Study for the Directorate-General for Energy (DGXVII)
of the Commission of the European Communities
Final Report – May 1999
Energy Efficiency
of Room Air-Conditioners
(EERAC)
Contract DGXVII4.1031/D/97.026
Co-ordinator: Jérôme ADNOT, ARMINES, France
PARTICIPANTS
Matthieu ORPHELIN, Cédric CARRETERO, Dominique MARCHIO
Centre d’Energétique, Ecole des Mines de Paris, France
Paul WAIDE
PW Consulting, UK
Michel CARRE
Ademe, France
Carlos LOPES
CCE, Portugal
Angel CEDIEL GALAN
IDAE, Spain
Mat SANTAMOURIS,
KLITSIKAS
University of Athens, Greece
Bill MEBANE, Milena PRESUTTO, Enzo RUSCONI
ENEA, Italy
Herbert RITTER
EVA, Austria
Sule BECIRSPAHIC
Eurovent Certification
Dominique GIRAUD, Edgard BOSSOKEN
INESTENE, France
Luigi MELI, Stefano CASANDRINI
CECED, Europe
Philippe AUFFRET
Electricité de France
With the additional participation of other experts from EdF (France), ENEL (Italy),
CECED and Eurovent Cecomaf (European manufacturers’ associations)
1
CONTENTS
EXECUTIVE SUMMARY AND RECOMMENDATIONS ......................................................4
1. INTRODUCTION ...........................................................................................................12
2. ROOM AIR-CONDITIONERS IN EUROPE: TECHNICAL DEFINITIONS .....................14
2.1. Basic definitions ....................................................................................................................................................... 14
2.2. RAC types ................................................................................................................................................................ 14
2.3. Test standards.......................................................................................................................................................... 18
2.4. RAC usage: a systems approach ............................................................................................................................ 19
2.5. Statistical databases, physical models created and information gathered.......................................................... 21
3. ROOM AIR-CONDITIONERS IN EUROPE ...................................................................23
3.1. European market for RACs in 1996 .......................................................................................... ............................ 23
3.2. EU production, imports and exports........................................................................................ .............................. 24
3.3. The stock of RACs in use in 1996 .......................................................................................... ................................. 25
3.4. Operation of the market................................................................................................... ....................................... 26
4. ENERGY EFFICIENCY ON THE EUROPEAN MARKET..............................................30
4.1. Reference lines and models ..................................................................................................................................... 30
4.2. Previous market-transformation efforts within EU Member States................................................................... 32
4.3. The Eurovent Certification programme................................................................................................................ 33
4.4. The European Commission’s efforts to raise RAC energy efficiency ................................................................. 35
5. MARKET TRANSFORMATION OUTSIDE THE EU ......................................................36
5.1. Minimum energy efficiency standards and labelling schemes in non-European countries............................... 36
5.2. Present situation in the USA.............................................................................................. ..................................... 36
5.3. Current schemes in Japan.................................................................................................. ..................................... 39
5.4. The European market situation compared with other OECD countries ............................................................ 44
6. PROJECTIONS TO YEARS 2010 AND 2020 (BAU SCENARIO).................................46
6.1 Computation of energy consumption of appliances............................................................................................... 46
2
6.2 Stock and market in 1990, 1996, 2010 and 2020 .................................................................................................... 49
6.3 Electricity consumption............................................................................................................................................ 52
6.4 – Environmental impact........................................................................................................................................... 55
7. TECHNICAL–ECONOMICAL STUDY OF OPTIONS....................................................58
7.2 Options and technical results................................................................................................................................... 59
7.3 Economic calculations for the screening of cost-effective measures..................................................................... 64
8. POLICY ACTIONS REALISABLE BY THE YEAR 2010 ...............................................70
8.2 Behavioural changes, controls, comfort conditions and thermal regulations...................................................... 75
8.3 Minimum energy efficiency standards.................................................................................................................... 76
8.5 Voluntary agreement of manufacturers, performance certification .................................................................... 84
8.6 – Actions by national bodies .................................................................................................................................... 86
8.7 – Summary of possible actions................................................................................................................................. 89
REFERENCES...................................................................................................................91
APPENDIX 1- INFORMATION ON REVERSIBLE AIR CONDITIONERS (EDF) ............93
Testing standards, terminology ................................................................................................. .................................... 94
Technical analysis............................................................................................................. .............................................. 97
Technical/economic analysis .................................................................................................... ...................................... 99
Conclusions .................................................................................................................... ............................................... 101
APPENDIX 2 - ADDITIONAL STATEMENTS OF MANUFACTURERS’ ASSOCIATIONS
.........................................................................................................................................102
3
Energy Efficiency
of Room Air-Conditioners (EERAC)
EXECUTIVE SUMMARY AND RECOMMENDATIONS
Room air-conditioners (RACs) constitute a growing electrical end-use in the European Union (EU),
yet the possibilities for improving their energy efficiency are not well known. RACs may be used in
households and can be bought or ordered directly by their occupants and as such they can be
classified as domestic appliances. However, the same appliances are also commonly used in offices,
hotels and small shops and therefore the impact of any policy measures proposed through this study
needs to be explored for both the domestic and tertiary sectors on an almost equal basis.
The European Standard EN 814 specifies the terms, definitions and methods for the rating and
performance of air- and water-cooled air-conditioners. The standard energy efficiency index is
the ‘energy efficiency ratio’ (EER):
EER = Pc (cooling) / Pe (electrical)
Where Pc is the cooling capacity of the air conditioner in watts and Pe is the electrical input also in
watts. A similar index is used for reversible units to describe their performance in the heating mode
and is conventionally called the ‘COP’ (coefficient of performance). To correspond to the European
industry’s definitions, the following products categories have been used and either included or
excluded from the study:
Included in the RAC Excluded from the RAC
Single-packaged units Spot air-conditioners
Split-packaged units Dehumidifiers
Multi-split packaged units Close-control air-conditioners
Single-duct air-conditioners Control cabinet air-conditioners
Evaporative coolers
Desiccant coolers
The study makes use of two databases containing technical information on RAC models on the
market in 1997/1998. The origin of these data were the European manufacturers’ associations
Eurovent and CECED. Data collected by Eurovent (multi-split, split, single-packaged and single-
duct RACs) and by CECED (split and single-duct RACs) were merged for most of the analyses
reported here and together represent about 80–90% of the models on the European market in
1997/98.
Market and stock
Italy is the largest RAC consumer and the largest RAC manufacturer in the EU (accounting for
about half of all European production). RAC penetration is also higher in Italy than in other EU
countries. Greece, Spain and Portugal manufacture far fewer RACs than they purchase and hence
are net RAC importers. France is a manufacturing country with a growing internal market. Germany
and Austria are experiencing market growth with increasing imports. The combined annual sales of
4
RACs in the EU was about 1 600 000 units in 1996.
