to be more appropriate for this first pan-EU compliance verification
exercise. Best-seller products have the highest impact on the market (high sale volumes
and many variants) and usually are the products where commercial pressure is highest
(i.e. higher pressure towards inappropriate use of tolerances). This choice has led to a
division of the market where the models to be tested fell under two categories: EU top-
sellers, i.e. the best sold models of the most relevant producers at EU27 level and the
rest of the market. This ensured that other manufacturers with a market share lower
than 0,5% or operating only nationally/regionally are targeted. By applying this criteria
and basing on a reliable database of the best sold models for each manufacturer
provided by the market research firm GfK, 80 Models produced by 40 manufacturers
were finally selected1.
The laboratories for testing identified models were selected by using a mix of knock-out
criteria and a score system. Knock-out criteria eliminate immediately least qualified
laboratories without any further investigation, while the score system allows ranking the
remaining laboratories according to the expected testing capability. The scoring system
was created by giving a “weight” (score) to the answers received in the questionnaire.
From an initial list of 23 laboratories that were contacted, 15 replied to the ad-hoc
1 In reality 82 models were tested at the end of the project, because for two appliances the three
additional units purchased for developing Step 2 belonged to a different model. The list of the tested
models is given in the Atlete web site at:
http://www.atlete.eu/index.php?option=com_content&view=article&id=121&Itemid=117
Compliance assessment methodology
Geographical and technical scope
Sampling criteria
Verification procedure
Actions after non-compliance
Reporting and dissemination
Laboratory selection criteria
Follow-up actions
Harmonised Standards
Compliance assessment methodology
Geographical and technical scope
Sampling criteria
Verification procedure
Actions after non-compliance
Reporting and dissemination
Laboratory selection criteria
Follow-up actions
Harmonised Standards
Atlete Project - Final Publishable Report | iii
Questionnaire, 10 were selected having passed the knock-out criteria and achieved a
defined score (25% highest ranking). Finally a call for tender was launched to these 10
laboratories of which four where finally selected.
The verification procedure, according to the essential requirements described in the EU
legislation on labelling/ecodesign was carried out in two testing steps plus an
intermediate phase of voluntary corrective action (this last voluntary corrective action
has been introduced in this project). The figure below outlines the followed procedure:
The novelty introduced by this project has been the "Protocol on the manufacturer’s
pro-active participation in the ATLETE Project". All manufacturers of the selected models
were invited to sign such a protocol whereby they accepted to take, proactively, remedy
actions should tested products they placed on the market fail to show compliance. This
protocol was signed by 27 out of 40 manufacturers and 18 remedy actions were carried
out after the results of the first step of tests.
All the results of the laboratories tests, both complaint or not complaint ones have been
eventually disclosed to the EU national Market Surveillance Authorities of the countries
where each model was reported to be sold as well as to the stakeholder experts via the
project public workshops and the general public: through the project website, media:
interviews, press releases, articles, etc.
It is worth noting that different opinions exist as to whether making the results of a
compliance verification action publicly available, disclosing also the model and the
supplier name. At EU level all Member States that have developed market surveillance
have not disclosed the model and the supplier name while the Australian authorities
fully disclose the results of the market surveillance activity once the procedure has been
completed and the final judgement on the tested models is reached. On the other side
in the USA it has been argued that an unfair market advantage can be given to products
and suppliers meeting the requirements over the competitors that have not been
selected for the verification action. It is thus recommended that some guidance on this
matter might be provided by the European Commission or the ADCO Group on
ecodesign and labelling.
1 unit tested(Step1)
Random Products selection
Pass?
Market SurveillanceAuthorities
in the countries where the product
was availableCorrective actions?
3 additional units tested (Step 2)
YES
YES
NO
NO
Notification of compliance
YES
Pass?
NO
Notification of non-compliance & remedy action
Notification of compliance
Notification of non-compliance
iv | Atlete Project - Final Publishable Report
As outlined before, the modularity of the proposed methodological approach makes
adaptation to other appliances and products feasible with (minor) adaptations. It is
obvious that each step of the procedure has to be adapted to the type of appliance or
Energy Using Product to be tested, but the most critical ones are probably the sampling
criteria (the market structure has to be carefully analysed) and the laboratories
selection.
Final results of the field work
As already outlined, 80 models of 40 manufacturers, available within the Single
European Market, have been tested to verify their compliance with the EU energy label
in four European laboratories. For each model the tests were conducted on the
following five technical parameters2:
Parameter Description
1. Energy consumption energy consumed by a refrigerating appliance over the
period of 24 hours)
2. Storage temperature (&
climate class)
the ability of an appliance to maintain simultaneously the
required storage temperatures in the different
compartments at a certain ambient temperature
3. Storage volume the space to store food in the appliance
4. Freezing capacity the amount of food, expressed in kilograms, that can be
frozen to a core temperature of -18 °C in 24 h
5. Temperature rise
time
the time period needed to raise the temperature of food in
the frozen food compartment from -18 °C to - 9 °C after the
operation of the refrigerated system has been interrupted
The final test results show that 80% of appliances subjected to testing and for which
testing has been concluded complied with the energy efficiency class declaration and
the two related key parameters: energy consumption and storage volume. But when all
five parameters are taken into consideration 57% of them do not comply with at least
one of the tested parameters.
The analysis of the tests results are summarized in the figure below, in which the overall
picture of the pan-EU compliance of refrigerators and freezers are presented. Although
the compliance rate for the single parameters is in the range 70-90%, the overall
compliance rate is unacceptably low at 43%, due to the combined effect of the single
non-compliances.
2 To this end the harmonized standards: EN 153:2006 & EN ISO 15502:2005, concerning the testing
normative for refrigerators and freezers ruled by the 2003/66/EC Energy Labelling Directive and that Eco-
labelling Regulation, have been applied by the selected laboratories
Atlete Project - Final Publishable Report | v
More in detail, the figure below shows the results on the critical parameter on the
compliance with the declared energy class. The figure below shows that with respect to
70 models for which the steps 1 and 2 have been carried out, 55 (that is the 79%), have a
correct energy efficiency class declaration (the green labels in the figure) while 10 tested
models (14%) were overrated by 1 class and 4 tested models (6%) were overrated of 2 or
more classes. thirteen models were unfortunately classified as undecidable (the yellow
labels) as no 3 equivalent models were found on the market at the time to carry out the
second testing step.
Energy consumption is another critical parameter. The results have actually shown that,
apart from the models largely exceeding the permitted tolerances, for some models the
difference between the rated energy consumption and the average measured energy
consumption of Step 2 is much smaller that the difference found in Step 1, but also the
contrary happens. This variation on this critical parameter, measured on same models,
justifies the need to allow a tolerance range for the energy efficiency compliance.
Non compliance analysis
The wealth of data provided by the test results of the ATLETE project allow, for the first
time, to run a semi-quantitative analysis of the non-compliance causes in order to
provide some guidance to better direct the selection of potentially non-compliant
79%
77%
90%
73%
84%
70%
43%
21%
23%
10%
27%
16%
30%
57%
0% 20% 40% 60% 80% 100%
Energy class
Energy consumption
Storage temperature
Storage volume
Temperature rise time
Freezing capacity
TOTAL
PASS
FAIL
vi | Atlete Project - Final Publishable Report
models. It is nonetheless worth noting that the conclusions of this analysis should be
considered only as rough guidance, given the small size of the considered sample. Four
important market and energy categories of the models have thus been analysed (i.e.
product category, price, manufacturing origin and energy efficiency class) and the table
below shows, per each of these issues, which are the possible clues non compliance
risks.
Models categories Possible indicator of non-compliance risk
The product category The higher non-compliance risks are
concentrated to low temperature
compartment refrigerators and to freezers
The purchasing price The results have shown that the purchasing
price is a good indicator of possible non-
compliance: more expensive models have
generally a higher probability to be compliant,
while cheaper models are often more non-
compliant.
The manufacturing place Although, again, the dimension of the sample
in not sufficient to draw statistically sound
conclusions, the trend is very clear: when the
country of manufacturing is not indicated or
indicated in a very generic way such as “EU”
the probability of non-compliance is higher
than when the country is clearly stated. The
same occurs when the products manufactured
in a specified EU Member State are compared
to the products manufactured in non-EU
countries.
The expected energy efficiency class Also here a further correlation appears to exist
between the declared energy efficiency class
and the compliance rate: the higher the
energy efficiency class the higher the
probability to find a compliant model.
All in all, since a correlation appears to exist between the purchasing price and the
energy efficiency class, in the end, the highest the purchasing price and energy efficiency
class, the higher the probability of compliance.
It is important to add at this point that the verification procedure of the tested models
did not always take place without problems. During the testing campaign several
problems arose of different nature, some attributable to a lack of respecting market
rules or to how to interprete Directives .
As far as the EU Labelling Directive interpretation is concerned, we have argued if the
model purchasing via on-line shops can be accepted as a true random selection on the
market. Actually it has been witnessed that when purchasing via the internet three
Atlete Project - Final Publishable Report | vii
additional units of a model for the step 2 verification, they all belonged to the same
batch and were produced very closely in time.
In another case we have experienced the importance of a clear description for the
rounding up of measured values for the compliance verification. This is of great
importance for the models that are “border line” with the energy efficiency class
thresholds or accepted tolerance or specific minimum conditions requested by the
standard. Actually it may come about that the compliance or non-compliance status
depends to the number of decimal digits with which the annual energy consumption has
been calculated. The following example, referring to an appliance with a declared annual
energy consumption of 124 kWh/year and in which the first figure refer to the hourly
energy consumption measured in the laboratory, clarifies this issue:
(a) 0,3904658 kWh/24h×365=142,52 kWh/year, rounded to the nearest integer is 143 kWh/year or
+15,3% higher than the rated value, the model is non-compliant for the energy consumption
(b) 0,39 kWh/24h×365=142,35 kWh/year, rounded to the nearest integer is 142 kWh/year or +14,8%
higher than the rated value, the model is compliant for the energy consumption.
This situation can happen whatever is the allowed tolerance and for any parameter to
be verified if the performance of the tested model is very close to the minimum
requirement or the threshold value.
In others cases, mostly depending to the inaccuracy of the manufacturers or "strange"
declarations of the importers, we have come across label layout belonging to the same
models but with different declared values for the energy consumption and/or the
storage volume, or same commercial code numbers for technically different appliance
models or even different suppliers (importers) for the same appliance model. In this last
situation, which of the suppliers will be – from the legal point of view – responsible for
the non-compliance or for the eventual correcting action or a possible sanction? These
problems were all solved (apart that concerning the different importers for the same
models) through interactions with the manufacturers, but at the expense of waste of
resources and time.
Main Critical Elements of the Proposed Procedure
After the completion of the field work and in the light of the achieved results, a critical
analysis of the applied evaluation procedure was developed with the aim of correcting
the aspects that have presented uncertainty or have been shown to be excessively
cumbersome. Some of the critical aspects have been already outlined in this brief
outline on the project outcomes, others were highlighted by the testing laboratories,
that were asked about their views about the verification procedure developed within
the ATLETE project and how it could be improved at the end of the field work.
The analysis has been focused on the models selection, purchasing and identification
and on the procedural and technical indication concerning the testing activity.
The problems concerning the way the models should be selected, purchased and
identified stem all from the difficulties met and the experience acquired during the
project. The challenges identified deal with the possibility to carry out all the evaluation
steps without running the risk of not finding three additional models on the market, or
viii | Atlete Project - Final Publishable Report
the necessity to clarify if, or to what extent, on-line purchasing can be considered valid
for the purchasing of the selected models or, even, to ensure that units apparently
belonging to the same model are actually the same product, in the case that they are
not univocally identified by their commercial code number.
With regard to the testing activities it has been highlighted that it is necessary to
adequately size the number of the selected laboratories in order to avoid saturating
them. Paying adequate remuneration taking into account possible actions beyond the
mere model testing and results reporting needs also to be considered. It would attract
top quality laboratories to participate in selection bids. Another requirement concerns
the necessity to provide a mandatory common test format for the reports the
laboratories have to provide (according to the applied standard(s) or legislation) in order
to better allow that results comparison among the laboratories themselves. Finally, on
the basis of the tests carried out during the ATLETE project, the laboratories, in
agreement with the project team, have suggested a list of improvements for the
refrigerating appliances standards.
Market surveillance procedures at EU and international level
In the project, a detailed analysis of the international (Australia and USA) and European
legislation and practice on the verification procedures has been performed on both the
basis of the technical literature and a survey carried out with the aim of assessing
surveillance practices in the EU Member States and in 4 other non-EU countries:
Australia, Japan, Turkey and the United States. The results of this analysis are that
Australia (with New Zealand) and the USA take into due consideration the surveillance
practices and, to varying extents, have not only developed good verification protocols
(actually not very different from those prescribed in the EU legislation), but are able to
seriously put them into practice. The figure on the following page shows for example the
two-stage verification testing process used by the US/DOE to carry out a pilot action on
261 models of different types of appliances.
In the EU, in accordance with the survey results, more than half Members States (i.e. 14)
declare performing tests in practice but only 5 countries declare testing appliances each
year (Denmark, Hungary, Greece, the UK and The Netherlands) and Sweden has been
running almost regularly appliance testing over the past years. All the other have
performed these tests just once in the last 10 years or very rarely.
Atlete Project - Final Publishable Report | ix
The two-stage verification testing process for the US/DOE pilot action
Nonetheless 12 Member States do not carry out at all appliance tests of energy labels’
accuracy. The main reason for not conducting the tests is that these procedures are
considered too expensive but also the lack of adequate laboratories and the fact that
the surveillance authorities are often undersized (being their main duty the safety
controls) are other relevant barriers.
The ATLETE project, without disregarding the problems claimed by the surveyed
countries, has in turn demonstrated that, using a formal and well tuned methodology,
surveillance activities for household appliances are technically possible and cost
effective.
Final conclusions
1. The ATLETE Project has brought specific positive examples of product test results
and has demonstrated that market surveillance activities for household appliance
energy labelling are essential, technically possible and cost effective. Lack of
market surveillance in this area leads to unfair competition and misleading of
consumers in their search for highly energy efficient products. It enables “free-
riders” to gain considerable market share which is undermining the market and
does not allow to achieve demanding energy efficiency targets.
2. The ATLETE Project has also proved that two steps approach is crucial for
completing proper verification procedure. Almost 25% of models tested in Step 2
3
3
6
The controlling institution has too many subjects to deal with
Finding an appropriate laboratory or independent company to proceed to the tests is difficult
Tests would be too expensive
number of countries
Why energy classification is not verified?
x | Atlete Project - Final Publishable Report
(that is suspected non-compliance in step 1) occurred to be fully compliant after
Step 2.