1996 Room air conditioner sales in the EU (Thousand units/year)
Country Split Multi-split Single-duct Single- Total sales Total stock in
packaged use
Austria 10 000 6 040 3 000 4 760 23 800 79 000
France 99 750 29 000 38 250 11 000 177 000 1 259 100
Germany 65 000 19 500 90 000 20 000 19 4500 526 100
Greece 138 000 12 880 600 1 000 150 880 744 830
Italy 363 360 20 350 42 127 13 653 439 490 2 111 740
Spain 250 000 0 39 000 29 000 318 000 1 369 000
Portugal 35 600 7 400 900 1 900 45 800 322 820
UK 104 000 0 26 000 3 800 133 800 674 412
Others 39 000 11 700 54 000 12 000 116 700 315 660
EU 1 104 710 106 870 293 877 97 113 1 599 970 7 402 662
Market share 69% 7% 18% 6% 100% -
The total number of RACs installed in Europe is around 7 500 000 units, which is equivalent in size
to the domestic market for one of the big US or Japanese manufacturers.
Energy efficiency
The distribution of RAC EERs about the average EER per type on the European market shows that
there are makred differences in energy efficiency performance and a significant margin for
overall improvement
40
35
30
25
models (%)
20
15
10
5
0
91-100
101-110
111-120
121-130
131-140
141-150
151-160
161-170
60
61-70
71-80
81-90
EER/EERaver (%)
Note that the present declared measurement tolerances are of the order of 6% in terms of the EER,
that variable speed units are not accomodated by the test procedure and that some uncertainties
about practical test conditions remain for single-duct RACs. Following one of the studies
recommendations, the Commission has prepared mandate M/274 to CEN, in order to improve the
accuracy and applicability of EN 814.
There have been some national RAC efficiency promotional programmes in the EU, such as EdF’s
5
Promotelec programme in France, while in Portugal thermal regulations in construction have been
developed that specifically address air-conditioning. The European manufacturers’ association
Eurovent has created a RAC performance certification programme called ‘Eurovent Certification’.
A high percentage of the EU market is already included in the Eurovent scheme (some 99% of
the EERAC combined database). The 10 most important European manufacturers are members of
the scheme. However, thus far, single-duct units, a growing segment of the EU RAC market, have
not been included in the Eurovent Certification programme.
International situation
A number of RAC market-transformation initiatives have been implemented in regions outside the
EU. The USA and some countries in Southeast Asia have had RAC energy labelling for many
years. In Japan, quasi-mandatory voluntary minimum energy efficiency thresholds that exclude
the sale of low-EER models are universally obeyed and are regularly reviewed and updated. The
two policy options of minimum energy efficiency standards (MEES) and energy labelling are
jointly applied in many countries, including several developing countries, although not all. An
important review of EER threshold values, energy labelling schemes and associated market-
transformation impacts has been conducted for the current study, providing a sound basis for the
determination of suitable values for future European legislation.
The principal conclusions of this comparative work are that there has been considerable legislative
activity to improve air-conditioner efficiency around the world and that a large proportion of RACs
currently available for sale in the EU would not satisfy efficiency minimum efficiency thresholds in
the countries that have introduced such requirements – including those in many developing
countries. This suggests that there is significant scope to improve RAC energy performance in the
EU and that to so does not require technological innovation, but merely implementation of well-
established higher-efficiency design options. European or national efforts reported previously in this
report appear very limited when compared to the schemes implemented in any of the countries or
group of countries considered here.
Projections of energy consumption
An evaluation of air conditioner demand and usage patterns has enabled estimates of an equivalent
number of hours of full-load air conditioner operation per year to be estimated by user sector and
country. These were used to predict RAC energy consumption and to evaluate the cost effectiveness
of the technical variations under investigation.
Estimated weighted-annual average hours of RAC usage by user sector in the EU
A base case, Business as Usual, scenario has been defined in order to analyse the technical and
economic potential of single or combined policy measures in several alternative energy efficiency
scenarios. The year 2010 has been chosen for projection purposes in accordance with the deadline
for the implementation of the Kyoto Protocol. Longer term porjections are given up to the year
2020. The base year for the study has 1996 but backwards projections have been made to 1990 to be
consistent with the base year used under the terms of the Kyoto Protocol. Very conservative
options have been taken about future market growth, nonetheless the estimated growth for
some countries is still very rapid.
6
Estimated RAC electricity consumption in the EU (GWh/year)
1990 1996 2010 2020
Austria 68.6 121.3 235.0 364.5
France 331.6 1 782.1 5 517.2 8 975.5
Germany 155.9 672.4 1 914.0 3 197.3
Greece 208.8 1 006.6 2 281.3 3 478.6
Italy 761.0 4 494.1 5 743.6 7 033.9
Portugal 162.4 713.8 1 806.8 2 552.2
Spain Not available 2 496.4 9 366.4 15 146.6
UK 120.0 446.0 1 135.7 1 783.8
Other EU 119.6 443.5 1 159.1 1 897.7
Total EU 1 927.9 12 176.2 29 159.1 44 430.2
Without policy intervention RAC associated CO2 emissions in the EU is conservatively projected to
increase by a factor of 11 from 1990 to 2010 (the timescales used in the Kyoto protocol) rising from
1 million tonnes to 11 million. The projected increment in RAC related CO2 emissions is about 10
M-tonnes (3000 M-tonnes being the initial CO2 emissions from all fuels in the EU). The direct
effect of the growth of the RAC market on total EU emissions is therefore projected to be +0.33%
without policy intervention, as compared with the -8% target for the EU as a whole.
Energy efficiency options
Among the seven RAC categories considered in this study, four specific models were selected close
to the centre of some representative ‘clusters’ for technical analysis: three very different models in
the largest category (splits = 69% of market) and one in a fast growing category (single ducts = 6%
of the stock but 18% of the market).
Two types of questionnaires, one for air-cooled and one for water-cooled systems, were compiled in
order to collect the technical data and parameters needed to model the air conditioners under
examination. The questionnaires were submitted to the European room air conditioner
manufacturers’ associations, which contacted the manufacturers of the selected representative units
and asked them to provide the technical data to the study group. A room air conditioner energy-
engineering simulation package was calibrated using the detailed technical data for each of the four
RAC models and then used to conduct simulations of a variety of known design modifications
leading to enhanced energy efficiency.
The full combination of options, using only proven design options, leads to a potential average
improvement in performance of ~50%, part of which wouldn't be economic. The economic analysis
was performed estimating the net benefits of the technological options of the various models to
consumers. This needs a careful study of costs and overcosts and the use of criteria like Net Present
Value of investment or minimum of Life Cycle cost. In all cases, such as in the graph below, the
least life cycle cost is found to occur for a 25% performance improvement, corresponding to options
(5+6+2b).
7
BASE +25% +36% +48%
Policy options
The USA-LBL model for the appliance industry was used to evaluate the impact of this technical
scenario (+25% efficiency) on European manufacturers. With the rather low elasticity and the
substantial benefits to consumers the unit shipments are predicted to increase, as much as 15% in
the combined case (5+6+2b options). This is the design option with 25% energy savings. The
policy option considered in the LBL model is quite strong (MEES at present average + 25%)
and, despite the important advantages that would accrue, scenarios and policy options have
been designed to give time and flexibility to the adjustment procedure.
It is recommended that municipalities, central administrations for southern EU Member States, etc.
should introduce regulations or at least advisory schemes on RAC sizing and inside set temperatures
and should strongly promote the development of regionally specific energy consumption
calculations. All the bases for the development of a common rule to easily compute RAC energy
consumption under local circumstances have been gathered in the present study. If an "installed
appliance" scheme is proposed, all the required technical elements can be found in the present report
A significant energy saving potential exists from the development and improvement of RAC
maintenance schemes. Terms of reference for such schemes could be defined with the national
installers (and retailers) associations. How much of the 20% drop in RAC efficiency due to fouling
could be avoided by any given maintenance scheme and for a given cost is an interesting subject for
national evaluation, prior to or independently of any common European measure.
At EU level, the introduction of Minimum Energy Efficiency Standards is recommended with
two sets of thresholds, one for enforcement before the year 2003, and another before 2010. The
recommended efficiency level of the second MEES is the present European market average EER +
10%, which is still short (by 15%) of the cost effective optimum. Its impacts, obtained by a detailed
stock simulation are given in the final table. The threshold values to be applied, in case they are
defined by types, are given hereunder.
Recommended minimum energy efficiency standards (MEES) for RACs in the EU (all values are for EER
levels (W/W))
Type of RAC First MEES Current highest EER Second MEES
(present market on the EU market (present market
average) average +10%)
MS,A: Multi-split, Air-cooled 2.63 3.74 2.89
8
PA,A: Single-packaged, Air-cooled 2.38 2.97 2.62
PA,W: Single-packaged, Water-cooled 3.32 5.42 3.55
SD,A: Single-duct Air-cooled 1.80 3.09 2.28
SD,W: Single-duct, Water-cooled 2.36 3.62 2.60
SP,A: Split, Air-cooled 2.48 3.56 2.73
SP,W: Split, Water-cooled 2.75 2.88 3.03
A comparison has been made with the MEES and target thresholds applicable in the largest RAC
producing countries. The proposed first and second MEES are just in the middle of the lines applied
by competitors up to 1998, but fall considerably short of the new efficiency targets in countries such
as Japan. The proposed MEES are clearly needed to avoid the outcome wherein the energy
efficiency legislation enforced in a number of exporting South East Asian countries results in
inefficient products that can no longer be sold locally being exported at discounted prices to the EU.
For energy labelling, four separate RAC categories could also be used for single-packaged units,
splits, multi splits and single ducts. The EER measured according to EN 814 (at the T1 conditions)
is generally representative of relative in situ energy consumption performance. Since the ‘unit’ EER
value is not representative of the energy consumption for water-cooled systems, specific limits
should be applied to their apparent EER. This leads to 7 RAC categories. The positive impact of
variable speed technologies, such as 'inverters', or simply multispeed drives, that can be applied to
any class of equipment, merits a positive correction for the nominal EER. Appraisal of this bonus
should be delayed; however, until the revision of EN814 provides more objective evidence on the
subject.
Potential energy labelling structure and thresholds: Air-cooled RACs
Label class limits EER/EERav Split Multi Split Packaged Single Duct
A starts over 150 3.72 3.95 3.57 3.11
B starts over 140 - 150 3.47 3.68 3.33 2.90
C starts over 130 - 140 3.22 3.42 3.09 2.69
D starts over 120 - 130 2.98 3.16 2.86 2.48
E starts over 110 - 120 2.73 2.89 2.62 2.28
F starts over 100 -110 2.48 2.63 2.38 2.07
G starts below 100 2.48 2.63 2.38 2.07
Potential energy labelling structure and thresholds: Water-cooled RACs
Label class limits EER/EERav Split Packaged Single Duct
A starts over 150 4.13 4.98 3.54
B starts over 140 - 150 3.85 4.65 3.30
C starts over 130 - 140 3.58 4.32 3.07
D starts over 120 - 130 3.30 3.98 2.83
E starts over 110 - 120 3.03 3.65 2.60
F starts over 100 -110 2.75 3.32 2.36
G starts below 100 2.75 3.32 2.36
9
The models found on the market compare appropriately with this classification structure: no specific
manufacturer has problems and no product category is specifically prohibited from the higher
labelling classes. A model of the label has been prepared and its effect forecast. Note that a labelling
structure has also been produced wherein all the RAC categories are merged into one category. This
could be deemed to correspond more closely to the interests of the consumer (an A appliance having
always the same performance level in absolute terms, whatever its porduct category) but it is not
favoured by some manufacturers.
T e c h n ic a l g r o u p :
S p lit, M u lti-s p lit, W in d o w ,
S in g le -d u c t
S p lit
A /C
E n e r g y c o n s u m p tio n :
A v e ra g e n b o f h o u rs * (1 0 0 W /m ²) /
EER
In d ic a tiv e e n e rg y c o n s u m p tio n
k W h /(y e a r.m ²)
A c tu a l e n e rg y co n s u m p tio n w ill d e p e n d o n
h o w th e a p p lia n c e is u s e d a n d o n c lim a te C o o lin g c a p a c ity :
a t T 1 c o n d itio n s
C o o lin g c a p a c ity [k W ]
C o o lin g o n ly / R e v e rs ib le C o o lin g o n ly
A ir c o o le d / W a te r c o o le d A ir c o o le d
H e a tin g m o d e (o r n o t)
C o o lin g m e d ia
A n y c e rty fy in g lo g o
E u ro v e n t o r o th e r
The impact of the labelling scheme may increase after 2003, but labelling can also help the market
to adjust to the proposed MEES regimes. Nontheless, the impacts of the first MEES and of energy
labelling combined are unlikely to be enough to make the market reach the economic optimum.
Stronger actions, such as the second MEES and the involvement of manufacturers, national agencies
and utilities will be necessary to get closer to the optimum. These actions could bring the market
average EER in to the neighbourhood of the best models available today by 2020, 2015 or
2010 depending on their vigour.
If voluntary agreements with manufacturers are proposed to obtain higher average EER levels, a
certification scheme similar to the one existing for splits, multi splits and single-packaged units
should be introduced by the manufacturers themselves for single duct RACs. Targets for all
segments could be defined over the second MEES level, half way toward the economic optimum.
In terms of average performance, if the intention is to achieve the economic optimum (+25%) by the
year 2015 through a voluntary agreement, the WEER (Weighted EER) should change in the
following way.