3. Cooperation of testing laboratories within the testing procedure is very helpful in
achieving fully comparable test results and exchanging the experience gained
during the testing procedure. Four meetings between the testing laboratories,
project partners and representatives of IAC committee enable also to provide
added value to the evaluation and recommendation phase in WP6
Atlete Project - Final Publishable Report | xi
Table of contents
1 THE PROJECT IN BRIEF: OBJECTIVES, ACHIEVEMENTS, OUTCOMES
AND MAIN FACTS ............................................................................................................................. 1
2 VERIFICATION PROCEDURE FOR THE COMPLIANCE MONITORING OF
THE DECLARED VALUES .................................................................................................................... 4
2.1 INTRODUCTION ...................................................................................................................................... 4 2.2 GEOGRAPHICAL AND TECHNICAL SCOPE OF THE TARGET PRODUCTS ................................................................... 5 2.3 SAMPLING CRITERIA................................................................................................................................ 6 2.4 LABORATORIES SELECTION CRITERIA ........................................................................................................... 7 2.5 VERIFICATION PROCEDURE ....................................................................................................................... 8
2.5.1 The two step compliance verification procedure ...................................................................... 8 2.5.2 Protocol on the manufacturer’s pro-active participation to the ATLETE
Project ..................................................................................................................................... 10 2.6 TESTED PARAMETERS AND HARMONIZED STANDARDS .................................................................................. 11 2.7 VERIFICATION RESULTS REPORTING AND DISSEMINATION .............................................................................. 11 2.8 ADAPTATION OF THE METHODOLOGY TO OTHER ENERGY RELATED PRODUCTS .................................................. 12
3 FINAL RESULTS OF THE FIELD WORK ............................................................................................... 15
4 NON COMPLIANCE ANALYSIS .......................................................................................................... 22
4.1 IDENTIFICATION OF THE NON-COMPLIANCE CASES ....................................................................................... 22 4.1.1 Models formally non-compliant before the laboratory tests .................................................. 22 4.1.2 Models that underwent a “manufacturers’ voluntary remedy action .................................... 23 4.1.3 Models (suspected to be) damaged or defective .................................................................... 23
4.2 ANALYSIS OF THE NON-COMPLIANCE MACRO-SOURCES ................................................................................ 24 4.2.1 Product category ..................................................................................................................... 25 4.2.2 Purchasing price ...................................................................................................................... 25 4.2.3 Place of manufacturing ........................................................................................................... 27 4.2.4 Declared Energy Efficiency class ............................................................................................. 28
4.3 FURTHER SOURCES OF UNCERTAINTY IN THE COMPLIANCE VERIFICATION .......................................................... 29 4.3.1 The EU labelling Directive interpretation (1), concerning the models selection
in the market ........................................................................................................................... 29 4.3.2 The EU labelling Directive interpretation (2): rounding of measured values .......................... 30 4.3.3 Market problem (1): different labels and declarations for the same appliance
model ...................................................................................................................................... 31 4.3.4 Market problem (2): same commercial code number for technically different
appliance models .................................................................................................................... 32 4.3.5 Market problem (3): different suppliers for the same appliance model ................................. 33 4.3.6 Improvement of the test method ............................................................................................ 34
5 MAIN CRITICAL ELEMENTS OF THE PROPOSED PROCEDURE ........................................................... 35
5.1 MODELS SAMPLING, PURCHASE AND IDENTIFICATION .................................................................................. 35 5.1.1 Models sampling ..................................................................................................................... 36 5.1.2 Models purchasing .................................................................................................................. 37 5.1.3 Models univocal identification ................................................................................................ 37 5.1.4 Laboratories selection and remuneration ............................................................................... 37 5.1.5 Final test report ....................................................................................................................... 37 5.1.6 Measured parameters in Step 1 and Step 2 of the compliance verification ........................... 38 5.1.7 Improvement of the test method ............................................................................................ 38
6 MARKET SURVEILLANCE PROCEDURES AT EU AND INTERNATIONAL
LEVEL AND COMPARISON WITH THE ATLETE METHODOLOGY. ....................................................... 39
6.1 MARKET SURVEILLANCE PROCEDURES FOLLOWED AT INTERNATIONAL LEVEL ..................................................... 40 6.1.1 Australia and New Zealand ..................................................................................................... 40 6.1.2 USA .......................................................................................................................................... 43
xii | Atlete Project - Final Publishable Report
6.2 BRIEF OVERVIEW ON THE FORMAL PROCEDURE FOLLOWED IN EU MEMBER STATES ........................................... 47 6.3 RESULTS OF TESTING ON REFRIGERATING APPLIANCES DEVELOPED IN RECENT YEARS AT MEMBER STATES LEVEL ........ 49
6.3.1 Sweden .................................................................................................................................... 49 6.3.2 Denmark .................................................................................................................................. 49 6.3.3 UK ............................................................................................................................................ 50
6.4 COMPARISON WITH THE EU AND THE ATLETE APPROACH ........................................................................... 53 6.4.1 Differences and common aspect in the applied verification procedures ................................. 53 6.4.2 Comparison of the pass/fail rates for refrigerating appliances .............................................. 54
6.5 THE MANUFACTURERS’ CONFORMITY ASSESSMENT IN THE EU MEMBER STATES ............................................... 55 6.5.1 Countries performing or not performing the tests in practice ................................................ 55 6.5.2 The laboratories issue ............................................................................................................. 57 6.5.3 The verification Tests .............................................................................................................. 57 6.5.4 Non-compliance and sanctions ............................................................................................... 58 6.5.5 Sharing the results and information ........................................................................................ 58 6.5.6 Perspectives on Ecodesign ....................................................................................................... 59
7 FINAL POLICY CONCLUSIONS AND RECOMMENDATIONS ................................................................ 60
Figures
FIGURE 1: METHODOLOGICAL STEPS FOR COMPLIANCE ASSESSMENT ............................................................................ 4
FIGURE 2: ENERGY LABEL FOR REFRIGERATORS AND FREEZERS ACCORDING TO THE DIRECTIVES
94/2/EC AND 2003/66/EC ................................................................................................................ 5
FIGURE 3: EU LABORATORIES CONTACTED FOR THE ATLETE PROJECT .......................................................................... 8
FIGURE 4: VERIFICATION PROCEDURE SCHEME AS DEVELOPED WITHIN ATLETE .............................................................. 9
FIGURE 5: OVERALL COMPLIANCE RESULTS OF THE ATLETE PROJECT ......................................................................... 19
FIGURE 6 EFFICIENCY CLASSES DECLARED / VERIFIED ............................................................................................... 20
FIGURE 7 DIFFERENCE (%) BETWEEN THE DECLARED AND THE MEASURED VALUE OF THE ENERGY
CONSUMPTION IN STEP 1 AND STEP 2 FOR THE 82 MODELS OF REFRIGERATING
APPLIANCES ...................................................................................................................................... 21
FIGURE 8 DIFFERENCE (%) BETWEEN THE DECLARED AND THE MEASURED VALUE OF THE ENERGY
CONSUMPTION IN STEP 1 AND STEP 2 (SINGLE UNITS AND AVERAGE VALUES) ................................................ 21
FIGURE 9 DIFFERENCE (%) BETWEEN THE DECLARED AND THE MEASURED VALUE OF THE STORAGE
VOLUME FOR THE REFRIGERATOR AND FREEZER COMPARTMENTS FOR THE 82 MODELS OF
REFRIGERATING APPLIANCES ................................................................................................................. 21
FIGURE 10: DEFECTIVE GASKET OF THE FREEZER COMPARTMENT DOOR ......................................................................... 24
FIGURE 11: PURCHASING PRICE FOR THE TESTED MODELS OF REFRIGERATORS AND FREEZERS ............................................. 26
FIGURE 12: PURCHASING PRICE FOR THE TESTED MODELS OF REFRIGERATOR-FREEZERS AND
REFRIGERATORS WITH A LOW TEMPERATURE COMPARTMENT ..................................................................... 27
FIGURE 13: PURCHASING PRICE FOR THE TESTED MODELS OF UPRIGHT AND CHEST FREEZERS ............................................. 27
FIGURE 14: PLACE OF MANUFACTURING AND COMPLIANCE WITH THE EU ENERGY LABELLING ............................................ 28
FIGURE 15: FIRST UNIT OF A MODEL PURCHASED FOR STEP 1 OF THE VERIFICATION ......................................................... 29
FIGURE 16 : ADDITIONAL THREE UNITS OF THE MODEL PURCHASED FOR STEP 2 OF THE VERIFICATION ................................. 29
FIGURE 17: LABEL STRIPS FOR THE 4 UNITS OF THE SAME MODELS USED IN STEP1 AND STEP2 OF THE
VERIFICATION PROCEDURE ................................................................................................................... 31
FIGURE 18: LABEL STRIPS FOR THE 4 UNITS OF THE SAME MODELS USED IN STEP1 AND STEP2 OF THE
VERIFICATION PROCEDURE ................................................................................................................... 32
FIGURE 19: RATING PLATE AND LABEL STRIP FOR THE FIRST UNIT OF THE REFRIGERATOR-FREEZER ....................................... 33
FIGURE 20: RATING PLATE AND LABEL STRIP FOR THE SECOND UNIT OF THE REFRIGERATOR-FREEZER ................................... 33
FIGURE 21: DIFFERENT SUPPLIERS FOR DIFFERENT UNITS OF THE SAME MODEL PURCHASED IN THE
SAME COUNTRY ................................................................................................................................. 34
FIGURE 22: TOTAL NUMBER OF CHECK TESTS CONDUCTED IN AUSTRALIA BY APPLIANCE CATEGORY,
1991-2010 ..................................................................................................................................... 42
FIGURE 23: THE TWO-STAGE VERIFICATION TESTING PROCESS FOR THE DOE PILOT ACTION ............................................... 46
Atlete Project - Final Publishable Report | xiii
Tables
TABLE 1: ELEMENTS TO BE MODIFIED TO TRANSFER THE METHODOLOGY FROM REFRIGERATING
APPLIANCES TO THE OTHER APPLIANCES .................................................................................................. 13 TABLE 2: OVERALL RESULTS OF THE LABORATORY TESTS ......................................................................................... 18 TABLE 3: SUMMARY OF THE RESULTS OF THE COMPLIANCE VERIFICATION OF THE DIFFERENT TYPES
OF REFRIGERATING APPLIANCES ............................................................................................................ 25 TABLE 4: AVERAGE, MIN AND MAX PURCHASING PRICE (€) FOR COMPLIANT AND NON-COMPLIANT
FOR THE APPLIANCE CATEGORIES DEFINED IN THE LABELLING SCHEME .......................................................... 26 TABLE 5: PLACE OF MANUFACTURING AS AND RATE OF COMPLIANCE WITH THE EU ENERGY
LABELLING ........................................................................................................................................ 27 TABLE 6: RELATION BETWEEN THE DECLARED ENERGY EFFICIENCY CLASS AND THE COMPLIANCE
RATE FOR REFRIGERATING APPLIANCES WITHIN THE ATLETE PROJECT .......................................................... 28 TABLE 7: FINAL NUMBER OF SELECTED MODELS FOR THE DOE VERIFICATION PILOT PROGRAMME .................................. 45 TABLE 8: PRELIMINARY RESULTS FOR DOE THE VERIFICATION ACTION ....................................................................... 47 TABLE 9: SUMMARY OF THE EU VERIFICATION SYSTEM AND TOLERANCES FOR ENERGY
CONSUMPTION IN THE PREVIOUS LABELLING AND MINIMUM REQUIREMENTS SCHEMES .................................... 48 TABLE 10: SUMMARY OF THE EU VERIFICATION SYSTEM AND TOLERANCES IN THE NEW LABELLING
AND ECODESIGN SCHEMES ................................................................................................................... 48 TABLE 11: RESULTS OF THE 2007 TESTS ON HOUSEHOLD APPLIANCES IN DENMARK ...................................................... 50 TABLE 12: ANALYSIS OF THE RESULTS OF THE 2005 ENERGY LABEL COMPLIANCE FOR
REFRIGERATING APPLIANCES IN UK ........................................................................................................ 52
xiv | Atlete Project - Final Publishable Report
Atlete Project - Final Publishable Report | 1
1 THE PROJECT IN BRIEF: OBJECTIVES, ACHIEVEMENTS, OUTCOMES AND
MAIN FACTS
Objectives
The purpose of the ATLETE Project was to increase European-wide implementation and
control of energy labelling and eco-design implementing measures for appliances. The
developed methodology, once validated, will be applicable with very minor adaptations
for any Energy-using Products (EuP).
Energy labels are a crucial driver for market transformation, orienting consumers'
choices towards more energy efficient appliances and thus realising the potential of
available technologies. Unfortunately, not all Member States (MS) apply a responsible
policy for controlling the correct labelling implementation. This needs to change.
ATLETE is designed to demonstrate, and has actually demonstrated, that market
surveillance and testing can be done in a systematic, effective and cost-efficient way,
thus helping to transform the market to ensure the highest benefit for consumers,
manufacturers and the environment.
Strong points of the ATLETE Project:
• addresses the issue of compliance testing relating to Energy labelling and Eco-
design requirements;
• ensures that the national authorities are informed of cases of non-compliance;
• identifies effective enforcement of existing legislation through national market
surveillance;
• provides the first pan-European testing results on a large number of refrigeration
appliances;
• gives concrete guidance to EU and National Authorities for effective labelling and
future eco-design requirements implementation;
• sets a shared procedure for the verification of the manufacturers' labelling/eco-
design declarations including a methodology for laboratories accreditation and
models selection;
• raises awareness among National Authorities of the effectiveness of the energy
labelling on national energy efficiency;
• engages key stakeholders that have an interest in the issue of compliance with
energy label requirements: the EU institutions, Government organisations,
manufacturers, retailers, associations, consumer groups, NGOs, the media and
general public.
2 | Atlete Project - Final Publishable Report
Achievements and Outcomes
ATLETE has largely achieved these objectives providing the following three important
results:
• setting out and validation of an overall procedure for the compliance monitoring
of the declared values for the EU policies;
• carrying out a European-wide market surveillance on a EU policy measure by
testing 80 randomly selected refrigerators;
• reviewing and comparing the EU and international legislation related to energy
labelling of household appliances especially for what concerns the identification
of the effective enforcement of existing legislation through national market
surveillance. This last analysis has been furthermore compared with the ATLETE
methodology showing the aspects that must be reinforced to both improve the
commercial fairness and better support the EC sustainable objectives.
The first of these results was made possible by the performance of the laboratory
testing campaign carried out by the project. In fact, the experience gained from this field
phase has allowed the project team to verify and amend the theoretical verification
procedure developed during the first phase of the project and to deliver the first issue of
the "Guidelines for Verification of EU Energy Related Products"3
In turn, the field work, beyond to provide meaningful hints for the laying out of the
Guidelines, has provided a twofold outcome:
• on one side it has demonstrated that that market surveillance is technically and
economically feasible, provided that adequate (but still at an acceptable level)
human and financial resources are made available,
• on the other side it allowed to evaluate the overall compliance of the selected
refrigerating appliances with the EU labelling scheme shedding light on a reality
until now little known by revealing a high non-compliance percentage with the
energy labelling Directive requirements (especially for what concerns the formal
requirements).
Dissemination activities
To ensure transparency, the final results of the project are made publically available via
the project website www.atlete.eu. Information made available includes the models
tested, the laboratories involved, final overall results and results by model and
parameter.
Moreover abundant dissemination material has been produced and distributed, several
seminars have been directly organized and in many other the ATLETE project has been
presented:
o Leaflets: more than 5,000 per country)
o Press Releases: two press releases published in all languages)
o Seminars: 19 events in total, 11 at international level
3 See The Guidelines for EU verification of Erp at
"http://www.atlete.eu/index.php?option=com_docman&Itemid=111"
Atlete Project - Final Publishable Report | 3
It is finally worth noting that about 93 articles have been monitored in total, including
some high level national and European daily or electronic press/media. All these articles
were published after the ATLETE EU final conference (12 April 2011 Brussels) and the
national events held in the Czech Republic, France and Italy.
During the period April – September 2011, that starting from the ATLETE final
conference and in parallel with the issuing of the above mentioned articles, the website
was visited by over 8700 visitors, which is a significant success, since the website was
not intended to the final consumers / general public, but to authorities, stakeholders,
experts and journalists.
Main Facts
• 5 Partners:
� ISIS (Institute of Studies for the Integration of Systems), Project
Coordinator
� CECED (European Committee of Domestic Equipment
Manufacturers), Responsible for the Laboratories Selection and
management of the Appliances Testing
� ENEA (Italian National Agency for new Technology, Energy and the
Environment), Responsible for the overall project methodology
and final results evaluation
� SEVEn (Energy Efficiency Center), Responsible for the
dissemination activities
� ADEME (French Environment and Energy Management Agency),
Responsible for the initial Background Analysis
• Target Groups
o Institutions, & Government organisations;
o Manufacturers & Retailers;
o Associations, Consumer Groups and NGOs;
o The media and general public
• Key Actors
o National Energy Agencies;
o Manufactures & Importers
• Duration 27 months4 : June 2009 – July 2011
• Budget:
o Total Budget: 1.012.057€
o Total EC financial contribution: 759.043€
4 The initial duration of the project was of 24 months.
4 | Atlete Project - Final Publishable Report
2 VERIFICATION PROCEDURE FOR THE COMPLIANCE MONITORING OF THE
DECLARED VALUES
As outlined in the previous chapter, one of the main result of this project was the setting
of an overall procedure for the compliance monitoring of the declared values for the EU
policies and the delivering of a comprehensive "Guidelines for EU Verification of Energy
Related Products" having the objective to provide a supporting tool valid at EU and
Member States level for the Authorities dealing with compliance and verification issues,
helping to optimize the available resources and minimize the inconvenience for
Authorities and industry while enforcing legislation concerning Energy Labelling and Eco-
design requirements
The developed approach, tested and validated through the field work carried out by
ATLETE on refrigerating appliances, is applicable for any Energy Related Products as such
or with some minor adaptations. This chapter summarises the main, relevant, steps of
the verification procedure applied in this project and finally shows which are the points
that have to be modified to transfer the methodology from refrigerating appliances to
the other appliances. To analyse more in depth this procedure please see the quoted
Guidelines that provides a detailed information on all steps that should be followed and
the documents to be prepared to carry out an accurate and effective compliance
verification.
2.1 Introduction
The overall verification assessment includes the steps shown in Figure 1. These steps are
described in details in the following paragraphs. Due to the time-frame of the
presentation of the project proposal and its approval by the IEE Programme, the
reference for the compliance verification was the energy label (Figure 2) and labelling
scheme established by Directives 94/2/EC and 2033/66/EC.
Figure 1: Methodological steps for compliance assessment
Compliance assessment methodology
Geographical and technical scope
Sampling criteria
Verification procedure
Actions after non-compliance
Reporting and dissemination
Laboratory selection criteria
Follow-up actions
Harmonised Standards
Compliance assessment methodology
Geographical and technical scope
Sampling criteria
Verification procedure
Actions after non-compliance
Reporting and dissemination
Laboratory selection criteria
Follow-up actions
Harmonised Standards
Atlete Project - Final Publishable Report | 5
Figure 2: Energy label for refrigerators and freezers according to
the directives 94/2/EC and 2003/66/EC
2.2 Geographical and technical scope of the target products
A compliance verification action can be run at local, national, EU or even international
level and can involve a single specific product or different products for which one or
more (common) aspects need to be verified. A decision about the level of the action and
the number of involved products is the pre-requisite for starting any actual procedure.
The geographical and technical scope of the ATLETE project was:
• Geographical scope:
− EU27 Member States as far as possible (GfK provided data for EU 24). If the
external market statistical sources have a more limited scope, it should be
evaluated that this limitation does not exclude specific markets
− National relevant supplier in the main manufacturing EU countries: BE, DE, DK,
ES, FR, IT, NL, PL, UK.
− Other suppliers randomly selected.
• Technical scope: refrigerating appliances (refrigerators and
freezers), divided into the following four categories5:
1. Bottom-mounted refrigerator-freezers (i.e. with freezer at the
bottom) also known as “combi” refrigerator-freezers
5 It is worth noting that the above four categories are different from the categories defined in the
previous energy labelling/minimum requirement directives and from the 10 categories defined in the
previous and new energy labelling scheme and Ecodesign Regulation, but they respond to the market-
related categories for which information are available from the major market research firm operating in
the EU.
6 | Atlete Project - Final Publishable Report
2. Top-mounted refrigerator-freezers (i.e. with freezer on top of the
appliance);
3. Freezers, upright and chest together;
4. all the other refrigerating appliances (mainly simple
refrigerator, but also side-by-side refrigerator-freezers).
2.3 Sampling Criteria
A “semi-random selection” procedure, focused on the best sold models was considered
the most appropriate for the EU.
In fact, although national Market Surveillance Authorities have usually rather limited
resources and would prefer to maximize the effectiveness of their compliance actions by
using a “maximum failure” selection procedure, the aim of the ATLETE project was to
show how much the energy labelling scheme is trustful and to give a full picture of EU
market. A high rate of non-compliance following a “maximum failure selection” would
not have given a realistic indication of market compliance.
Another alternative for models selection could have been to go for the models with a
high energy efficiency (A++ or even better), but these models are usually under the spot
light also from competitors and although having a high visibility represent limited sale
volumes on the EU market. Again the semi-random selection was felt more appropriate
for the first pan-EU compliance verification exercise. Best-seller products have the
highest impact on the market (high sale volumes and many variants) and usually are the
products where commercial pressure is higher (i.e. higher pressure towards
inappropriate use of the tolerance).
In practice, to guarantee that the tested products cover all manufacturers and brands
operating within the Community market of refrigerating appliances, half of the models
were selected among the “EU top-sellers” according to the market share of the relevant
manufacturers/importers, i.e. the best sold models of the most relevant producers at
EU27 level; in this way bestseller models of all major brands in Europe were targeted
and tested. The second half of the models was selected randomly within the remaining
producers active on the EU27 market, thus ensuring that other manufacturers with a
market share lower than 0,5% or operating only nationally/regionally are targeted. The
selection was based on the Market Share of each supplier (including all the owned
brands) at European level according to the data provided by the market research firm
for January 2010:
Atlete Project - Final Publishable Report | 7
• if market share (MS) is MS >10% or 5%<MS<10% or 1%<MS<5%, 15 top best
seller models in each of the 4 product categories are considered
• if market share is 0,5%<MS<1% and 0,1%<MS<0,5% and for major national
manufacturers, 10 top best seller models are considered without product
category distinction.
Out of the list of the models identified for each market share range, a random selection
was performed, to choose the 80 models to be tested.
The updated information about the manufacturers’ market share at EU and at national
level and a reliable database of the best sold models for each manufacturer was
purchased - after a public Call for Tender open to all market data providers in Europe -
from the well-known market research firm GfK.
In reality 82 models were tested at the end of the project, because for two appliances
the three additional units purchased for developing Step 2 belonged to a different
model6.
2.4 Laboratories selection criteria
Reliable and transparent selection criteria were used, including a mix of knock-out
criteria and a score system. Knock-out criteria eliminate immediately least qualified
laboratories without any further investigation, while the score system allows to rank the
remaining laboratories according to the expected testing capability.
A Questionnaire was used, including questions about the lab experience, testing
capability, whether accredited for testing appliances, to EN 17025, available
instruments, etc. For laboratories passing the knock-out criteria the final score was
calculated, allowing for a ranking against a rating scheme. The scoring system was
created by giving a “weight” (score) to the answers to the Questionnaire. From an initial
list of 23 laboratories that were contacted (Figure 3), 15 replied to the ad-hoc
Questionnaire, 10 were selected having passed the knock-out criteria and achieved a
defined score (25% highest ranking).
6 The list of the tested models is given in the Atlete web site at:
http://www.atlete.eu/index.php?option=com_content&view=article&id=121&Itemid=117 .
8 | Atlete Project - Final Publishable Report
Figure 3: EU laboratories contacted for the ATLETE project
The final selection of the 4 laboratories iPi (Germany), LCOE (Spain), RegenT (The
Netherlands) and VDE (Germany) was done on the basis of the offers received to a Call
for Tender and the a final inspection of project expert and an interview with the
laboratories representatives.
2.5 Verification procedure
2.5.1 The two step compliance verification procedure
The verification procedure is outlined in Figure 4. According to the essential
requirements described in the EU legislation on labelling/ecodesign it foresees two
testing steps plus an intermediate phase of voluntary corrective action (this last
voluntary corrective action has been introduced in ATLETE Project only).
Atlete Project - Final Publishable Report | 9
Figure 4: Verification procedure scheme as developed within ATLETE
In particular:
• Sample gathering: test laboratories searched for and purchased the needed units of
each selected model
• Step 1: testing of 1 unit per selected model
− if all labelling declarations comply with the outcome of the tests, the model is
considered as compliant
− if the unit fails even only one of the labelling declarations (or is not able to
accomplish the test for the specific product), Step 2 is triggered
− the relevant producer is informed about “suspected non-compliance” and asked
for checks of possible declaration mistakes. The producer could choose to accept
a non-compliance (and to correct the labelling declaration accordingly) or to
proceed to Step 2.
• Step 2: three other units of the same model are purchased from the market and
tested again (against a reduced tolerance for energy consumption)
• Possibility to assist to the test: project partners, IAC members or a representative of
the manufacturer of the model under test (under supervision of the test laboratory)
had a possibility to assist the test of given model. In case of dissent with the testing
conditions followed in the laboratory, the project partner/IAC
member/manufacturer reported back to the project leader (later to the market
surveillance Authority) before the result of the Step is known.
• Test reports: the testing laboratories reported the test results to the project leader
with a copy of the test report for Step 1 (and Step 2), and considered such results
confidential.
• Actions for non-compliant models: were foreseen for both Step 1 and Step 2:
‒ after Step 1: notification to national Market Surveillance Authorities, together
with voluntary actions taken i.e. for manufacturer accepting the non-compliance,
no legal actions after voluntary correction of the labelling declaration(s) were
taken
1 unit tested(Step1)
Random Products selection
Pass?
Market SurveillanceAuthorities
in the countries where the product
was availableCorrective actions?
3 additional units tested (Step 2)
YES
YES
NO
NO
Notification of compliance
YES
Pass?
NO
Notification of non-compliance & remedy action
Notification of compliance
Notification of non-compliance
10 | Atlete Project - Final Publishable Report
‒ after Step 2: notification to national Market Surveillance Authorities for legal
actions after the end of all tests: non-compliance was disclosed.
2.5.2 Protocol on the manufacturer’s pro-active participation to the ATLETE Project
All manufacturers of the selected models were invited by CECED to sign a protocol on
the manufacturer’s pro-active participation to the project where they accepted to take
proactively remedy actions would the tested products fail to show compliance. The
protocol stated that, in case the energy labelling/eco-design declarations of one of
appliance models were found to exceed the permitted verification tolerance of the
relevant testing standard after the first or second Step of the verification procedure, and
without prejudice to the ability to challenge the factual findings made in the testing
procedure, before its result is made available, the manufacturer had the possibility to:
• take within 30 (calendar) days after being informed of the results of the testing
procedure all the remedy actions necessary to correct the energy labelling/eco-
design declarations of the appliance model/s concerned in accordance with the
results of the testing procedure; and
• inform within the same timeframe the ATLETE project leader of the remedy
actions taken to correct the energy labelling/eco-design declarations, and
provide the ATLETE project leader with the appropriate evidence of these
remedy actions, such as a copy of a letter sent to the trade, with a proof of the
sending thereof, and a copy of brochures or leaflets marketing or advertising the
concerned appliance model/s with the correct energy labelling/eco-design
declarations, without disclosing any information which would be confidential.
In particular the protocol included:
• Manufacturer “obligations”:
o acceptance ex-ante of the reliability of the testing laboratory
o consideration of Step 1 results for possibly setting immediate remedy
actions
o set remedy actions for non-compliant models.
• Manufacturer “positive feedback”:
o possibility of assisting the test of own models and to raise concerns on
the regularity of the test before the result is known
o possibility to have a more lenient (or even no) sanction for non-compliant
declarations.
In some cases the voluntary remedy actions notified by the manufacturers to the ATLETE
project leader were verified through additional laboratory tests during the project field
work.
Atlete Project - Final Publishable Report | 11
2.6 Tested parameters and harmonized standards
The tests were conducted on the following five parameters:
Parameter Description
6. Energy consumption energy consumed by a refrigerating appliance over the
period of 24 hours)
7. Storage temperature (&
climate class)
the ability of an appliance to maintain simultaneously the
required storage temperatures in the different
compartments at a certain ambient temperature
8. Storage volume the space to store food in the appliance
9. Freezing capacity the amount of food, expressed in kilograms, that can be
frozen to a core temperature of -18 °C in 24 h
10. Temperature rise
time
the time period needed to raise the temperature of food in
the frozen food compartment from -18 °C to - 9 °C after the
operation of the refrigerated system has been interrupted
To this end the harmonized standards: EN 153:2006 & EN ISO 15502:2005, concerning
the testing normative for refrigerators and freezers ruled by the 2003/66/EC Energy
Labelling Directive and that Eco-labelling Regulation, have been applied by the selected
laboratories.
2.7 Verification results reporting and dissemination
Different opinions do exists about the opportunity to make the results of a compliance
verification action publicly available, disclosing also the model and the supplier name.
A compliance verification action is needed to prevent manufacturers who break the law
from gaining a competitive advantage over those that adhere to it. In this respect the
full disclosing of the verification results, once the procedure has been completed and
the final judgement on the tested models is reached, can only made public aware of
those products and suppliers that do not follow the rules. This is the approach followed
since many years at international level by the Australia’s market surveillance authority.
On the other side the USA has different experience. An unfair market advantage can
derive to products and suppliers meeting the requirements over the competitors not
having been selected for the verification action.
At EU level all Member States having developed market surveillance have not disclosed
the model and the supplier name. Although this was in part due to the fact that in some
cases only the first step of the verification procedure was run and therefore no
conclusion could be drawn on the actual compliance of the tested models.
Although each national Market Surveillance Authority may decide a specific position on
this issue, due to the subsidiarity nature of the controls within each Member State. It is
nevertheless recommended that some guidelines are given by the European
Commission or the ADCO Group on ecodesign and labelling.
12 | Atlete Project - Final Publishable Report
For what concerns the ATLETE project all compliant and non-compliant models have
been disclosed to:
• relevant manufacturers and project partners
• EU national Market Surveillance Authorities of the countries where each model
was reported to be sold
• stakeholder experts via the Final Conference, held on 12th
April 2011 in Brussels
and the national events (such as the national workshop held in Rome on 21st
June)
• general public: through media: interviews, press releases, articles, workshops,
etc.
Most important media dissemination actions have been uploaded to the project website
www.atlete.eu
2.8 Adaptation of the methodology to other Energy Related Products
The modularity of the proposed methodological approach makes the adaptation to
other appliances and products feasible with (minor) adaptations.
In Table 1 the elements to be modified in order to transfer the methodology from
refrigerating appliances to the other energy related products covered by a
labelling/ecodesign legislation are qualitative described.
Atlete Project - Final Publishable Report | 13
Table 1: Elements to be modified to transfer the methodology from refrigerating appliances to the other appliances
Products Geographical
scope
Technical
scope
Sampling
criteria
Laboratory
selection
criteria
Verification
procedure
Actions
after non-
compliance
Harmonised
standard Reporting
REFRIGERATING APPLIANCES OK OK OK OK OK OK OK OK
WASHING MACHINES if necessary,
to be adapted
to be
adapted to
the specific
product
if
necessary,
to be
adapted
procedure
valid,
selection
criteria to
be adapted
to the
specific
product
OK OK
specific
standard to be
used for
testing
to be adapted to the specific
product
DISHWASHERS if necessary,
to be adapted
to be
adapted to
the specific
product
if
necessary,
to be
adapted
procedure
valid,
selection
criteria to
be adapted
to the
specific
product
OK OK
specific
standard to be
used for
testing
to be adapted to the specific
product
SIMILAR HOUSEHOLD
APPLIANCES
if necessary,
to be adapted
to be
adapted to
the specific
product
if
necessary,
to be
adapted
procedure
valid,
selection
criteria to
be adapted
to the
specific
product
OK OK
specific
standard to be
used for
testing
to be adapted to the specific
product
14 | Atlete Project - Final Publishable Report
Products Geographical
scope
Technical
scope
Sampling
criteria
Laboratory
selection
criteria
Verification
procedure
Actions
after non-
compliance
Harmonised
standard Reporting
TV SETS if necessary,
to be adapted
to be
adapted to
the specific
product
if
necessary,
to be
adapted
procedure
valid,
selection
criteria to
be adapted
to the
specific
product
OK OK
specific
standard to be
used for
testing
to be adapted to the specific
product
HCV HOUSEHOLD
APPLIANCES
if necessary,
to be adapted
to be
adapted to
the specific
product
probably to
to be
adapted
procedure
valid,
selection
criteria to
be adapted
to the
specific
product
OK OK
specific
standard or
calculation
methodology
to be used for
testing
major changes to be
adapted to the specific
product
OTHER PRODUCTS if necessary,
to be adapted
to be
adapted to
the specific
product
probably to
to be
adapted
Probably
both
procedure
and
selection
criteria to
be adapted
OK OK
specific
standard or
calculation
methodology
to be used for
testing
major changes to be
adapted to the specific
product
Atlete Project - Final Publishable Report | 15
3 FINAL RESULTS OF THE FIELD WORK
As already outlined, 80 models of 40 manufacturers, present on the European Union
market, have been tested to verify their compliance with the EU energy label in four
European laboratories. For each model the tests were conducted on the five parameters
described in paragraph 2.6, two of which (freezing capacity and temperature raise time)
have rarely been examined by when the energy label has been introduced in 1995.
The final test results show that 79% of appliances subjected to testing and for which
testing has been concluded complied with the energy efficiency class declaration and
the two related key parameters: energy consumption and storage volume (77% and 90%
respectively). But when all five parameters are taken into consideration 57% of them do
not comply with at least one of the tested parameters
In accordance with the test outcome, the selected models have been classified in the
following 8 categories
Models classification by
tests outcome
Description
Models non-compliant for
formal reasons
Products are non-compliant for “formal” reasons when the Label and/or
the product fiche are lacking or not fulfilling the declared energy
efficiency class when it is re-calculated using the declared values against
the formula provided in the labelling scheme. The ATLETE team decided
to consider a model as non-compliant if one or both of the above
informative fiches were missing or were not correct, whatever are the
final results of the tests
Models found (or
suspected) to be damaged
or defective
During both Step 1 and Step 2 a unit of a model could result damaged or
defective either before a test is run (at a visual inspection or suspected)
to be damaged or defective as a consequence of a deviating behaviour or
an unexpected test result. Although this event was thought to be very
uncommon, we experienced it during the ATLETE project. Depending on
the time when the defect or damage is found or suspected the reaction
could be the repetition of Step 1 (or Step 2) on a different unit(s) of the
same model or to ignore Step1 results and go directly for the Step 2
(should a sufficient number of additional units of the same model be
available on the market).
Models compliant after
Step 1
If the products are considered fully compliant with the EU energy label no
further action is needed. The models are immediately notified to the
national Market Surveillance Authorities of the Member States where
they were reported to be sold.