Weighted EER corresponding to achieving the techno-economic optimum target by 2015 where the market
shares of the various RAC types used for overall averaging are their shares in 1996.
Year 1996 Check point Check point Check point Target 2015
2010
2000 2005
Splits (air) 2.48 2.79 3.10 3.10 3.10
Splits (water) 2.75 3.10 3.44 3.44 3.44
10
Multi-splits 2.63 2.96 3.29 3.29 3.29
Packaged (water) 3.32 3.74 4.15 4.15 4.15
Packaged (air) 2.38 2.68 2.98 2.98 2.98
Single ducts (air) 2.07 2.33 2.59 2.59 2.59
Single ducts (water) 2.36 2.66 2.95 2.95 2.95
All RACs 2.44 2.75 3.06 3.06 3.06
One can judge the depth of a voluntary agreement by comparing the WEER reached with the WEER
of the table above. Voluntary agreements may cover also some additional aspects such as advice on
sizing, agreement to only offer high performance equipment to installers, indications of acoustic
power on labels, etc…
All electric utilities and agencies can have an important role. Northern utilities can promote high
efficiency reversible units; see annex 1. Southern utilities could support the European scheme
because of the savings it will induce in their investments
Aside from the +25% efficiency goal, what defines the target is the date at which this optimum is
reached. The energy consumption and environmental impacts associated with various policy targets
have been estimated using the market stock model for the year 2010, the deadline for the
satisfaction of the Kyoto protocol targets and for the year 2020 and are summarised in Table 8.15.
The average RAC life expectancy (10-13 years) and hence replacement rate is such that reaching the
target in 2010 or 2015 requires almost immediate action, while targeting the year 2020 to reach
today’s economic optimum would allows progressive measures.
Summary of targets and scenarios
Scenario Policy measures TWh TWh CO2 CO2 Annual Annual Avoided
saved saved saved in saved in gain in consum- risk in
(labelling is an per year per year 2010 - 2020 Manuf. er 2020
accompanying in 2010 in 2020 % of M-tonnes Revenue benefit (southern
measure in all cases) 1
(M-Euro (M-Euro utilities)
1990
in 2010) in 2010)
BAU No measure - - - - - - -
"First Step" 1st level MEES in 0.6 1.6 60% 0.7 12 72 88
2003 +labelling
"Target MEES in 2003 and 2.8 8.6 120% 3.4 57 336 411
2020" 2007 +labelling
"Target MEES in 2000 and 3.3 9.4 100% 3.7 67 396 485
2015" 2005 +labelling
"Target 2nd level MEES in 4.4 10.3 140% 4 90 528 646
2010" 2000 +labelling
High energy savings are possible, as well as significant CO2 emissions reductions, at no (in fact
negative) cost, such that all parties (manufacturers, consumers and utilities) would find a benefit in
the marketing of efficient RACs.
1
Around 1 Mt CO2 emitted in 1990 due to RAC on a total of fuel related emissions of 3000 MtCO2; the percentage is
expressed in the terms of the Kyoto protocol, i.e. related to the total 1990 emissions (like the 8% target), but only the
RAC emissions. In terms of the total emission the values are between 0,02 and 0,07%.
11
1. INTRODUCTION
SAVE II is an EU programme to promote the rational use of energy within the EU. The EERAC
working group began a study in December 1997 to investigate the technical and economic potential
of measures to raise the energy efficiency of individual (or ‘room’) air-conditioners (RACs). This
report gives the final result of work to May 1999.
The objectives of the study were:
• to estimate the electricity consumption of RACs
• to estimate potential energy savings deriving from the use of more efficient RACs
• to investigate ways in which these savings can be realised, including behavioural changes and
policy options (labelling, minimum efficiency standards, procurements, incentives, etc.)
• to make appropriate recommendations, on the basis of a cost–benefit analysis.
RACs can be used in households and are usually bought or ordered directly by the user. However,
the same appliances are also frequently used for office and tertiary-sector building air-conditioning.
As a result, the consequences of any measure need to be studied for both sectors on an equal basis.
The working party has been chaired directly by Paolo Bertoldi from DGXVII and co-ordinated by
Jérôme Adnot of Armines.
Choice of technical experts
• Armines is a research association supported by the Ecole des Mines de Paris and is especially
active in the field of energy efficiency, with activities ranging from technological development
to socio-economic investigations.
• PW Consulting is a UK consultancy specialising in appliance energy-efficiency initiatives and
programmes around the world.
• INESTENE is the leading consultancy on demand-side management (DSM) in France.
• Eurovent Certification is a creation of the Eurovent/Cecomaf manufacturers’ association for the
certification of the energetic and acoustic performance of air-conditioning and ventilation
equipment.
• The University of Athens, in particular the Group of Building Environmental Studies, is very
active in the field of solar cooling and energy conservation in buildings; the group carries out
research, specialised studies, application projects, education, and dissemination of information.
Represention of Mediterranean countries
National representivity was increased with the participation of the following national agencies.
• Ademe, the French energy and environment agency.
• CCE, the Portuguese energy-conservation agency.
• ENEA, the Italian national agency for energy and the environment.
• EVA, the Austrian energy research and policy institution in which the federal and provincial
administrations (‘Bund’ and ‘Länder’, respectively) and some 30 important institutions and
corporations from a variety of economic sectors co-operate.
• IDAE, the Spanish energy-conservation agency.
12
Direct representation of utilities
Utilities bring an important contribution and stimulation to the study of this growing end-use of
electricity. Two were directly present:
• EdF (Electricité de France), the French electricity utility, represented by Hervé Rivoalean and
Philippe Auffret,
• ENEL, the Italian electricity utility, represented by Salvatore Russo and Alfredo Previ.
Close involvement of manufacturers’ associations
Two manufacturers’ associations are participating directly in the study. Not only have they
participated actively in the plenary meetings of the working party, but they have also held a number
of associated working meetings with their members (the manufacturers), including those at Roissy
(February 1998), Milan (October 1997) and Rome (May 1998).
CECED is the European Council of Household Appliances, represented by its General Secretary,
Luigi Meli, its Technical Secretary for energy performance, Stefano Casandrini, the Chairman of its
ad hoc working group, Sergio Zanolin, and the Manager of its Italian branch, Antonio Guerrini.
Eurovent/Cecomaf is the manufacturers’ association for refrigeration, air-conditioning and
ventilation equipment, represented by its Director of Operations, Sule Becirspahic, the Chairman of
its group on RACs, Jan Cluyse, and one of its members, Ing Tavazzi.
Co-ordination was sought with the team that made previous efforts to prepare the application of the
‘labelling’ Directive, namely that led by TNO of the Netherlands (TNO 19982). However, the
objectives of the current study differ from those of that previous study.