Models suspected of
non-compliance after
Step 1
If a product, having passed the initial check for the formal compliance,
has been found either not fulfilling the conditions established by the
standard or exceeds the permitted tolerances for this verification Step, it
is suspected to be not compliant and further action is needed to confirm
or negate this suspect. In this case the supplier of each model is
contacted and invited to:
a. accept the suspected non-conformity and go for a voluntary
remedy action
b. do not accept the suspected non conformity and ask for the
development of the second phase (Step 2) of the verification. In
16 | Atlete Project - Final Publishable Report
Models classification by
tests outcome
Description
this case the laboratory was asked to purchase three additional
units of the models and to test them according to the Step 2 of
the verification procedure
c. comment the test report asking to the laboratory to modify the
test conditions (considered for example not in line with the
standard). In some few cases the ATLETE team considered the
remarks legitimate and the test was repeated.
Models suspected of
non-compliance after
Step 1 that have
underwent
“manufacturers’
voluntary remedy action”
(point "a" above)
Such action was in the form of the correction of the wrong declaration or
of the instruction manual, or even to go as far as stopping the production
of that specific model. After the remedy action in general no further
action is needed, but in some selected cases the remedy action has been
further checked through an additional laboratory test. In any case, the
products were considered non-compliant with the EU energy label and
were immediately notified to the national Market Surveillance Authorities
of the Member States where they were reported to be sold, along with
the description of the remedy agreed by the supplier and the evidence of
its actual application
Models suspected of
non-compliance after
Step 1 but for which 3
additional units are no
more available on the
market
These models are considered “non-decidable” because, despite all the
attempts done by the relevant laboratory the 3 additional units essential
for the development of Step 2 could not be found on the EU market.
Therefore no final judgement about the compliance of the models could
be reached. These products were in any case notified to the national
Market Surveillance Authorities of the Member States where they were
reported to be sold
Models non-compliant
after the complete
verification procedure
(Step 1+ Step 2)
For the models having failed Step 1 and (i) for which the suppler has
asked to go for Step 2 or (ii) if they belong to manufacturers that did not
sign the protocol on the manufacturer’s pro-active participation to the
ATLETE project, the verification Step 2 was developed. Models for which
at least one of the 3 additional tested units does not fulfil all the
conditions established by the standard or if even one of the tested
parameter exceeds the permitted tolerances for this verification Step are
considered non-compliant with the EU energy labelling.
Models fully compliant
after the complete
verification procedure
(Step 1+ Step2)
Models for which all measure parameters one the three units fulfilling the
conditions established by the standard and do not exceed the permitted
tolerances for this verification Step are considered fully compliant with
the EU energy label and no further action is needed. The models are
immediately notified to the national Market Surveillance Authorities of
the Member States where they were reported to be sold.
The number of models covered in each of the 8 categories above outlined is shown in
Atlete Project - Final Publishable Report | 17
Table 2:
18 | Atlete Project - Final Publishable Report
Table 2: Overall results of the laboratory tests
Number of models tested models 827
Models formally non-compliant* 10
Models compliant after Step 1 26
• of which re-tested on a different unit, due to a defect or
suspected damage 1
• of which re-tested due to test conditions non fulfilling the
standard 3
Model for which Step 1 was not developed 2
Models suspected to be non-compliant after Step 1, Step 2 needed 54
• of which voluntary remedy actions was agreed before running
Step 2 18
→ of which the voluntary remedy action was tested in
laboratory 2
• of which Step 2 testing was planned 36
→ of which Step 2 was run 22
� of which only Step 2 was run 2
→ of which Step 2 was not run 14
� of which Step 2 not run, 3 additional units not available 10
� of which Step 2 not run, 3 units of another model
received 2
� of which Step 2 not run, lack of reference values on
label 2
Models compliant after Step 2 4
Models non-compliant after Step 2 18
The analysis of the tests results are summarized in , in which the overall picture of the
pan-EU compliance of refrigerators and freezers is presented. Although the compliance
rate for the single parameters is in the range 70-90%, the overall compliance rate is
unacceptably low at 43%, due to the combined effect of the single non-compliances.
The presented results have been calculated on the 70 models for which the verification
was completed, while the single compliance rates have been calculated on the number
of models for which the specific parameter applies, because temperature rise time and
freezing capacity are relevant only for freezers and refrigerator-freezers. When these
two parameters not directly linked with the energy consumption (temperature rise time
and freezing capacity) are excluded from the statistic, the overall compliance rate raises
to 50%, due to the fact that 5 models have been found non-compliant only for the
freezing capacity and/or the temperature raise time.
7 In reality 82 models were tested at the end of the project instead of 80, because, for two appliances, the
three additional units purchased for developing Step 2 belonged to a different model.
Atlete Project - Final Publishable Report | 19
Figure 5: Overall compliance results of the ATLETE project
These results can be broken down by the different attributes in base of which the tested
models can be classified (i.e. by country, model type, energy efficiency class). A very
detailed analysis is provided in the deliverable: "Outcome of the pan-EU compliance of
refrigerators and freezers" available in the ATLETE website8." Here some spots are
provided mainly focusing on the energy efficiency and the storage volume compliance
rate.
For what concerns in particular the compliance with the Efficiency Classes, Errore.
L'origine riferimento non è stata trovata. shows that with respect 70 models for which
the steps 1 and 2 have been carried out, 55, that is the 79%, have correct energy
efficiency class declaration (the green labels in the figure) while 10 tested models (14%)
were overrated by 1 class and 4 tested models (6%) were overrated of 2 or even more
classes. 13 models were unfortunately classified as undecidable (the yellow labels) as no
3 equivalent models were found in the market at the time to carry out the second
testing step.
8 See: http://www.atlete.eu/index.php?option=com_docman&Itemid=111
79%
77%
90%
73%
84%
70%
43%
21%
23%
10%
27%
16%
30%
57%
0% 20% 40% 60% 80% 100%
Energy class
Energy consumption
Storage temperature
Storage volume
Temperature rise time
Freezing capacity
TOTAL
PASS
FAIL
20 | Atlete Project - Final Publishable Report
Figure 6 Efficiency classes declared / verified
More in detail, Figure 7 shows the percentage difference between the declared and the
measured value of the energy consumption in Step 1 and Step 2 (for Step 2 the average
of the three measured values) for the 82 models of refrigerating appliances.
A part from the models largely exceeding the permitted tolerances, it is worth noting
that for some models (for example models N. 6, 18 and 50) the difference between the
rated energy consumption and the average measured energy consumption of Step 2 is
much smaller that the difference found in Step 1, but also the contrary happens (for
example models N. 36, 37, 49 and 72).
For sake of details and to understand the importance to allow a tolerance range for the
energy efficiency compliance, it is worth mentioning also Figure 8 that shows the energy
consumption difference between the rated energy consumption and the measured one
for Step 1 and for the single units of Step 2. It is evident that for some models, for
example in models N. 23, 31, 50 but also in models 35 and 57 there is a significant
difference among the four tested units.
Another critical parameter influencing the energy consumption is the storage volume. It
is actually clear that the energy consumption of a model can be increased (in order to
include the model in a higher energy efficiency class) by delivering a storage volume
lower than that actually declared.
Figure 9 shows the percentage difference between the declared and the measured value
of the storage volume, for the 82 models of refrigerating appliances. The black line
highlights the 3% max allowed tolerance. To this end it is worth noting that it is not by
chance that the major differences concern the freezer volume, that is the component
that consumes the more in the refrigerating appliances
Legenda: Undecidable Compliant Non Compliant
Atlete Project - Final Publishable Report | 21
Figure 7 Difference (%) between the declared and the measured value of the energy consumption in Step
1 and Step 2 for the 82 models of refrigerating appliances
Figure 8 Difference (%) between the declared and the measured value of the energy consumption in Step
1 and Step 2 (single units and average values)
Figure 9 Difference (%) between the declared and the measured value of the storage volume for the
refrigerator and freezer compartments for the 82 models of refrigerating appliances
-30,0-25,0-20,0-15,0-10,0
-5,00,05,0
10,015,020,025,030,035,040,045,050,055,060,0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
Dif
fere
nce
in e
nerg
y co
nsum
ptio
n (%
)Step 1 Step 2
-35,0-30,0-25,0-20,0-15,0-10,0-5,00,05,0
10,015,020,025,030,035,040,045,050,055,060,065,070,075,0
6 7 18 23 30 31 33 35 36 37 44 45 49 50 55 56 57 58 62 72 81 82
Dif
fere
nce
in e
nber
gy c
onsu
mpt
ion
(%)
Step 1
Step 2, 1st unit
Step 2, 2nd unit
Step 2, 3rd unit
Step 2, average
-30,0
-25,0
-20,0
-15,0
-10,0
-5,0
0,0
5,0
10,0
15,0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
Dif
fere
nce
in s
tora
ge v
olum
e fo
r th
e re
frig
erat
or a
nd fr
eeze
r co
mpe
rtm
ents
(%)
refrigerator volume
freezer volume
22 | Atlete Project - Final Publishable Report
4 NON COMPLIANCE ANALYSIS
4.1 Identification of the non-compliance cases
In chapter 3 a classification of the tested models in accordance with the outcomes of the
tests themselves has been outlined. A case by case analysis of the tests results and the
consequent actions undertaken on these models is provided in detail in the quoted
deliverable on the "Outcome of the pan-EU compliance of refrigerators and freezers9".
To provide a good overview on the work carried out, the decision taken and the
observed outcomes of the testing campaign, it is enough to highlight here just three of
the eight cases in which the tested models have been classified. These cases refer:
• the first to the models formally not compliant with the Energy Labelling Directive
rules, that shows the numerous manner in which these rules can be not
observed;
• the second to the remedy actions undertaken by the manufacturers that have
signed the ATLETE "Protocol for manufacturers’ proactive participation”, that
shows that good results can be achieved when working in close collaboration
with the direct involved actors;
• the third to damaged or defective models due to external accident or
production/assembling problems. These cases are not common but they exist
(we had two cases of this type) and it is important to correctly deal with them.
4.1.1 Models formally non-compliant before the laboratory tests
Before starting the actual test, the laboratory checked each model for the compliance
with the more formal, but nevertheless mandatory, elements of the EU energy labelling.
Among the 82 models tested within the ATLETE project 12 models have shown some
formal non-compliance:
• lack of the label (the strip) or the product fiche in the technical documentation
accompanying the model
• presence of the label and the fiche but lack of one or more of the parameters to
be mandatory declared; the missing value(s) can be supplied at a later stage by
the manufacturer
• declared energy efficiency class not in line with the other declared parameters
(i.e. the energy efficiency class does not correspond to the Energy Efficiency
index calculated using the formulae provided in the labelling scheme with the
values of the declared parameters.
One or more of the above formal non-conformities can be present in the same model. In
particular:
• no “temperature rise time” in the fiche: 4 cases
• the energy efficiency class declaration is not correct: 2 cases
9 See: http://www.atlete.eu/index.php?option=com_docman&Itemid=111
Atlete Project - Final Publishable Report | 23
• label not available with the appliance: 4 cases
• label with incorrect declarations: 2 cases.
The decision of the ATLETE team was to proceed with the laboratory testing to evaluate
the compliance with the labelling declarations as much as possible, but whatever would
it been the outcome of the tests, to consider in any case the model as non-compliant (at
least from the formal stand point).
4.1.2 Models that underwent a “manufacturers’ voluntary remedy action
Manufacturers of models suspected of non-compliance after Step 1 and having signed
the “Protocol for manufacturers’ proactive participation” were given the possibility to
accept the non-compliance results of Step 1 and go for a “voluntary remedy action”. As
said, in 18 cases this remedy action was put in place and documented:
• in 1 case the correct energy efficiency class was declared on the label, in the
product fiche and the other documentation accompanying the model
• in 5 cases the correct value of the freezing capacity was declared in the product
fiche and the other documentation accompanying the model
• in 3 cases the correct temperature raise time was declared in the product fiche
and the other documentation accompanying the model
• in 3 cases the correct value of the storage volume of the freezer compartment
was declared on the label, the product fiche and the other documentation
accompanying the model
• in 1 case the correct value of the storage volume was declared on the label, the
product fiche and the other documentation accompanying the model
• in 2 cases several parameters were non-compliant and the correct valued were
declared on the label, the product fiche and the other documentation
accompanying the model.
• in 3 cases the production of the model was discontinued, and in one of them the
manufacturer committed himself to declare the correct values of the non-
compliant parameters for the equivalent models still placed on the market.
The remedy action was supported by the manufacturer with some evidence, such as a
copy of the new and correct label, fiche or instruction manual.10
4.1.3 Models (suspected to be) damaged or defective
One or more units of a model under test could result damaged and/or defective either
before a test is run (at a visual inspection) or suspected to be damaged / defective as a
consequence of a deviating behaviour or an unexpected test result. Depending on the
time when the defect or damage is found or suspected the reaction could be the
repetition of Step 1 (or Step 2) on a different unit(s) of the same model or to ignore Step
10
The ATLETE had not the possibility nor the responsibility to verify that the new and correct
documentation has been actually enclosed in all models on the market or that the model claimed to be
discontinued was not sold with a different name/commercial code number
24 | Atlete Project - Final Publishable Report
1 results and go directly for the Step 2 (should a sufficient number of additional units of
the same model be available on the market).
Although this event was thought to be very uncommon, we have experienced it twice:
• the unit did not pass Step 1 due to an apparent malfunctioning of the thermostat
that could be due to a damage of the specific unit or to a production defect. It
was decided to run Step 2 with three additional units as the best approach to
state that the non-compliance was due to the damage of the first tested unit and
not to a systematic defect of the model. The test run on the three additional
units confirmed that the first tested unit was damaged, but showed also that the
model was non-compliant.
• the model tested did not pass Step 1 for the freezing capacity (declared at 7kg).
Since a defect was found in the mounting of the freezer compartment gasket ()
that could have been the reason for the failure, it was decided to run again Step
1 on an additional unit of the same model.
Figure 10: Defective gasket of the freezer compartment door
This second unit, with no defective gasket, resulted fully compliant. But unfortunately,
due to the transition between the old and the new labelling scheme, the second unit,
although having the same commercial code of the first unit, showed a different declared
value for the freezing capacity (4kg) and bore the new energy label. Therefore, although
the model represented by the second unit is fully compliant to the energy labelling, a
doubt still remains on the model represented by the first unit, mainly because the failing
parameter for this first unit is the freezing capacity that has been modified (reduced) in
the new declaration on the energy label.
4.2 Analysis of the non-compliance macro-sources
The wealth of data provided by the test results of the ATLETE project allow, for the first
time, to run a semi-quantitative analysis of the non-compliance causes. Four main issues
can be analysed:
• the product category;
• the purchasing price;
• the manufacturing place;
• the expected energy efficiency class.
Atlete Project - Final Publishable Report | 25
4.2.1 Product category
The 82 tested models belong to the following types:
• refrigerator-freezers (4 stars): 44 models
• simple refrigerators: 11 models
• chest freezers: 9 models.
• upright freezers: 8 models
• refrigerators with a low temperature compartment11
: 4 models
The compliance rate has been calculated for each of these types to evaluate if a specific
product type presents a more marked compliance rate compared to the other types. The
results of this analysis are presented in Table 3 where it is evident that the higher non-
compliance risks are concentrated to the low temperature compartment refrigerators
(but this indication has to be taken with care because the sample of this appliances type
is very small) and to the freezers.
Table 3: Summary of the results of the compliance verification of the different types of refrigerating
appliances
Product types PASS FAIL
Simple refrigerators
‒ for 1 model having failed Step 1 three additional units
were not available
70% 30%
refrigerators with a low temperature compartment
& refrigerator-freezers
‒ for 4 models having failed Step 1 three additional units
were not available
34% 66%
upright freezers
‒ for 5 models having failed Step 1 three additional units
were not available
45% 55%
chest freezers
‒ for 2 models having failed Step 1 three additional units
were not available
60% 40%
TOTAL 43% 57%
4.2.2 Purchasing price
The average price of the 80 tested models is 396,45 € as show in
Table 4 along with the disaggregation per appliance category (categories with similar
characteristics have been grouped).
It is worth noting that the average price of the compliant model is significantly higher
(+125,85 € ) than the average, while on the contrary the average price of the non-
compliant models is lower than the average (-80,42 € ) and the same happened for the
11
From 0 to 2 stars
26 | Atlete Project - Final Publishable Report
models where 3 more units were not available on the market (-59,95 € ). The same
pattern results for the different product categories: the compliant models are always
more expensive than both non-compliant ones and models where 3 additional units
could not be found on the market. This trend is more clearly appreciated in Figure 11 to
Figure 13: compliant models are usually more expensive than the average, but there are
also appliances with a purchasing price below the average that can be compliant,
although less frequently.
Although due to the limited dimension of the ATLETE models sample the conclusion of
the project could not be fully applied to all refrigerating appliances sold annually on the
Community market but, from this first analysis, it seems that the purchasing price is a
good clue of possible non-compliance: more expensive models have a higher probability
to be compliant, while on the contrary cheaper models are more often non-compliant.
Table 4: Average, min and max purchasing price (€) for compliant and non-compliant for the appliance
categories defined in the labelling scheme
Category All models Refrigerators and
freezers Low temp. compartment Freezers
Compliance Yes No 3 n.a. Yes No 3 n.a. Yes No 3 n.a. Yes No 3 n.a.
average price 396,45 451,15 337,91 294,08
522,30 316,03 336,50 621,40 360,33 425,75 454,88 239,83 317,14 387,00 176,25 115,00
min 89,00 107,00 115,00 232,00 139,00 243,00 234,00 123,00 124,00 89,00 107,00 115,00
max 2.085,00 890,00 636,00 2.085,00 890,00 613,00 771,00 445,00 636,00 1.096,00 222,00 115,00
second min 174,00 123,00 124,00 -- -- -- -- -- -- -- -- --
second max 1.096,00 776,00 613,00 -- -- -- -- -- -- -- -- --
Difference to
the average 125,85 -75,56 -59,95 170,25 -90,82 -25,40 116,97 -98,08 -20,77 92,92
-
117,83
-
179,08
Figure 11: Purchasing price for the tested models of refrigerators and freezers
12
Note: in red the average purchasing price of all models of the group
12
The compliant model with the highest price of 2.085,00 Euro has been excluded from the picture
0,00
200,00
400,00
600,00
800,00
1.000,00
1.200,00
0 1 2 3 4
Pur
chas
ing
pric
e (E
uro)
3 n.a. No Yes
Atlete Project - Final Publishable Report | 27
Figure 12: Purchasing price for the tested models of refrigerator-freezers and refrigerators with a low
temperature compartment
Figure 13: Purchasing price for the tested models of upright and chest freezers
Note: in red the average purchasing price of all models of the groups
4.2.3 Place of manufacturing
The manufacturing country or region is often, but not always, reported in the nameplate
attached at the back of each model. An analysis was ran comparing the manufacturing
place with the rate of compliance. Results are presented in Table 5 and Figure 14
Table 5: Place of manufacturing as and rate of compliance with the EU energy
labelling
Country of
manufacturing Compliant Non-compliant 3 n.a. Tot.
EU-Member State 14 7 3 24
EU generic 2 5 1 8
n.a. 6 16 6 28
non-EU countries 8 12 2 22
Total 30 40 12 82
0,00
100,00
200,00
300,00
400,00
500,00
600,00
700,00
800,00
900,00
1.000,00
0 1 2 3 4
Pur
chas
ing
pric
e (E
uro)
3 n.a. No Yes
0,00
100,00
200,00
300,00
400,00
500,00
600,00
700,00
800,00
900,00
0 1 2 3 4
Pur
chas
ing
pric
e (E
uro)
3 n.a. No Yes
28 | Atlete Project - Final Publishable Report
Figure 14: Place of manufacturing and compliance with the EU energy labelling
Although, again, the dimension of the sample in not sufficient to draw statistically sound
conclusions, the trend is very clear: when the country of manufacturing is not indicated
(n.a. in Table 2 and Figure 6) or indicated in a very generic way such as “EU” the
probability of non-compliance is higher than when the country is clearly stated. The
same occurs when the products manufactured in a specified EU Member State are
compared to the products manufactured in non-EU countries (China, Korea, Thailand
and Turkey in the ATLETE project).
4.2.4 Declared Energy Efficiency class
A further correlation appears to exist between the declared energy efficiency class and
the compliance rate: the higher is the energy efficiency class the higher is the probability
to find a compliant model. This is especially true for classes A+ and A (Table 6). For the
latter, being de facto the lowest and the most populated energy efficiency class
currently available on the market the compliance rate is also the lowest, when
compared to higher classes A+ and A++.
Since a correlation appears to exist between the purchasing price and the energy
efficiency class, as shown in previous Figure 11, in the end, the highest is the purchasing
price and the energy efficiency class the highest is the probability of compliance.
Table 6: Relation between the declared energy efficiency class and the compliance
rate for refrigerating appliances within the ATLETE project
Declared energy
efficiency class Compliant Non-compliant 3 n.a. Total
A++ 2 0 1 3
A+ 14 8 2 24
A 12 31 8 51
B 1 1 0 2
C 1 0 1 2
D and lower 0 0 0 0
Total 30 40 12 82
0123456789
101112131415161718
EU-Member State EU generic n.a. non-EU countries
Num
ber
of t
este
d m
odel
s
Compliant
Non-compliant
3 n.a.
Atlete Project - Final Publishable Report | 29
4.3 Further sources of uncertainty in the compliance verification
Several problems of different nature, some attributable to a somewhat market with no
rules or other to the interpretation of same Directives prescriptions, have been met
during the testing campaign. These problems, that have not hindered the
implementation of the verification procedure and the provision of the final results, are
highlighted in the following pages of this report to put in evidence how certain rules of
the reference Directives show their limits when challenged with the market reality.
4.3.1 The EU labelling Directive interpretation (1), concerning the models selection in
the market
According to the verification procedure of the EU labelling scheme13
the units of a model
to be verified shall be “randomly selected on the market”. The reason is that randomly
selected units should better represent the overall production of a model, and this is
particularly important when Step 2 is run.
But how the random selection concept can be applied to the actual purchasing of the
products models to be verified? During the ATLETE project the testing laboratories
purchased the appliances to be tested via on-line shops, because of the lower price
compared to other distribution channels and the relative easiness of the purchasing.
We nevertheless wonder if the models purchased in this way were really “randomly
selected”, since a situation where units belonging to the same production batch are
selected could frequently happen as shown in Figure 15 and Figure 16 in the following
page. In the second Figure the three units clearly belong to the same batch and were
produced very closely in time.
Figure 15: First unit of a model purchased for Step 1 of the verification
Figure 16 : Additional three units of the model purchased for Step 2 of the
verification
13
Described in the harmonised standard EN 153 for directive 94/2/EC and in Annex VII of delegated
Regulation 2010/1060/EU.
30 | Atlete Project - Final Publishable Report
4.3.2 The EU labelling Directive interpretation (2): rounding of measured values
During the ATLETE Project we have experienced the importance of a clear description of
the rounding for the measured values for the compliance verification. This is of great
importance for the models that are “border line” with the energy efficiency class
thresholds or accepted tolerance or specific minimum conditions requested by the
standard. The importance of rounding is highlighted by the following case:
− Annual energy consumption declared on the label: the EU labelling directive 94/2/EC
establishes (Annex I, point 2.V) that the annual energy consumption of a model is
declared on the label with no digits; it is calculated from the measured energy
consumption in 24h multiplied by 365 and the resulting value is then rounded. Since
no rounding criterion is described the rounding to the nearest integer should be
applied.
Annex I, point 2.V
V. Energy consumption in accordance with standards referred to in Article 1 (2) but
expressed in kWh per year (i.e. per 24 hours × 365).
The harmonised standard EN 15505:2005 prescribes that the measured energy
consumption per 24h is declared with two digits (Clause 15.4), but the measured
value is usually reported:
In laboratory practice the energy consumption per 24h is measured and reported
with a number of digits higher than two. The use of two or more digits in the
compliance verification calculation can lead to completely different outcome for an
appliance that with a declared annual energy consumption of 124 kWh/year:
(c) 0,3904658 kWh/24h×365=142,52 kWh/year, rounded to the nearest integer is 143 kWh/year or
+15,3% higher than the rated value, the model is non-compliant for the energy consumption
(d) 0,39 kWh/24h×365=142,35 kWh/year, rounded to the nearest integer is 142 kWh/year or +14,8%
higher than the rated value, the model is compliant for the energy consumption.
This situation can happen whatever is the allowed tolerance and for any parameter
to be verified if the performance of the tested model is very close to the minimum
requirement or the threshold value.
Atlete Project - Final Publishable Report | 31
4.3.3 Market problem (1): different labels and declarations for the same appliance
model
Apart from the specific and rare case (described in the following paragraph) when the
labels ruled by different legal provisions are shown on the units of the apparently the
same model, a more frequent situation can be finding units of the same model bearing
the same label layout but with different valued declared for the same parameters or
even with different parameters declared, as shown in Case 1 and Case 2 below.
a) Case 1: Same labels layout but with different declared values for the same parameters
as shown in Figure 17
Step1 Step 2
Label strip of unit 1 Label strip of unit 2 Label strip of unit 3 Label strip of unit 4
Serial N. 094403645 Serial N. 104104696 Serial N. 095004226 Serial N. 102203466
Figure 17: Label strips for the 4 units of the same models used in Step1 and Step2 of the verification
procedure
b) Case 2: Same labels layout but different declared values and different parameters
In this second case the four units of the model not only presented a different value for
the annual energy consumption, but also according to the label strip for the second
appliance used in Step 2 (units 3 in Figure 18) a 15 litre 4-star freezer compartment is
present, that instead is not present in the appliance.
32 | Atlete Project - Final Publishable Report
Step1 Step 2
Label strip of unit 1 Label strip of unit 2 Label strip of unit 3 Label strip of unit 4
Y2010W10 Y2010W10 B32 Y2010W37
Figure 18: Label strips for the 4 units of the same models used in Step1 and Step2 of the verification
procedure
4.3.4 Market problem (2): same commercial code number for technically different
appliance models
A refrigerator freezer with a freezing capacity of 7kg/24h and a declared energy
consumption of 219 kWh/year (Figure 19), failed Step 1 for the freezing capacity, but the
gasket of the freezer door was found defective. Therefore another unit of the same
model (same commercial code number) was purchased from the market by the
laboratory; but this second unit was labelled according to the new labelling scheme with
a freezing capacity of 4kg/24h and an energy consumption of 230 kWh/year (Figure 20).
The manufacturer was asked for about the differences of the label declarations of the
two units, apparently belonging to the same model. The answer was that with the new
regulation 1060/2010/EU the manufacturer has switched to the new energy label
declaration and has optimised most of its appliances according to it. For the specific
model under investigation this optimization has caused technical modifications which
then lead to a change of the energy consumption and freezing capacity compared to the
old label. The testing laboratory has bought the second unit at the end of May 2011 and
therefore tested a model with the new energy label.
Although this specific case was due to the change in the EU legislation on energy
labelling, a supplier is currently permitted to apply technical modifications and to
declare different values for any parameter of an appliance without necessarily change
the name and identification number of the model.
Atlete Project - Final Publishable Report | 33
Figure 19: Rating plate and label strip for the first unit of the refrigerator-freezer
Figure 20: Rating plate and label strip for the second unit of the refrigerator-freezer
4.3.5 Market problem (3): different suppliers for the same appliance model
For one of the models for which both verification steps were run, a unique supplier
could not be identified because the same appliance with the same commercial code
number was imported and commercialised by two different suppliers in different EU
Member States.
The four rating plates are shown in Figure 21: the first tested unit belong to a specific
importer while the other three (used in Step 2) to a different importer.
34 | Atlete Project - Final Publishable Report
First unit of a model purchased for Step 1 of the verification
Additional three units of the model purchased for Step 2 of the verification
Figure 21: Different suppliers for different units of the same model purchased in the same country
In this situation, which of the two suppliers will be – from the legal point of view –
responsible for the non-compliance or for the eventual correcting action or a possible
sanction? Should each suppler be responsible only for the units placed on the market
under its own responsibility?
4.3.6 Improvement of the test method
The lesson learned from the test campaign together with the input and suggestions
received from the selected laboratories (four meetings were actually organized) as well
as the outcome of the workshops held with the ATLETE advisory group14
have allowed
the ATLETE working group to finally formulate the following improvements for the
refrigerating appliances standard:
• to shorten the time required for the completion of the verification process, the
specific setting(s) used by the supplier for the measurement of the parameters at
the basis of the labelling declaration and the compliance with the ecodesign
requirements should be specified either in the labelling/ecodesign technical
report or to the booklet of instructions or to any other technical documentation
accompanying the products
• storage volume measurement is still critical, at least for some manufacturers and
products configurations. The need for further clarification should be evaluated by
the standardisation experts
14
An ad hoc Advisory Group composed by relevant stakeholders and representative of the ADCO group
has advised and supported the ATLETE project over all its life
Atlete Project - Final Publishable Report | 35
• the 2 star compartments identification is controversial and requires a further
clarification
• the use of “cold plates” (eutectic accumulators) should be ruled and the impact
on the load plan and the appliance volume measurement described. The use of
eutectic plates can be accepted but only in accordance with the already
established standard conditions for the volume measurement:
• the appliance storage volume should exclude the space needed for the plates
• the load plan should show the position of the plates, that cannot be placed
directly over the stacks
• rounding rules have to be both improved and better specified: not only for the
declaration of the parameters but also in the intermediate calculations
• the relation between the instruments uncertainty (usually described in the
standards) and the measured values is apparently not clear at least for some
manufacturers. A manufacturer claimed that the failing to comply – of 0,5K - with
a storage temperature cannot be accepted because the accuracy of the
measurement instrument is of the same order of magnitude. The answer of the
ATLETE team was that storage temperature must be respected beyond the
accuracy of the measurement instrument, once this accuracy is in accordance
with the standard specifications
• devices designed specifically to underestimate energy consumption under test
conditions should be clearly prohibited.
5 MAIN CRITICAL ELEMENTS OF THE PROPOSED PROCEDURE
After the completion of the field work and in the light of the achieved results, a critical
analysis of the applied evaluation procedure (see chapter 2) was developed with the aim
to correct the aspects that have presented uncertainty or shown to be excessively
cumbersome.
Some of the critical aspects have been already outlined in chapter 4, others were
highlighted by the testing laboratories, that at the end of the field work were asked
about their views about the verification procedure developed within the ATLETE project
and how it could be improved.
5.1 Models sampling, purchase and identification
In general, models scouting, purchase and logistic is a resource-consuming (human,
time, financial) task that requires market-skilled personnel and appropriate timing. The
scouting of the appliances to be tested should not be in charge of the testing
laboratories, but either of the national Market Surveillance Authority or of a delegated
subject.
36 | Atlete Project - Final Publishable Report
5.1.1 Models sampling
The sampling criteria should be defined to tackle also models placed on the market in
“small” batches by manufacturers and/or importers, to avoid that the lack of 3
additional units to be tested will impede the completion of the compliance verification.
Possible solutions are:
• immediate purchase of 4 units for each of the selected models, at least for
suppliers with a low market share. This approach will increase the costs of the
verification and in case a model passes Step 1 the other 3 units will be useless.
Nevertheless, if 3 units will be left untouched in the original packaging the
national Market surveillance Authority could give them back to the retailer or
even purchase the first unit and have the other 3 as a “free sample” to be paid
only if tested; this approach could be proposed to retailers only if the time to
finalise the test is acceptable, on the contrary the 3 units could remain unsold
• purchase of 3 units for each of the selected models, at least for suppliers with a
low market share, to run Step 2 as the unique verification stage; this approach
will increase slightly less the purchasing costs, but will increase the testing costs
since only Step 2 could be run
• delete from the list of selected models those not available in at least 4 units: this
approach will create areas with de-facto no compliance verification
• use “equivalent” models: this approach is applicable a sufficient number of units
of equivalent models(s) are available and if the “equivalent” models are actually
equivalent.
For the same reason, once the total amount of models to be verified has been decided,
the actual selection - including the publication of the selected models list - should not be
done at the beginning of the verification action but in several phases. This approach will
reduce the time-to-test (i.e. the time from the announcement that a specific model will
be tested to the time of the actual completion of the tests including the purchase of the
1+3 units) to a minimum, thus reducing the possibility for the model to disappear from
the market.
The actual meaning of “random” selection and if necessary the minimum criteria to
assure a random selection should be defined as much as possible, at least in a qualitative
way. In this respect:
• can units purchased in the same market place be considered randomly selected?
• are models purchased in an on-line shop eligible for a verification action? In fact
sometimes the time for delivering the on-line purchased goods is quite long and
may indicate that the products are directly looked for from the
manufacturer/importer. This event should be carefully avoided because to
assure a fully independent verification action.
• Models selection could be based not on “market analysis”, but on complaints or
evidence obtained in a pre-evaluation process (such as a risk analysis) to detect
products highly suspected of non-compliance, that are eventually selected for
testing.
Atlete Project - Final Publishable Report | 37
5.1.2 Models purchasing
Purchase of the selected models: the purchase of products in foreign countries and on
internet was applied for the ATLETE project for the two fold reason that consumers
often buy products on-line, and that four laboratories should procure models from all
European countries. But at national Member State level this purchasing process could be
not in line with the legal and administrative procedures. For example, the Consumer Law
in Spain that regulates the market surveillance procedures for the Energy Labelling
Authorities require to select the sample in the market (retailer warehouse, or shop) and
directly by an Authority Inspector that identify and mark the samples to be sent to the
specific testing laboratory. In addition this Law allows, under requirement of the
manufacturer or importer involved in the market surveillance that they may be present
when the laboratory opens the appliance package.