2
Common final report to contract nos. XVII/4.1031/Z/95-055 ‘Energy labelling of room air conditioners’ and
XVII/4.1031/Z/96-024 ‘Energy labelling of domestic air to air heat pumps’, under the EC SAVE programme, also
financially supported by: NOVEM, the Dutch organisation for energy and environment, and NUTEK, the Swedish
national board for technical development.
13
2. ROOM AIR-CONDITIONERS IN EUROPE: TECHNICAL DEFINITIONS
2.1. Basic definitions
The term ‘room air conditioner’ (RAC) is widely used in the USA and other English-speaking
countries and helps to distinguish these appliances from central (or ‘ducted’) air-conditioners. It is
perhaps more precise to call RACs ‘individual’ or ‘autonomous’ air-conditioners, but the usual term
‘room air-conditioner’ or ‘RAC’ is used in this study to avoid confusion. A ‘room air-conditioner’,
as opposed to an ‘air-conditioning system’, is an appliance that can be bought by a household, with
a direct link between the customer and the selection of the purchased good – either direct purchase
by the household or through an installer with whom negotiation and specification of the appliance
takes place.
On the other hand, ‘central air-conditioning systems’ have a central refrigerating unit and make use
of a fluid to transport cold. They are specified by engineers or technicians, who define the system
assembly without the direct influence ofthe customer. They will not be considered in this study.
As illustrated in Figure 2.1, an RAC cools the room by ejecting heat outside of the room, either
condensing some moisture or not.
Figure 2.1. Essential quantities in the process of room air-conditioning.
P r
T i
P c
P e
The terminology in this study corresponds to that used in the European test standard EN 814-1. The
appliance extracts heat from inside the room (Pc), leading to a lower temperature (Ti) through the
use of electricity (Pe). Usually the heat ejected outside (Pr) has an energetic value similar to Pe +
Pc. The accepted energy performance index is called the ‘energy efficiency ratio’ (EER) and is
defined as:
EER = Pc / Pe
A similar index, the coefficient of performance (COP), is applied to indicate the performance of
reversible units in the heating mode.
2.2. RAC types
Air-cooled and water-cooled units must be included in the study. Evaporatively cooled units (which
are not frequently found within the study’s upper 12 kW cooling-capacity range) need not be treated
separately in the study. Some single-duct units use a very limited volume of water for heat
exchange. Note that these units are those that use water evaporation as the condenser heat-transfer
medium, not evaporative cooling units (which evaporate water directly or indirectly in the process
air). The few evaporatively cooled units are similar to air cooled units and will be assimilated with
them. On the other hand, the classic water-cooled units are a separate class: they use potable
(drinkable) water to eject the heat, a fact that should not be forgotten in assessing their
environmental impact. The water used in an RAC could in principle be non-potable water, but this
14
is seldom available.
For most of the available technologies, two types of operating mode are possible: cooling only and
reverse cycle. The techno-economic assessment of reverse-cycle units in this study focuses on the
cooling mode as it is the air-conditioning function that is the area of interest. None the less, some
data on the heating mode will be presented in an appendix.
Some systems have a variable refrigerant volume (by way of varying the compressor speed), which
results in better part-load performance. They are called ‘inverters’ and will be studied as part of
technical options.
As this study addresses air-conditioners that will be potentially used in the domestic sector, and as
the German market has many RACs using a 400 V, 3-phase supply but with a cooling capacity of
5 kW, RACs with any input voltage will be included in the terms of the study.
Split-packaged units (split systems)
A split-packaged unit is defined as a factory assembly of components of a refrigeration system fixed
on two or more mountings to form a matched functional unit. This type of appliance comprises two
packages (an indoor and outdoor unit) connected only by the pipe thattransfers the refrigerant. The
indoor unit includes the evaporator and a fan, while the outdoorunit has a compressor and a
condenser.
Indoor units can be either fixed – whether mounted high on a wall (Figure 2.3), floor-mounted or as
‘cassette’, ceiling-suspended, built-in horizontal or built-in vertical mountings – or, sometimes,
mobile (Figure 2.4). The outdoor unit can be either fixed (see Figures 2.3 and 2.4) or mobile (Figure
2.5). Whether the indoor or outdoor unit (or both) is mobile has no discernible influence on the
system’s cooling capacity or energy performance. Ducted systems are a subclass among the split
systems (Figure 2.6). These air-conditioners can deliver cool air to several rooms or to several spots
within a single room; the smallest of these appliances currently have 7 kW of cooling capacity.
Ducted systems are uncommon in Europe but are included in our study.
Multi-split-packaged units (multi-splits)
Multi-split-packaged units comprise several interior units connected to one exterior unit (Figure
2.7). Multi-split systems will be investigated in the study, but there is a small problem in that many
configurations are possible (in terms of the number and type of indoor units that can be connected to
the outdoor unit). We shall consider here the maximum configuration, but the results can be
extended as soon as EN 814 is extended.
Single-packaged units (window air-conditioners)
Single-packaged units, commonly known as ‘window’ or ‘wall’ RACs, are strictly defined as a
factory assembly of components of a refrigeration system fixed on a common mounting to form a
single unit.
This type of equipment comprises a single package, one side of which is in contact with the outside
air for condensation, while the other side provides direct cooling to the inside air with a fan (Figure
2.8). The two sides of the appliance are separated by a dividing wall, which is insulated to reduce
heat transfer between the two sides.
This kind of unit often fits under or above a window or above a door. A distinction is generally
made between those units having louvered sides (designed to be installed in a window opening) and
those without louvered sides (designed to be installed in an opening in the exterior wall).
In general, window systems on the EU market are significantly less efficient than split systems;
15
however, the ‘window’ market is small in Europe. According to some, they are less efficient partly
because of their size but mostly because this is an old technology which has not been updated.
Single-duct units
These are appliances in which the condenser ejects hot air through a duct to the exterior (Figure
2.9). They are generally movable, but in order to operate they must be set close to a window or a
door through which the duct eliminates the hot air. In principle, a dedicated hole should be made in
the envelope just for the appliance, as the use of doors and windows for the duct allows hot air to
infiltrate. There are difficulties in taking the thermal effect into account when measuring single-duct
energy performance. Furthermore, such penetration in the building envelope has an acoustic impact.
Other appliance types
Spot air-conditioners are comparable to single-duct RACs, except that the heat is directly ejected
from the back of the unit into the ambient air (if the window is closed, this appliance actually heats
the room air while cooling the targeted ‘spot’). It is not a real air-conditioner.
Dehumidifiers are not included within the remit of EN 814. They do not permit temperature control
and do not seem to have large market share, except in the UK. As a result they cannot be considered
as RACs and will be excluded from the current study. A further study would be required to
investigate these appliances.
Neither evaporative nor desiccant cooling systems are directly studied in the EERAC project, but
they may be useful for comparison in Chapter 7. Evaporative cooling is an air-conditioning process
in which the evaporation of water is used to decrease the dry-bulb temperature of the air. Some
small appliances of this type are manufactured for the domestic sector (particularly in Greece). In a
desiccant cooling system, the air is dried before being re-cooled by evaporative cooling (or a cooling
coil). This configuration is only used in industrial applications.