5.1.3 Models univocal identification
The ATLETE field work has shown that not always appliance models are univocally
identified by their commercial code number (or a similar coding). The challenge is to
ensure that units apparently belonging to the same model are actually the same
product. Or in other words how far two apparently identical units of a model can have
different characteristics (different components, technical details, performance
declarations, etc.) and still be considered two units of the same model? In this respect a
better system of univocal identification of models should be developed.
5.1.4 Laboratories selection and remuneration
The number of selected laboratories for a verification action should be adequate:
• to the dimensions of the sample to be tested, to avoid saturating the laboratories
for some weeks/months
• to the number of existing laboratories able to test the same product, that in turn
influences the overall number of models that can be actually tested.
Adequate remuneration should be paid to the laboratories, including for actions beyond
the mere models testing and results reporting if such actions are under the responsibility
of the laboratory. Knowing that - if selected - their work will be adequately remunerated
a larger number of laboratories will participate to the selection procedure thus
guaranteeing that the most skilled laboratories are selected and a stronger motivation
for laboratories to improve their testing equipment and ability.
5.1.5 Final test report
Laboratories provide test reports according to the applied standard(s) or legislation. But
test reports can be quite different in format and content and can provide a different
amount of information (i.e. temperature graphs, copy if the label and of the booklet of
instructions, etc.). This makes the comprehension not immediate and the comparison of
the tests run in different laboratories on the same products quite complex.
38 | Atlete Project - Final Publishable Report
In this respect, is a mandatory common test report format, i.e. something more detailed
and structured that what is currently described in the standard, be adequate? Or will it
prevent the necessary flexibility?
In the case of the ATLETE refrigerating appliances all laboratories have verified also the
compliance with the energy efficiency class. In some cases stating that considering the
allowed tolerance it was fulfilled, although using the measured values of the parameters
it was often lower. This is in our view exceeding the role and responsibility of the
laboratories, that should limit their action to only testing and certifying the compliance
of the measurable parameters.
The final responsibility for the model compliance with the (labelling/ecodesign)
legislation requirements should be under the responsibility of the national Market
Surveillance Authority or its delegated subject.
5.1.6 Measured parameters in Step 1 and Step 2 of the compliance verification
Within ATLETE project the more restrictive interpretation of the measurements to be
done in Step 1 and Step 2 of the verification procedure was applied, asking laboratories
to re-test four of the five parameters (the storage volume has not been re-measured
since the same result would have been achieved) of a model in Step 2. Although some
of the laboratories had a different opinion, we believe that only this approach can prove
beyond any doubt the compliance of a product.
The consequences of this approach falls also on the test reports to be delivered by the
test laboratories: in the Step 2 report the laboratory should provide the average values
of all the tested parameters and not only of the elementary test results for each unit.
The appropriateness of our decision was in the end confirmed by the results of the field
work: some of the models failed Step 2 not only for the parameters having failed Step 1
but also for additional parameters that before were instead compliant.
5.1.7 Improvement of the test method
The test procedure needs to be shared by all market actors before any enforcement or
verification testing begins. In addition, testing laboratories may have different
interpretations of specific test conditions. Since without consistency in test methods it
will be impossible to run an effective verification and enforcement programme the
standard should be unambiguous. If necessary, interpretations should be prepared to
clarify the still unclear conditions. On the other side a care should be taken by the
national Authorities and test laboratories that a claim of “ambiguity” is not being used
as a way out from too stringent testing conditions.
In particular, suggested improvements for the refrigerating appliances standard are:
• to shorten the time required for the completion of the verification process, the
specific setting(s) used by the supplier for the measurement of the parameters at
the basis of the labelling declaration and the compliance with the ecodesign
requirements should be specified either in the labelling/ecodesign technical
Atlete Project - Final Publishable Report | 39
report or to the booklet of instructions or to any other technical documentation
accompanying the products
• storage volume measurement is still critical, at least for some manufacturers and
products configurations. The need for further clarification should be evaluated by
the standardisation experts
• the 2 star compartments identification is controversial and requires a further
clarification
• the use of “cold plates” (eutectic accumulators) should be ruled and the impact
on the load plan and the appliance volume measurement described. The use of
eutectic plates can be accepted but only in accordance with the already
established standard conditions for the volume measurement:
o the appliance storage volume should exclude the space needed for the
plates
o the load plan should show the position of the plates, that cannot be
placed directly over the stacks
• rounding rules have to be both improved and better specified: not only for the
declaration of the parameters but also in the intermediate calculations
• the relation between the instruments uncertainty (usually described in the
standards) and the measured values is apparently not clear at least for some
manufacturers. A manufacturer claimed that the failing to comply – of 0,5K - with
a storage temperature cannot be accepted because the accuracy of the
measurement instrument is of the same order of magnitude. The answer of the
ATLETE team was that storage temperature must be respected beyond the
accuracy of the measurement instrument, once this accuracy is in accordance
with the standard specifications
• devices designed specifically to underestimate energy consumption under test
conditions should be clearly prohibited.
6 MARKET SURVEILLANCE PROCEDURES AT EU AND INTERNATIONAL LEVEL
AND COMPARISON WITH THE ATLETE METHODOLOGY.
In the quoted deliverable "Outcome of the pan-EU compliance of refrigerators and
freezers" as well as in the WP 2 deliverable: "Overview of the methodology and the
results achieved in the conformity assessment action at EU Member States and
international level15
", a detailed analysis of the international (Australia and USA) and
European legislation and practice on the verification procedures has been carried out. A
synthesis of the outcome of this analysis, as well as of a survey carried out at the
beginning of the ATLETE project to assess the manufacturers’ conformity practices in the
EU Member States, is outlined in this report to provide the picture of the EU and
International situation on the surveillance procedures and compare them with the
ATLETE methodology.
15
See for both deliverables: http://www.atlete.eu/index.php?option=com_docman&Itemid=111
40 | Atlete Project - Final Publishable Report
The extra-EU countries analysed have been Australia, New Zealand and the US. while for
three EU countries: Sweden, Denmark and UK (the only ones for which the data on the
surveillance activity are public available) the results achieved by market surveillance
actions developed for refrigerators and freezers in the last decade are briefly outlined.
6.1 Market surveillance procedures followed at international level
6.1.1 Australia and New Zealand
The Check Testing Process
In April 2011 the report “Performance Verified” was published in Australia. It presents
the results of 1.000 laboratory tests conducted between 1991 and 2010 undertaken to
verify the performance of appliances and equipment regulated for energy efficiency in
Australia and New Zealand. The two relevant types of regulatory programmes are
minimum efficiency requirements (called Minimum Energy Performance Standards) and
energy labelling, both of which are part of the Equipment Energy Efficiency (E3)
programme.
In Australia and New Zealand all products within the scope of energy labelling and
minimum efficiency requirements regulations must be registered and meet the specified
performance requirements. In the case of labelling, products are also required to display
the correct label indicating the product’s performance. Some individual products have
additional requirements.
Amongst other tasks, the E3 Committee administers a monitoring, verification and
enforcement programme to maximize compliance with energy efficiency regulations and
ensure that the overall efficiency goals of the programme are being met. These activities
are designed to raise the perceived risks of non-compliance by stakeholders in the E3
Programme, following the principles of deterrence theory16
Activities undertaken by the E3 Committee to monitor compliance include:
• aiding compliance through education, stakeholder forums and other
communication activities
• maintenance of a registration database and on-line registration facility (see
www.energyrating.gov.au)
• in-store surveys to check that the correct labels are being displayed
• checks to ensure that products on the market are registered
• and check testing: check testing or verification testing, is undertaken according
to the methodology specified by a test standard or protocol and is the only
means to confirm whether individual models meet the mandatory performance
requirements and that energy labels indicate the correct level of performance.
Check testing provides several important functions with respect to the energy
efficiency regulatory programme:
16
“Deterrence theory… maintains that there must be a credible likelihood of detecting
violations, swift, certain, and appropriate sanctions upon detection; and a perception
among the regulated firms that these detection and sanction elements are present”.
Atlete Project - Final Publishable Report | 41
o it helps to ensure that the projected energy and greenhouse gas savings
are delivered
o it safeguards the integrity of the programme, maintaining consumer and
industry confidence in the energy performance labels, and
o it protects the investment made by manufacturer’s and product suppliers
producing compliant equipment from being undercut by non-compliant
products.
The check testing process is similar to that envisaged in the EU Directives. It actually
comprises a stage 1 test (also known as the “screen test”) performed on one sample of
the model and of a stage 2 check testing to be carried out if the model is found to be
non-compliant. The sample to be tested is generally independently purchased (usually
through a retail outlet), paid for by the regulatory authority and tested by a laboratory
accredited for check testing17
. If, after test 1 the model is found to be non-compliant the
supplier has the choice of either cancelling the registration for the model or proceeding
to stage 2 check testing. Stage 2 involves the testing of either two further units (where
the failure relates to minimum efficiency requirements) or three units (where failure
relates to claims on labels) paid for by the registration holder (i.e. the subject that places
the product on the Australian market). The actual units to be tested in stage 2 are
randomly selected from the stock by a representative of the regulatory authority.
Guidelines18
explaining how the national mandatory scheme for energy labelling and the
minimum energy efficiency requirements are administered have been issued by and ad
hoc committee of the Australian and New Zealand Minerals and Energy Council.
Although not legally binding, the purpose of the Guidelines is to act as a guide to
facilitate uniform and consistent practice among central and regional regulatory
agencies and to explain to stakeholders the responsibilities of relevant State and
Territory regulatory agencies and the responsibilities of industry. The Guidelines include,
inter alia, a detailed description of the programme compliance monitoring through
laboratory check testing.
The E3 check testing programme aims to cover most major categories and types of
regulated appliances and equipment. Individual models are not randomly selected from
the stock of models sold, but are chosen according to a set of criteria; the criteria are
based on established ‘risk factors’ that indicate that a product has a higher probability of
failure compared to the market19
. These factors can be summarised as:
• suppliers and models with a demonstrated record of check testing non-
compliance because of the likelihood of a continuation of such historical trends
• where a third party, such as competitors, consumers, consumer groups or
regulatory agencies, provide evidence of non-compliance, for example the results
of in-house or independent laboratory tests
• products that appear as new brands on the market or from suppliers that do not
have any check testing track record
17
The test can be carried out only by laboratories accredited to the NATA (National Association of Testing
Authorities) or by other bodies with a mutual recognition agreement with NATA 18
“Administrative Guidelines for the Appliance and Equipment Energy Efficiency Program of Mandatory
Labelling and Minimum Energy Performance Standards”, Edition 5, June 2005, downloadable from:
http://www.apec-esis.org/www/UploadFile/Australian_admin_guidelines.pdf 19
Criteria are published on the public website at: http://www.energyrating.gov.au/checktest-criteria.html
42 | Atlete Project - Final Publishable Report
• models tested in previous years of the check test programme are excluded from
any further testing unless specific evidence becomes available to suggest that a
re-test is warranted.
Amongst products selected according to ‘risk factors’, products that are likely to have
the greatest adverse impact on energy and greenhouse gas savings are prioritised for
check testing. These include models belonging to the following groups:
• models with high volumes of sales because of their greater potential to impact
on energy usage as compared to models with low sales volumes
• models with the highest claims for energy efficiency (e.g. high star ratings)
because of the market’s higher expectations with respect to the performance of
these models as compared to models with low ratings
• newer models will normally be given preference when considering models for
check testing because of their potential to remain on the market for a longer
period as compared to older models.
Results of the check testing process
The years 2005-2010 have seen considerable expansion in the check test programme,
driven by a corresponding growth in the number of regulated product categories and
the number of individual products registered, which have increased considerably over
the past ten years. The number of stage 1 check tests undertaken each year has varied
considerably, from less than 10 to about 150, and this is partially explained by changing
budget allocations and the mix of products tested, due to the variation between the per
unit test costs for different product categories; 65% of all check tests have been
conducted on refrigerators and freezers, air conditioners, and clothes washers, which
were amongst the first products to be regulated (Figure 22)
Figure 22: Total number of check tests conducted in Australia by appliance category, 1991-2010
An overview of the results by product category , referring to 1,000 appliances tested, is
presented below.
Atlete Project - Final Publishable Report | 43
• 73% (730 products) were found to meet the energy performance requirements
contained in energy efficiency regulations
• the registration of 207 (21%) models was cancelled by regulators
• for an additional 1% of cases, action by regulators included the referral of three
products for cancellation or to the Australian Competition and Consumer
Commission.
• a further 4 failed products were found to be older stock imported or
manufactured prior to the introduction of new regulations and allowed to be
sold under grandfathering provisions
• 22 (2%) failed tests are either still awaiting resolution, the registration has
expired or were found not be registered at the time of testing
• suppliers either withdrew or corrected the registration details, or re-registered a
total of 39 (4%) products
• as far as the specific Stage 1 results are concerned, of the 388 products that
failed at stage 1 testing, 92 (24%) passed the stage 2 test, 19 (5%) were found to
be no longer sold and 2 (1%) are still awaiting a decision by the regulator.
6.1.2 USA
The Energy Policy and Conservation Act
The U.S. Department of Energy (DOE) regulations require manufacturers of products
covered under the Energy Policy and Conservation Act (EPCA) to submit compliance
statements and file certification reports to DOE verifying that the products meet the
relevant energy efficiency or water conservation requirements.
In late 2009, the Office of the General Counsel issued a new legal interpretation of its
enforcement authority under the EPCA which viewed improper certifications or failure
to file certifications as EPCA violations "subject to enforcement action, including the
imposition of civil penalties. Concurrently DOE has undertaken a new initiative to
aggressively enforce energy efficiency and water conservation requirements for a
variety of products, including air conditioning and heating systems, kitchen and other
household appliances, televisions, lighting, bathroom fixtures, dehumidifiers and battery
chargers.
In particular in October 2009, DOE announced three new steps to strengthen its ability
to enforce energy efficiency requirements:
• the formation of an enforcement team within the Office of the General Counsel
that comprises of lawyers with extensive litigation and regulatory experience.
• a programme to randomly review manufacturers' compliance with DOE
certification requirements. The enforcement team will be initiating a compliance
review of certification reports for covered consumer products. DOE will randomly
select previously filed certification reports for review, request certification
records as needed, and hold manufacturers accountable for failing to certify
covered products according to DOE rules
• the issuing of guidance further detailing its energy efficiency enforcement
regulations. In its new guidance on Energy Efficiency Enforcement Regulations,
44 | Atlete Project - Final Publishable Report
the Department confirms that under existing DOE regulations, DOE can take
enforcement action and assess civil penalties if a manufacturer fails to properly
certify a covered product and retain records. Specifically, the Department
clarifies that any failure to certify covered products according to DOE's rules
violates the Energy Policy Conservation Act of 1975 and DOE's regulations.
A new set of rules, overhauling the past ones, and concerning the enforcement of DOE’s
efficiency requirements for appliances, lighting and other products were then issued by
DOE in 2011. These new rules are designed to encourage compliance and prevent
manufacturers who break the law from gaining a competitive advantage over those that
adhere to it. Among other changes, the rules will - for the first time - allow the
Department to test products on its own initiative to determine whether they comply
with DOE’s efficiency standards. Using its reinforced enforcement authority, DOE has
already issued civil penalties and testing orders for covered products from several
manufacturers
The verification of the Energy Star Programme
American consumers, businesses, and federal agencies rely on the Energy Star
programme to identify products that decrease greenhouse emissions and lower energy
costs. In addition, the federal government and various states offer tax credits and other
incentives to encourage the use of energy-efficient products including Energy Star
products. Specifically, approximately 300 million USD from the American Recovery and
Reinvestment Act will be used for state rebate programmes on energy-efficient
products.
The Energy Star programme is overseen jointly by the U.S. Department of Energy and
the U.S. Environmental Protection Agency (EPA). The programme is currently based on
self-certifications by manufacturers with after-market tests and self-policing that ensure
standards are maintained.
Given the millions of dollars allocated to encourage use of Energy Star products and
concerns that the Energy Star programme is vulnerable to fraud and abuse, the U.S.
Government Accountability Office (GAO) was asked to conduct proactive testing to
obtain Energy Star partnership status for bogus companies and submit fictitious
products for Energy Star certification. GAO's investigation showed that Energy Star is for
the most part a self-certification programme vulnerable to fraud and abuse20
. In fact
GAO found that for the bogus products certification controls were ineffective primarily
because Energy Star does not verify energy-savings data reported by manufacturers.
On the basis of a new Memorandum of Understanding between DOE and EPA, the
programme was shifting toward a more rigorous up-front screening process. The first
phase of the initiative started in January 2011, with the launch of a new rule requiring
products to be tested for compliance by an EPA-approved third-party certification body.
Phase two involves verification testing, to make sure that products which are certified as
Energy Star-compliant stay that way.