Evaporative and desiccant coolers are not really in direct competition with RACs. The market share
of evaporative cooling is small within the EU, and desiccant cooling is even less common.
Summary of appliances considered as room air-conditioners
Included in the RAC Excluded from the RAC
♦ Single-packaged units ♦ Spot air-conditioners
♦ Split-packaged units ♦ Dehumidifiers
♦ Multi-split-packaged units ♦ Close-control air-conditioners
♦ Single-duct air-conditioners ♦ Control cabinet air-conditioners
♦ Evaporative coolers
♦ Desiccant coolers
Only appliances of less than 12 kW cooling capacity are included in this study.
16
Figure 2.3. Diagrammatic representation of a non- Figure 2.4. Diagrammatic representation of a non-
ducted fixed split-packaged unit. ducted split-packaged unit with mobile indoor unit.
Figure 2.5. Diagrammatic representation of a non- Figure 2.6. Diagrammatic representation of a ducted
ducted split-packaged unit with mobile indoor and split-packaged unit.
outdoor units.
Figure 2.7. Diagrammatic representation of a multi-
split-packaged unit. Figure 2.8. Diagrammatic representation of a single-
packaged unit.
Figure 2.9. Diagrammatic representation of a single-
duct air-conditioner.
17
2.3. Test standards
Overview of international and national RAC test standards
European standards (EN) are called ‘CEN standards’ within the sense of the ISO/IEC definition of
‘regional standards’ (Figure 2.2). Harmonised EU product testing standards are established by the
European agencies CEN and CENELEC, and Member States are obliged to implement them
through their adoption as national standards.
Figure 2.2. Links between RAC test standards.
S ta n d a rd s o n T e s tin g a n d R a tin g
W O RLD STAN D A RD S
I S O 5 1 5 1 (1 9 9 4 ) IS O 1 3 2 5 3 (1 9 9 5 )
n o n -d u c te d A C a n d H P d u c te d A C a n d H P
R E G IO N A L S T A N D A R D S
E u ro p ean Japan U SA
E N 8 1 4 - 1 ,2 ,3 (1 9 9 7 ) J IS C 9 6 1 2 (1 9 9 4 )
A C a n d H P c o o lin g m o d e A N S I /A S H R A E
RAC
1 6 -1 9 8 3 (1 9 8 3 )
R A C & Packaged
T e rm in a l R A C
E N 2 5 5 - 1 ,2 ,3 ,4 ( 1 9 9 7 )
A C a n d H P h e a tin g m o d e
A C = a ir c o n d itio n in g , H P = h e a t p u m p , R A C = ro o m a ir c o n d itio n e r
The full list of standards on performance, security and sound levels can be found in the references.
For test procedures on acoustic (noise) testing, refer to Directive 86/594/EEC of 01/12/86 (OJ
L344,6.12.1986, P24). Some industrial testing conventions are found in Eurovent documents 6/6
(Thermal Testing), 8/1 and 8/4 (Acoustic Testing), but these are not test standards in their own right.
Description of standard EN 814-1, 2, 3 (1997)
Within this standard:
• there is no restriction on voltage (i.e. 110, 220 and 340 V units are considered)
• there is no restriction on maximum cooling capacity
• ducted and non-ducted units are considered
• multi-split units are excluded
• continuously variable capacity control units are excluded.
The last restriction means that EN 814 excludes ‘inverter’-type RACs because it is difficult to
design comparative tests for inverter units: the energy performance of the inverters may be worse
than that of a ‘classic’ RAC at full load but can be appreciably better at part load. The existing
standards only define the tests at full load (maximum total cooling capacity at standard conditions).
Single-duct appliances are covered by EN 814 but occurrences of their testing are still very limited.
The exclusion of multi-splits is only because of the necessity of defining testing when only a subset
of the indoor units are connected, though in the meantime they can be tested with all the indoor
units connected. Work on these questions has been initiated in CEN following a mandate from the
Commission.
18
Basic principles of the EN 814 performance test
The EN 814 test standard defines many possible testing points, but normally only results recorded
under T1 and, in some cases, T2 test conditions are available (Table 2.1). Results may be
appreciably different from one test condition to another as the TNO report has shown.
The EN 814 standard specifies that two test methods can be used: measurements can be performed
either using a calorimeter room or through an air-enthalpy method. The cooling capacity must be
determined within a maximum uncertainty of ±5% independently of the individual causes of
measurement uncertainty. The electrical power must be measured within a maximum uncertainty of
±1%. Thus, overall there is a permitted maximum uncertainty of ±6% in the measured EER.
A rating plate is fixed on each unit by its manufacturer to comply with the EN 814 standard. The
rating plate contains at least the following information (in addition to compulsory information
required by safety standards):
• the manufacturer or supplier
• the manufacturer’s model designation and serial number
• the EER to two significant figures and the test condition (normally T1)
• the cooling capacity in kW, to one decimal place but not more than three significant figures.
Table 2.1. Test conditions under EN 814.
Rating Type of unit
Conditions Comfort air- Spot air- Single-duct air- Control cabinet Close-control
conditioner or conditioner a air-conditioner air-conditioner
conditioner
heat pump
Air-cooled or air/air units
Mandatory test points
b A35(24)/ A35(24)/ A27(19)/ A35(24)/ A35(24)/
T1
A27(19) A35(24) A27(19) A35(24) A24(17)
b A27(19)/ – – A50(30)/ –
T2
A21(15) A35(24)
Water-cooled or water/air units
Mandatory test points
T1 W30/A27(19) – – W15/A35(24) W30/A24(17)
Abbreviations: A = air; W = water.
Notes
a
Single-duct discharges to outside air leaving the condenser: in order to maintain atmospheric pressure in test room
during the test, outside air at 35 °C (24 °C) will be introduced and measured by pressure-equalising device of
calorimeter.
b
The wet-bulb temperature condition on the condenser is not required when testing units that do not evaporate
condensate.
2.4. RAC usage: a systems approach
Figure 2.10 describes the commercial, physical and legal operative framework of RACs. This kind
of appliance is directly related to human comfort, or more precisely the absence of discomfort, and
is influenced by local norms and building thermal standards. Actions on RACs interfere with other
19
systems, such as heating in the case of reversible units, and with the environment.
Figure 2.10. Factors inflencing RAC use
Building Codes climate
local emissions
loads Interior conditions + heat
emissions
control
Utility bill RAC
European
Policy
demand
National Agencies
merchant
manufacturer
influence, support installer
determination of national policies
The conditions of comfort that are realised depend on the attributes of the RAC, as well as on the
subjective control of the user and on the building fabric. The price paid for electricity in the summer
is certainly a factor influencing the appliance’s use. Climatic conditions differ greatly from one EU
region to another, as do indoor comfort expectations. Some of the need for cooling arises from
buildings’ solar gains and the heat transmitted through insulated walls. The amount of insulation
used to minimise the need for heating also varies greatly between countries; and existing building
codes generate large differences in cooling requirements. The efforts of architects towards a
bioclimatic architecture, taking advantage of natural heat in winter and natural cool in summer, can
be very effective in some climates.
The so-called ‘single-duct’ appliances raise an important question from a systems perspective. They
play two roles: while cooling the air, they bring into the conditioned space air either from the
outside or from another room.
In addition to energetic performance information, RAC customers are also likely to be interested in
comfort performance aspects such as acoustic comfort (noise levels) and dehumidification. It should
also be an option to indicate acoustic intensity on any RAC energy label because a low noise level is
a highly desirable quality and is also positively correlated with RAC efficiency. None the less, a
detailed study of acoustic comfort is obviously outside the scope of this study.
There is a range of ‘cooling needs’ in the residential sector, depending on users’ subjective feelings,
values and decisions, that is much wider than the range of heating needs. Definitions of standard
comfort conditions for conditioned spaces are based on state-of-the-art knowledge; however, the
levels of discomfort that lead to the purchase of an air-conditioner have not been modelled. The
quality of air filtration provided by an RAC has been neglected during the modelling of their
demand, as have RAC noise impacts and the avoidance of exposure to outside noise as a result of
being able to keep windows closed (except in the case of single-duct RACs, which may give rise to
both sources of noise in some cases).
20
Interaction with some other energy policies
It can also be noted that public policies already have a large influence on the use of air-conditioning.
Building codes relating to space heating have already been mentioned, but the increasing need for
air-conditioning has been a concern for many national agencies responsible for saving electricity.
Electricity utilities themselves may or may not welcome such increases in demand, depending on
whether they have summer or winter maximum power peaks and on their fuel mix.
Quality labels and promotional campaigns by manufacturers’ associations have already proved
effective. Specific sections of building codes in some countries, e.g. France and Portugal, deal with
air-conditioning. Although they are frequently installed by professional fitters, RACs are
increasingly more mobile and can be installed by the end-user following their direct purchase. As a
result, the inclusion of RACs in the list of appliances under study at DGXVII was consistent with all
these efforts.
Local heat emissions of all kinds in Mediterranean cities in summer have already given rise to a
heat-island effect, as in the southern USA, and the phenomenon has been measured in Athens. RAC
performance can deteriorate greatly when local temperatures are higher than the normal regional
climatic conditions, but this factor could not be taken into account in this study.
An important issue is whether or not the heating-mode energy performance of reversible units
should be taken into account in the overall evaluation of RAC energy performance. The study
authors are in agreement with the following statement by TNO (1998): ‘In practice, reversible
appliances are distinctly marketed either as reversible air-conditioners or as reversible heat pumps.’
As a result the study group believes it is more appropriate for the heating mode of reversible units to
be treated as an entirely separate issue from their cooling-mode performance. Air-conditioner
performance should be judged in the same manner regardless of whether the same appliance can
also operate as a heat pump.
Future studies will evaluate the gains attributable to reversibility. Information available in the public
domain has been summarised in an appendix to this report.
2.5. Statistical databases, physical models created and information gathered
The conclusions of this study are based upon the information gathered from two technical databases,
a number of computer models and national surveys performed in the countries that are directly
represented.
Technical databases
Two technical databases of models manufactured in 1997/98 were created using data originating
from two manufacturers’ associations, CECED and Eurovent. The CECED data are directly
reported by manufacturers and have not been independently verified. By contrast, the Eurovent data
were certified following the process described in section 4.2 and are published. CECED
manufacturers and Eurovent manufacturers are not mutually exclusive sets since Eurovent
Certification works with all manufacturers ready to have their performance checked by outside
parties, including Eurovent/Cecomaf members as well as CECED members. The data collected by
Eurovent (on multi-split, split, packaged and single-duct RACs) and by CECED (split and single-
duct RACs) cover 80–90% of the RACs of up to 12 kW cooling capacity on the European market in
1997/98.
At the beginning of the study, RACs were distinguished according to their voltage (230 V and
400 V), their mode of integration into the building and their differences in heat-transfer fluid (air or
water) and cycle type (cooling-only or reversible). This led initially to 19 RAC categories within the
21
database, as shown in Table 2.2, though this number decreased as evidence was gathered.
Table 2.2. Classification and size of the RACs in the EERAC technical database
Category Description No. of models Percentage of
total
MS1 Multi-split, 230 V, cooling-only, air-cooled 83 4.1
MS2 Multi-split, 230 V, reverse, air-cooled 49 2.4
S1 Split, 230 V, cooling-only, air-cooled 610 30.1
S2 Split, 230 V, reverse, air-cooled 543 26.8
S3 Split, 230 V, cooling-only, water-cooled 6 0.3
P1 Single-packaged, 230 V, cooling-only, air-cooled 32 1.6
P2 Single-packaged, 230 V, reverse, air-cooled 15 0.7
P3 Single-packaged, 230 V, cooling-only, water-cooled 68 3.4
P4 Single-packaged, 230 V, reverse, water-cooled 52 2.6
SD1 Single-duct, 230 V, cooling-only, air-cooled 58 2.9
SD2 Single-duct, 230 V, cooling-only, water-cooled 13 0.6
MS3 Multi-split, 400 V, cooling-only, air-cooled 20 1.0
MS4 Multi-split, 400 V, reverse, air-cooled 6 0.3
S4 Split, 400 V, cooling-only, air-cooled 228 11.2
S5 Split, 400 V, reverse, air-cooled 185 9.1
P5 Single-packaged, 400 V, cooling-only, air-cooled 15 0.7
P6 Single-packaged, 400 V, reverse, air-cooled 20 1.0
P7 Single-packaged, 400 V, cooling-only, water-cooled 18 0.9
P8 Single-packaged, 400 V, reverse, water-cooled 6 0.3
Total 2 027 100
The database contains statistics on more than 2 000 models on the European market. It is
statistically analysed later in this report (section 4).
National surveys
Separate national surveys have been conducted by members of the study group to provide additional
information such as:
• the historical evolution of the RAC market
• the total number of RACs in use in 1996
• the air-conditioned surface and the average rate of insulation in 1996
• RAC trade marks, manufacturers and market shares
• the different means of distribution and associated costs
• the evolution of the equipment ratios of RACs
• the potential of RAC sales as a percentage of available floor area or houses up until 2020
• the U values of walls, and the comfort expectations of consumers
• the generation fuel mix and customer tariffs.
22
3. ROOM AIR-CONDITIONERS IN EUROPE
3.1. European market for RACs in 1996
The 1996 market share for each RAC type has been estimated firstly from the distribution of models
on the market in the technical database (assuming the market share is proportional to the number of
models available) (Table 3.1). Such figures will henceforth be called model-weighted figures.