20
For further information on this investigation see: http://www.gao.gov/products/GAO-10-470
Atlete Project - Final Publishable Report | 45
Starting January 1, 2011, new products that meet Energy Star specifications must be
certified by an EPA-recognized Certification Body (CB) before the product can be labelled
with the Energy Star mark (Figure 4). Upon certification of a product based on test data
provided by an EPA-recognized laboratory, the CB will notify the partner that the
product meets the Energy Star requirements and will submit the qualified product data
to EPA for listing on the Energy Star website21
. In addition to certifying products as
Energy Star, the CB is required to operate a verification programme ensuring on an
annual basis that a certain percentage of basic models it has certified continue to meet
the requirements through verification testing again using an EPA-recognized laboratory.
Testing laboratories are either directly EPA accredited or are EPA-recognized through
enrolling in a Certification Body's WMTL (witnessed manufacturers’ testing laboratory)
or SMTL (supervised manufacturers’ testing laboratory) programme. The list of
Certification Bodies and the accredited laboratories is available on the Energy Star
website.
In addition to the CB-run verification programmes, the DOE verification testing
programme will target certain Energy Star products that are also covered by DOE’s
regulatory programme such as: refrigerators, freezers, washing machines, dishwashers,
water heaters, room air conditioners, central air conditioners and heat pumps, furnaces
and televisions.
In March 2010, DOE launched a pilot action to verify the energy efficiency and water-use
characteristics of clothes washers, dishwashers, freezers, refrigerators, room air
conditioners, storage water heaters, and tankless water heaters through laboratory
testing. The pilot action was used also to validate the basic aspects of the verification
procedure.
The available results of the tests show that so far 110 models have gone through initial
energy testing. The agency plans to test an additional 157 models and expand testing to
other Energy Star products under long-term programs.
DOE randomly selected 20% of basic models (i.e. a model that includes all variations of a
unit made by a manufacturer that have the same energy-use characteristics) for testing,
and then modified the selection based on strategic considerations:
• basic models picked through random selection
• omitted basic models that were not available for sale through normal retail
distribution channels
Table 7shows the appliances types and the number of models to be verified.
Table 7: Final number of selected models for the DOE verification pilot programme
Appliance Number of basic
models
Target number of
tests
Refrigerators 405 81
Freezers 121 26
Washing machines 200 40
21
For additional information: www.energystar.gov/index.cfm?c=partners.enhanced_test_verification
46 | Atlete Project - Final Publishable Report
Appliance Number of basic
models
Target number of
tests
Dishwashers 54 11
Thankless WH 59 12
Storage WH 55 11
Room air conditioners 398 80
Total 1.292 261
The verification testing process begins with a Stage I test, then proceeds to a Stage II
test, if warranted (Figure 23):
‒ Stage I testing: screening tests, leading to no action or further actions
‒ Stage II: determining the compliance if action is required after Stage I
Figure 23: The two-stage verification testing process for the DOE pilot action
If the results of testing indicate that a product does not meet applicable minimum
Federal energy conservation requirements, DOE will initiate an enforcement action. This
enforcement action could lead to further testing, and ultimately, to the government
requiring that the product no longer be distributed in commerce in the U.S. If a product
does not meet Energy Star programme requirements DOE will refer the matter to EPA
for enforcement.
Table 8 shows the preliminary results of the pilot action. It is worth noting that about
17% of the tested models require further action.
Atlete Project - Final Publishable Report | 47
Table 8: Preliminary results for DOE the verification action
Appliance Total models Tested
models Units failing Stage 1
Refrigerators 82 66 8
Freezers 24 10 3
Washing machines 40 32 5
Dishwashers 11 10 2
Thankless WH 12 10 0
Storage WH 11 9 3
Room air conditioners 87 39 9
Total 267 176 30
6.2 Brief overview on the formal procedure followed in EU Member
States
The verification procedure for the old energy labelling of major household appliances -
still in force until when the provisions of the new delegated Regulations will mandatorily
apply - is based on a specific Annex of the applicable harmonised standard(s), which in
turn is mentioned in the product specific directive(s).
In the new labelling delegated Regulations, as well as in the ecodesign Regulations, the
essential elements for the verification procedure are instead explicitly mentioned (in
one if the Annexes) among the legislation provisions.
Whether in the case of an old implementing directive or a new delegated regulation, the
verification procedure is based on a two-step approach: in Step 1 the check is performed
on one sample of the model; in case of non-compliance Step 2 is developed, checking
three additional samples of the same model.
Depending on the parameter to be verified, a verification tolerance (that takes into
consideration the uncertainty in the laboratory measurements) is applied to both Steps.
In this respect it is worth noting that while in the old labelling directives the tolerance
accepted in Step 1 was larger than that accepted in Step 2, in the new delegated
regulations the allowed tolerance is the same in both Steps in most cases.
A summary of the EU verification system for the energy consumption for the old and
new energy consumption declarations in the energy labelling and energy efficiency
requirement schemes is presented in Table 9 and Table 10.
48 | Atlete Project - Final Publishable Report
Table 9: Summary of the EU verification system and tolerances for energy consumption in the previous
labelling and minimum requirements schemes
Appliance Implementing
Directives
Harmonised
standard
Verification procedure
Step 1 Step 2 Units Tolerance Units Tolerance
(n) (%) (n) (%)
(old) Energy labelling scheme
Refrigerators&freezers 94/2/EC/2003/66/EC EN 153 1 15% 3 10%
Washing machines 95/12/EC/96/89/EC EN 60456 1 15% 3 10%
Tumble dryers 95/13/EC EN 61121 1 15% 3 10%
Washer-dryers 96/60/EC EN 50229 1 15% 3 10%
Dishwashers 97/17/EC/99/9/EC EN 50242 1 15% 3 10%
Air conditioning 2002/31/EC EN 14511 1 15% 3 10%
Ovens 2002/40/EC EN 50304 1 40Wh+10% 3 10%
Efficiency requirements scheme
Refrigerators&freezers 96/57/EC EN 153 1 15% 3 10%
Table 10: Summary of the EU verification system and tolerances in the new labelling and ecodesign
schemes
Appliance Implementing
regulation Standard
Verification procedure
Step 1 Step 2 Units Tolerance Units Tolerance
(n) (%) (n) (%)
(new) Energy labelling scheme
Refrigerators&freezers 1060/2010/EU EN 153 1 3/10% 3 3/10%22
Washing machines 1061/2010/EU EN 60456 1 10%23
3 6/10%24
Dishwashers 1059/2010/EU EN 50242 1 10/19%25
3 6/10/19%26
Televisions 1062/2010/EU IEC 62087* 1 7% 3 7%
Ecodesing requirements
Refrigerators&freezers 643/2009/EC EN 153 1 3/10% 3 3/10%27
Washing machines 1015/2010/EC EN 60456 1 4/10%28
3 4/6/10%29
Dishwashers 1016/2010/EC EN 50242 1 10/19%30
3 6/10/19%31
Televisions 642/2009/EC IEC 62087* 1 7% 3 7%
*EN standard in preparation by European standardisation bodies under a mandate from the European
Commission
22
3% for gross volume and storage volume, 10% for all other parameters (except noise). 23
10% for all parameters (except noise). 24
6% for annual energy consumption, 10% for all other parameters (except noise). 25
19% for drying efficiency, 10% for all other parameters (except noise). 26
6% for annual energy consumption, 19% for drying efficiency, 10% all other parameters (except noise). 27
3% for gross volume and storage volume, 10% for all other parameters. 28
4% for washing efficiency, 10% for all other parameters. 29
4% for washing efficiency, 6% for energy consumption, 10% for all other parameters. 30
19% for drying efficiency, 10% for all other parameters. 31
6% for annual energy consumption, 19% for drying efficiency, 10% for all other parameters.
Atlete Project - Final Publishable Report | 49
6.3 Results of testing on refrigerating appliances developed in recent
years at member states level
The results achieved by market surveillance actions developed for refrigerators and
freezers in the last decade in some EU countries (and applying the EU verification
procedure before outlined) are here collected.
6.3.1 Sweden
In Sweden the Swedish Energy Agency has been running almost regularly appliance
testing over the past years. The report “Ten Years of Energy Labelling of Domestic
Appliances 1995–2005”32
states the conclusion of ten years energy labelling and showed
also the result of appliance testing from one single test:
• 101 refrigerators and freezers, 15 deviated more than allowed (14,9%).
• 19 ovens, 2 deviated more than allowed (10,5%).
• 28 dishwashers, 13 deviated more than allowed (46,4%).
• 48 washing machines, 20 deviated more than allowed (41,7%).
• 14 tumble dryers, 2 deviated more than allowed (14,3%).
Since only the Step 1 of the two-stage verification procedure of the labelling scheme has
been completed, it is not possible to draw conclusions about the actual compliance rate
of the tested products. Additional information at: www.energimyndigheten.se.
6.3.2 Denmark
On behalf of the Danish Energy Agency, the Energy Labelling Denmark checks
compliance with legislation. Energy Labelling Denmark is responsible for administration
of sample checks of products, including the selection of the products to be checked, the
retrieval and review of technical documentation from manufacturers and the follow-up
of test results with manufacturers and suppliers.
Every year tests are run on 5-20 items of each product group, equivalent to 1-5% of the
national market. The preconditions for the annual work are established by the Danish
Energy Authority in cooperation with Energy Labelling Denmark and the laboratories
that test the products. The models to be checked are selected either as a random
sample or according to set criteria that might include: discrepancies in the information
on the energy label, previous unacceptable results for the same supplier, the desire to
check a number of appliances that have features in common or the appliance type’s
market share; efforts are also made to include all suppliers and brands in the check,
though not necessarily every year. An annual report is prepared showing the results
achieved in the previous year and also the followings of the verifications run in the
previous years
32
The Swedish Energy Agency, Ten Years of Energy Labelling of Domestic Appliances 1995–2005, ER
2006:18.
50 | Atlete Project - Final Publishable Report
The Annual Report of the activities for 200733
and 200834
(covering the activities from 1st
January to 31st
December of each year) show (Table 11) that the compliance testing
action initiated in 2007 has not yet completed in 2008. For refrigerators and freezers the
volume is measured in addition to the energy consumption.
Table 11: Results of the 2007 tests on household appliances in Denmark
2007 tested products Models
number
Non-compliant
models, Step 1
Models
to Step 2
Non-compliant
models, Step 2
Household refrigerators, free-
zers and their combinations 30 12 7 not completed in 2008
refrigerators 10 1 0 --
refrigerator-freezers 10 5 4 not completed in 2008
upright freezers 6 2 0 --
chest freezers 5 4 3 not completed in 2008
Washing machines 7 4 0 --
Washer dryers 3 2 0 --
Dishwashers 10 4 4 not completed in 2008
Electric ovens 5 + 4* 3 2 not completed in 2008
Air-conditioning systems 4 1 1 not completed in 2008
*four models came from a Norwegian product compliance action
Unfortunately the follow up of the test results is not fully understandable from the
Annual Reports: in fact it is not clear if – and how many - models have been re-tested in
Step 2 and confirmed non-compliant or if the action taken by the supplier (label
declaration modification, technical modification of the product, discontinue the model
production) were the consequence of a bilateral discussion with the Market Surveillance
Authorities over the results of the tests done on the first unit or of a final assessment of
the model non-compliance after three additional units were tested.
6.3.3 UK
The overall UK market surveillance system is coordinated by BIS (department of Business
Innovation and Skills). Within this system DEFRA - the UK government department
responsible for policy and regulations on the environment, food and rural affairs - has
carried out since 2004 compliance verification tests on household appliances.
Enforcement and compliance are in fact considered essential components to deliver the
desired energy savings but also create a level playing field for industry. At present the
rate of non-compliance in the UK is estimated to be around 10 to 15% at manufacturing
33
Annual Report 2007 Report on the work of Energy Labelling Denmark on checking energy labelling of
household appliances, air- conditioning systems and household lamps in Denmark, Energy Labelling
Denmark, 2008. 34
Annual Report 2008, Report on the work of Energy Labelling Denmark on checking energy labelling of
household appliances, air conditioning systems and household lamps in Denmark, , Energy Labelling
Denmark, 2009.
Atlete Project - Final Publishable Report | 51
level (failure to meet the claim on the label) and 20% at retail level (absent or incorrect
labelling).
In 200535
a set of energy label tests were carried out on 8 ovens, 10 washer driers, 20
tumble dryers and 20 refrigerating appliances for DEFRA via the Market Transformation
Programme (MTP) in order to monitor compliance to the EU labelling directive
provisions.
The specific research for refrigerators and freezers involved the purchase of 20 domestic
refrigerating appliances from high street traders and testing them to measure whether
the products complied with the values declared for the energy consumption and the
storage volume on their energy labels displayed at the time of purchase: the scope was
limited to testing one sample of each appliance (i.e. only Step 1 of the verification
procedure was carried out). The following protocol was used in producing the list of
brands for purchasing:
• for the refrigerator-freezers, 10 models were selected from different brands that
were being promoted under a British Gas EEC scheme in a retailer during January
2005. At that time only a limited number of ‘class A’ energy rated models from a
limited number of brands were promoted through this scheme. The scheme was
subsequently extended to include all ‘class A’ energy rated refrigerator-freezers,
but the brand list was not changed as it covered a representative selection of
suppliers. The majority of the samples were bought from the same retailer, but
three were sourced from other retailers
• the remaining five refrigerators and freezers were chosen from brands not
already represented in the refrigerator-freezer list. The selection included three
retailer own brands.
Although the general summary of the verification outcome was that 13 models were
within the measurement tolerance allowed for energy consumption, the report stated
also that 14 out of 20 appliances tested did not comply with the energy class as claimed
on the energy label: 11 models were one class worse and 4 were two or more classes
worse; 10 models out of these 15 fifteen non-compliance appliances have energy
consumption and / or volumes at least 10% worse than that claimed on the energy label;
8 models were in excess of the allowed 15% tolerance on energy consumption.
Since Step 2 test (on additional 3 units) was not run, no final conclusions can be drawn
on the actual compliance rate for the 15 refrigerating appliances considered as having
failed Step 1. Re-analysing more carefully the available test results, it has been possible
to argue that (see Table 12 that provides the final results of this testing campaign):
• Model EC5 has been excluded from the statistic and models having passed the
energy consumption test and having the measured energy efficiency class one
class worse than the declared have been considered as complying with the label
class parameter (it is not clear if the energy efficiency class has been calculated
from the measured values or from the rated ones).
• It can be assumed that for the 12 models having failed the volume measurement
there is little scope to re-measure the same volume for 3 additional units in Step
35
Nicola King, Market Transformation Programme, 2005 Energy Label Compliance Testing Post-
Consultation Report, September 2005.
52 | Atlete Project - Final Publishable Report
2, therefore 11 out of these 12 models fail the verification. For the 12th model
(EC8) the outcome of the volume measurement is apparently controversial, since
the laboratory and the supplier have a different opinion about the use of the
drawers in the volume measurement.
• For the remaining 3 models (EC3, EC4, EC5), having failed Step 1 for the energy
consumption but compliant for the storage volume, Step 2 would have been
fundamental for a final decision, as well as the testing of the other two
parameters included in the labelling scheme, the freezing capacity and the
temperature raise time.
• For model EC5, probably damaged, the repetition of Step 1 should have been
done.
The outcome of this new analysis is thus that 5 models (26,3%) passed Step 1 for the
three measured parameters and 14 models (73,7%) failed.
Table 12: Analysis of the results of the 2005 energy label compliance for refrigerating appliances in UK
Sample
code
EE
class
Energy
consumption
Storage volume of the
compartment Step1 result
refrigerator freezer
EC 1 P P P F F
EC 2 F(?) P P F F
EC 3 (?) P P P P
EC 4 F F P P F
EC 5 F (?) F P P F (unit damaged)
EC 6 F (?) P P F F
EC 7 F F F/P (second comp.) F F
EC 8 P P P F (?) F (?)
EC 9 F (?) P P F F
EC 10 F (?) P P F F
FR 1 P P F F
FR 2 P P P P
FR 3 (?) P P P P
FR 4 P P P P P
FR 5 F F F F F
FZ 1 F F F F
FZ 2 F F F F
FZ 3 F F P F
FZ 4 F (?) P F F
FZ 5 (?) P P P
Pass 13 13 11 2 5
Fail 6 6 3 5 14
Pass (%)
68,4 68,4 78,6 28,6
26,3 8 or 42,1%
Fail (%) 31,6 31,6 21,4 71,4
73,7 11 or 57,9%
Atlete Project - Final Publishable Report | 53
6.4 Comparison with the EU and the ATLETE approach
The comparison of the above described verification procedures as well as the pass/fail
rates achieved for refrigerating appliances with the approach followed for the ATLETE
project is not easy because:
• the verification testing was run on a different time scale: for example some
Australian data are more than 10 years old
• an incomplete analysis is sometimes reported: in some cases only preliminary
partial data and in general only the results of Step 1 are available
• the applied measurement methods and verification tolerances are non-
homogeneous:
• it is not possible to distinguish the data derived from the verification of energy
efficiency minimum requirements and the verification of the energy labelling
efficiency classes and other performance parameters
• the results of the verifications are not disaggregated by the tested parameters
and in some cases only the energy consumption or the energy consumption and
the volume has been verified.
6.4.1 Differences and common aspect in the applied verification procedures
The major differences among the verification testing procedures applied in the USA and
Australia/New Zealand with the procedure followed in the EU and the ATLETE project
are:
• applicable standard with the specific testing conditions: the test conditions used
in the EU (and therefore in the ATLETE project), in the USA and in Australia/New
Zealand are currently quite different so that the results of the tested parameters
cannot be compared;
• use of accredited laboratories, listed in specific publicly available lists, as the only
laboratories that can develop a formal verification procedure, and the specific
accreditation criteria;
• accredited laboratories are used in Australia and since the beginning of 2011 in
USA;
• Instead in the EU Member States national Market Surveillance Authorities can
(and in some cases must) use any national laboratory they consider appropriate.
Within the ATETE project we have defined a specific procedure with specific
selection criteria to choose the most skilled test laboratories;
• criteria for the selection of the models: random sampling or selection according
to a specific set of criteria that might include: discrepancies in the information on
the energy label, previous unacceptable results for the same supplier, the desire
to check a number of appliances that have features in common or the appliance
type’s market share, inclusion of all suppliers and brands in the check, selection
54 | Atlete Project - Final Publishable Report
of products with an above average likelihood of non-compliance, other strategic
considerations, etc;
• the ATLETE project applied a semi-random selection on the best sold models, but
this is justified by the fact that this was the first pan-EU compliance verification
exercise and not previous experiences were available;
• partial verification of the labelling/ecodesign requirements: in some verification
action only the energy consumption and the storage volume only were tested,
probably considering them the most important parameters and trying to reduce
the time and costs of the verification action. In our view and in the light of the
ATLETE results, all parameters to be verified have the same importance from the
market surveillance point of view. Giving a lower importance to a parameter may
result in lower attention by the supplier and a higher non-compliance rate.
The common aspects to all verification actions are:
• in all countries/regions a two-stage verification procedure with specific
tolerances (linked to measurement uncertainty) is applied;
• this reassures us that despite the recurrent complaining of some European
stakeholders that no verification tolerances should be allowed, the EU procedure
is in line with the protocols followed worldwide in the major markets;
• a “commenting stage” is foreseen: non-compliant test results are discussed with
the relevant supplier to evaluate if a possible defective/damage unit might have
been tested, if the standard test conditions were fully respected by the
laboratory or if an incorrect declaration has been done by the supplier
• in the UK compliance verification 25% of the selected models showed that an
erroneous energy class was claimed from the declared energy and volume
values; within ATLETE we found the same incorrect although for a very limited
number of models
• in Australia, when Stage 2 verification is run all parameters are re-tested on the
additional units, and not only those failed Stage 1; this reassures us that this
approach, followed within ATLETE is internationally shared
• in USA part of the models were randomly selected, although then some
additional corrective conditions were also applied.
6.4.2 Comparison of the pass/fail rates for refrigerating appliances
Despite the described differences, the comparison of some test results achieved at
Member State and international level results with the ATLETE outcome can be
attempted:
• in the UK tests, 1 model (or 5%) out of the 20 tested was probably broken. This is
in line with the 2 models (or 2,4%) out of 82 found defective or suspected to be
damaged in the ATLETE project
• in the Australian tests, 114 models that failed Step 1 were tested in Step 2 and 26
models (or 22,8%) passed Step 2. Within the ATLETE project 4 models (or 18,2%)
Atlete Project - Final Publishable Report | 55
out of the 22 for which Step 2 was run passed the second step of the
verification.
In general. the pass/fail rate of the verification testing is significantly influenced by the
followed procedure. For example: if the models sampling is based on the selection of
products with an above average likelihood of non-compliance, the resulting fail rate will
be higher than if tests were to be conducted on a random sample of products available
in the marketplace. In the latter case in fact it would be expected that the majority of
tests would be compliant. The use of accredited laboratories would allow to reduce the
measurement uncertainty of the measured value, thus allowing to reduce the
verification tolerances.
6.5 The manufacturers’ conformity assessment in the EU Member States
A world-wide survey, covering 26 EU member states and 4 other (non-EU) countries:
Australia, Japan, Turkey and the United States, been carried out by the ATLETE project
with the purpose to concretely assess the applied methodology and the achieved results
in the conformity assessment activities. The results shed light on a multifaceted reality
with lights and shadows, even if, on average, the situation is not satisfactory.
The survey has investigated the main relevant points concerning the surveillance
activities starting from the analysis of the countries that do perform them and those that
do not carry out any type of surveillance. Apart this first obligatory screening, the survey
has then investigated the following issues:
• the laboratories accreditation
• the critical points of the verification tests
• the sanction for non compliance (and their effect)
• the way in which the results of the surveillance activities are made public.
The results of the survey are in deep described in the WP2 deliverable: "Overview of the
methodology and the results achieved in the conformity assessment action at EU
Member States and international level 36
" while a summary of the outcome of this work
are outlined in the following paragraph of this paper.
6.5.1 Countries performing or not performing the tests in practice
Countries performing the tests
All in all More than half EU Member States declare to perform tests in practice in order
to verify the correctness of energy labels’ classification. These are: Austria, Denmark,
Estonia, Finland, France, Germany, Greece, Hungary, Italy, Latvia, The Netherlands,
Spain, Sweden and the United Kingdom but with different extent. Actually the test
frequency is different among these countries and, in most cases they are rarely
conducted (i.e. one or two time during the last ten years). In other cases other countries
(or regions like the some Bundesländer in Germany ) are said to be quite active in testing
36
See: http://www.atlete.eu/index.php?option=com_docman&Itemid=111
56 | Atlete Project - Final Publishable Report
appliances. Appliances are also tested each year in Denmark, Hungary, Greece, the UK
and The Netherlands. Also the average number of appliances tested per year differs
from one country to another, with figures up to 200 appliances (Hungary) tested per
year. A significant number of tests (50-100) are conducted in Denmark, The Netherlands
and the UK.
The triggering factor for conducting a test could be either a compliant (i.e. from the
consumer associations) or a suspicion (from the enforcement authority). This is for
instance the case in France. Finland also mentioned that complaints represent less than
1% of all tests undertaken, suspicion and random selection being the main sources of
appliance testing. In Spain, whereas tests are undertaken after complaints by the
regional governments, they are only conducted upon suspicion by IDAE, on its own
criteria. In Denmark, The Netherlands and the UK, tests are not carried out after
complaints.
Finally appliances to be tested are picked up at retail shops. IDAE (Spain) indicated that
it would pick up the required appliances from a market store in order to reproduce the
casual circumstances encountered by any consumer. In Denmark, appliances are
obtained from central stocks where all types of appliances are available (except for
lamps which are bought at retail shops).
Countries not performing the tests
Twelve EU Member States (Belgium, Bulgaria, Cyprus, Czech Republic, Lithuania,
Luxembourg, Malta, Poland, Portugal, Romania, Slovakia and Slovenia) do not carry out
appliance tests of energy labels’ accuracy. The main reason for not conducting the tests
is that these procedures are considered too expensive but also the lack of adequate
laboratories and the fact that the surveillance authorities are often undersized (being
their main duty the safety controls) are other relevant barriers.
Nonetheless all the interviewed countries except for the Czech Republic and Slovakia
(focusing on other priorities than testing appliances) states that they intend to conduct
label verification tests in the near future.
Although tests are not being conducted, most of these countries indicate that they
provide the opportunity to other stakeholders to submit observations on product
compliance. Several countries underlined that consumers (usually through consumers
associations) and any other organisations could easily contact the enforcement
authorities. However, despite offered opportunities, no experience (Portugal) or few
cases (Slovenia, Belgium) of submitting observations on product compliance by other
3
3
6
The controlling institution has too many subjects to deal with
Finding an appropriate laboratory or independent company to proceed to the tests is difficult
Tests would be too expensive
number of countries
Why energy classification is not verified?
Atlete Project - Final Publishable Report | 57
organisations occurred: Belgium for instance mentioned that the Ministry received and
followed-up two to three complaints from consumers and one complaint from a Region
6.5.2 The laboratories issue
As already outlined, the presence of accredited laboratories is one of the main barriers
for the carrying out of the surveillance activities. Countries performing tests report
dispose of very few accredited laboratories: only one accredited laboratory in Estonia
and Hungary; 2 laboratories in Denmark and Greece; 3 in Austria, Finland and The
Netherlands; 4 accredited laboratories in Spain (but each one of them is not running all
kinds of tests). For what concerns the laboratories provenience and legal status they are
usually national in the EU Member States and most frequently private (but in Hungary
there is only a single testing laboratory, which is a public laboratory).
Four countries (Austria, The Netherlands, Sweden and the United Kingdom) declare
making use of both national and foreign laboratories. This also depends on which kind of
appliances are considered: the Electrical engineering section in Austria, for example,
sometimes needs to turn to foreign laboratories as Austrian laboratories are not
accredited to conduct tests on the complete range of electrical appliances.
6.5.3 The verification Tests
Several problems have to be faced when undertaking a surveillance activity. The major
difficulty relate to financial costs but also other problems as the time required to carry
out a complete testing procedure, the manufacturers objections to the tests results, the
fact that very often the tests are not directly carried out by the surveillance authority
(for lack of resources) severely limit the MSs capacity to carry out a strict a strict
monitoring action.
For what concerns the total costs, for example, only very few countries provided
information concerning the annual cost of monitoring manufacturers’ compliance37
. The
cost of a single test vary from 1,800 € to 2-3,000 € depending to the appliances tested
but the cost of a meaningful testing campaign may easily arrive to 300,000 € (i.e.
Denmark, The Netherlands, Sweden and the United Kingdom)38
.
Another limit in performing labelling verification tests is the time required for getting
complete tests’ results. The time required from the decision to launch a test to the
results varies from 2 to 6 weeks in Greece (depending on the type of appliance tested);
about 1 month in Hungary and Estonia; 1 to 2 months in Finland; from 3 to 4 months in
Denmark and the United Kingdom.
For what concerns the time required from the decision to launch a test to the results
varies from 2 to 6 weeks in Greece (depending on the type of appliance tested); about 1
month in Hungary and Estonia; 1 to 2 months in Finland; from 3 to 4 months in Denmark
and the United Kingdom. According to the ATLETE experience, a complete verification
cycle may take more than 6-8 months.
37
Source: Fraunhofer 2008 survey. 38
The cost of the testing campaing in ATLETE has been of more than 500,000 €
58 | Atlete Project - Final Publishable Report
6.5.4 Non-compliance and sanctions
Once non-compliance is confirmed, remedy actions may be a topic of negotiations
between manufacturers and enforcement authorities. Usually the first mean of action is
an order to rectify the situation, the manufacturer being asked to correct the energy
class (a timetable of one month is imposed in Hungary). The importers may also be
contacted as this is the case in The Netherlands. Past experiences in this country showed
that importers promptly took action by sending the non-compliant appliances back to
their manufacturers (cases on air-conditioning appliances for instance).
Typical sanctions are fines, product withdrawal from the market and even court
prosecution. These sanction are rarely administered. Fines are moreover considered not
high enough to increase compliance and no cases of product withdrawal have been
reported in this survey. This may depend to the fact that as high rates of compliance or
fast corrective actions are observed in the countries performing tests, few information
was provided about sanctions applied in practice39
. Some Member States mentioned
court prosecution. However, very few of them were really applied in practice. According
to the Fraunhofer 2008 survey, seven EU countries (Cyprus, Denmark, Finland, France,
Luxembourg and Sweden) reported “zero” prosecutions and sanctions issued. About 10
prosecutions, and as much sanctions, were applied in Hungary and Latvia. There is no
available data on the number of prosecutions for the other EU countries.
It is finally worth noting that another dissuasive mean to reduce non-compliance is the
publication of non-compliance cases through press releases (as happened in Sweden).
Publishing general reports was also used in the UK in the past framework. Similarly in
Spain, non-compliant appliances are withdrawn from IDAE’s national and public
database of household electrical appliances (published through a website). Regional
governments can also send the results to retailers’ associations, which is considered as a
dissuasive measure with harmful consequences on the appliance’s distribution in shops.
6.5.5 Sharing the results and information
Among the EU Member States conducting tests in practice, only 5 countries (Denmark,
Estonia, Hungary, Sweden and the United Kingdom) do make results public. Results are
not per se made public in Spain; however, they can be sent to retailers’ organisations
which may then withdraw non-compliant appliances (see above).
The main explanation provided to explain the low effort in disseminating the results is
that it was not required within the EU Directive. In Greece, for example, only non-
compliant manufacturers are informed about tests’ results.
39
Since remedy actions can be discussed at local levels (in Germany and Spain for example), there could be no
statistics at the national level about the measures undertaken in order to return to compliance.
No6
Yes5
N/A3
Are test results made public?
Atlete Project - Final Publishable Report | 59
It is also worth adding that countries performing verification tests on energy labelling
usually do not compare tests’ results with laboratories / enforcement authorities from
other countries. The comparison of tests’ results can actually reveal differences between
European countries, i.e. different results can be noticed on the same appliance tested by
two different laboratories. This difference could be a consequence of testing procedures
used by laboratories (Greece’s experience) or of slight differences among products from
the same model according to the manufacturing chain in different plants.
Limited resources (Hungary) or the lack of national manufacturers (Estonia) are
suggested to explain the non-comparison of tests’ results across countries.
6.5.6 Perspectives on Ecodesign
We deem useful close this brief overview on the way the MSs face and carry out the
surveillance activities with a brief note on implementation and monitoring of the
Ecodesign Directive.
Nine EU countries (Bulgaria, Cyprus, the Czech Republic, Denmark, Estonia, Slovakia,
Slovenia, Sweden and the United Kingdom) have declared to have already
modified/reinforced their surveillance actions in view of verifying Ecodesign
requirements. However, since it is a recent issue, some countries (Finland, France, and
The Netherlands) are currently working on the relevant procedures to be implemented
and the required changes in the national legislations.
Nonetheless, more than half EU countries have not implemented actions to reinforce
market surveillance on Ecodesign measures. In Austria, for instance (like in Belgium,
Poland, Romania...), the enforcement body has just been founded. More emphasis will
be placed on the surveillance actions related to Ecodesign once the implementing
measures start to come into force.
No14
Yes9
Ongoing
process
3
Have surveillance actions on Ecodesign been already modified?
60 | Atlete Project - Final Publishable Report
7 FINAL POLICY CONCLUSIONS AND RECOMMENDATIONS
1. The ATLETE Project has shown, beyond any doubt, that market surveillance is
essential to guarantee the compliance of the specific products with energy
labelling legislation.
2. The ATLETE Project has demonstrated that market surveillance is essential,
technically possible and cost effective. Lack of market surveillance in the area of
energy labelling and ecodesign of energy related products leads to unfair
competition enabling “free-riders” to gain potentially considerable market
advantage over the competitors, which in turn creates market distortion and
undermines the possibility to achieve the demanding EU energy efficiency
targets. Lack of market surveillance also undermines the trust of the consumers
in the EU energy efficiency policies and makes their search for high energy
efficient products somehow ineffective.
3. The verification procedure followed in the ATLETE Project is based on the
standard procedure applied in EU energy labelling and ecodesign legislation.
Although it has been validated against the old energy labelling and in the specific
case of refrigerators and freezers, it is fully applicable to the new delegated
regulations. The procedure is based on a two-step approach: in Step 1 the check
is performed on one sample of the model; in case of suspected non-compliance
Step 2 is conducted, testing three additional samples of the same model.
Depending on the parameter to be verified, a verification tolerance (uncertainty
in the laboratory measurements) is applied to both Steps.
4. The ATLETE Project has proven that this two-step approach is necessary for the
proper assessment of the product compliance. Almost 25% of the refrigerating
appliance models suspected of non-compliance in Step 1 and tested in Step 2
ended up being fully compliant with energy labelling requirements.
5. Cooperation of testing laboratories and the exchange of the experiences gained
during the product testing foreseen within the procedure developed in the
ATLETE project has proven to be helpful in achieving fully comparable test
results. The four meetings with the testing laboratories, project partners and
international experts have provided useful suggestions for the fine tuning of the
verification procedure.
6. Market Surveillance should be conducted both at country level as well as at EU
level. The exchange of experiences between the national Market Surveillance
Authorities is also needed for a better planning and coordination of the national
efforts.
7. National Market Surveillance Authorities should guarantee that testing
laboratories assure the lowest possible measurement uncertainty. This will in
turn allow comparability of the verification results.