Table 3.1. Number of models by RAC type in the EU in 1996.
RAC type Number %
Multi-split 158 7.8
Split 1 572 77.6
Packaged 226 11.1
Single-duct 71 3.5
Total 2 027 100
The 1996 EU RAC market has also been derived from national surveys of sales carried out for this
study, as summarised in Table 3.2. Such figures can be called sales-based figures.
Table 3.2. EU RAC market share by type in 1996, as determined in national surveys.
Country Total sales Split Multi-split Single-duct Single-
packaged
Austria 23 800 10 000 6 040 3 000 4 760
France 177 000 99 750 29 000 38 250 11 000
Germany 194 500 65 000 19 500 90 000 20 000
Greece 150 880 138 000 12 880 600 1 000
Italy 439 490 363 360 20 350 42 127 13 653
Spain 318 000 250 000 0 39 000 29 000
Portugal 45 800 35 600 7 400 900 1 900
a
UK 133 800 104 000 0 26 000 3 800
a
Others 116 700 39 000 11 700 54 000 12 000
EU 1 599 970 1 104 710 106 870 293 877 97 113
Market share 100% 69% 7% 18% 6%
a
Estimated
From these two sources it is possible to compare the model-weighted estimate of the relative sales-
weighted importance of the RAC categories on the European market with the sales-based estimate,
as shown in Table 3.3.
Table 3.3. Model-weighted and sales-weighted estimates of EU market shares for each type of
RAC in 1996.
Model-weighted Sales-weighted
Multi-split packaged units 8% 7%
Split-packaged units 78% 69%
23
Single-packaged units 11% 6%
Single-duct units 3% 18%
The two estimates diverge seriously only in the case of the single-duct RACs, which are known to
be under-represented in the database and which also probably have higher sales per model than
other types. The figures show that a good consideration of both split units and single-ducts is
necessary for adequate coverage of the European market, not just the split-unit segment.
3.2. EU production, imports and exports
As with other domestic appliances, a number of EU countries have specialised in the manufacture of
RACs, either because of the proximity of the market or because of their competitive advantage. Not
all the figures are known, but it is likely that the trade is mostly intra-European. Europe maintains a
high degree of self-sufficiency (some 70% of the market being produced locally, even if partly
exported), despite the relative smallness of its market in comparison with the USA and Japan.
Table 3.4 presents the number of units of RACs produced in some Member States. A decrease in the
market in 1993 affected Italian production. The number of units manufactured doubled between
1993 and 1995.
Table 3.4. Number of units manufactured between 1992 and 1996.
Country 1992 1993 1994 1995 1996
France 124 480 137 050 150 900 176 100 182 300
Greece 10 000 12 000 17 000 21 300 20 500
Italy 344 550 311 448 386 674 564 146 459 867
These countries do not saturate the European market. A substantial flow of RACs is bought from
and sold outside the EU, as shown in Table 3.5. These figures are largely indicative because
customs statistics are not detailed enough to differentiate RACs from other air conditioning
equipment.
Table 3.5 Comparison of RAC production, imports and exports by country in 1996 (thousands).
Country Market Domestic Imports from Imports from Exports to EU Exports to non-
production EU countries non-EU countries EU countries
countries
Austria 24 0 23 7 6 0
France 177 182 40 32 47 30
Germany 195 123 30 45 3 0
Greece 151 20 106 25 0 0
Italy 439 460 60 242 183 140
Spain 318 127 25 166 0 0
Portugal 46 0 28 18 0 0
UK 134 60 22 52 0 0
Others 117 152 20 60 115 0
Total 1 601 1 124 354 647 354 170
Figures in italic are estimated values
24
The main net exporters of RACs are Italy and France (Table 3.6).
Table 3.6. Number of RAC units exported from France and Italy between 1992 and 1996.
Country 1992 1993 1994 1995 1996
France 47 439 53 511 60 360 72 201 76 566
Italy 169 284 141 833 185 871 330 711 323 296
3.3. The stock of RACs in use in 1996
Based on the national surveys conducted in this study, the average life span of the four principal
RAC types is as shown in Table 3.7. Experts indicate that replacement is largely due to the desire
for a new unit rather than equipment failure, due to good reliability and low annual usage.
Table 3.7. Average life span for each type of RAC.
Category Average life expectancy
(years)
Split-packaged 12.6
Multi-split-packaged 12.6
Single-packaged 12.5
Single-duct 10.3
Given these sales figures and life spans, it was possible to calculate annual stock figures by country,
as shown in Table 3.8.
Table 3.8. Stock of RACs in use by EU country from 1990 to 1996.
Country 1990 1991 1992 1993 1994 1995 1996
Austria 8 600 22 600 29 100 42 600 58 300 70 300 79 000
France 369 200 502 000 645 000 760 900 908 100 1 082 100 1 259 100
Germany 144 000 186 400 261 500 294 400 352 200 444 400 526 100
Greece 76 000 148 010 223 450 313 020 439 750 593 950 744 830
Italy 198 900 396 450 641 050 871 400 1 000 000 1 672 640 2 111 740
Spain n.a. n.a. 350 000 549 000 777 000 1 051 000 1 369 000
Portugal 136 670 159 900 199 250 230 210 261 400 295 930 322 820
a
UK 153 112 222 612 290 212 354 312 431 612 540 612 674 412
Othersa 86 400 111 840 156 900 176 640 211 320 266 640 315 660
Total 1 172 882 1 749 812 2 796 462 3 592 482 4 439 682 6 017 572 7 402 662
a
Estimates
Overall, Italian owners accounted for 28% of the total number (7.4 million) of room air-conditioners
in use in the EU in 1996.
A breakdown of the figures showing the proportion of each of the four RAC types in use within
each country in 1996 is shown in Table 3.9.
25
Table 3.9. Numbers of each type of RAC in use within each EU country in 1996.
Country Split Multi-split Single- Single-duct Total
packaged
Austria 33 400 21 300 16 600 7 700 79 000
France 752 000 183 850 106 500 216 750 1 259 100
Germany 198 600 59 600 74 500 193 400 526 100
Greece 138 000 51 830 555 000 * 744 830
Italy 1 504 697 90 177 134 860 382 006 2 111 740
Spain 972 000 * 245 000 152 000 1 369 000
Portugal 267 157 30 143 17 720 7 800 322 820
UK 516 790 * 54 867 107 755 674 412
Others 119 160 31 100 44 700 116 040 315 660
Total 4 501 734 468 000 1 249 747 1 183 451 7 402 662
Percentage of 61% 6% 17% 16% 100%
total stock
* Included in the split-packaged units
The fraction of the stock comprising split RACs (61%) in 1996 is lower than that for new
appliances on the market (69%), which confirms the continuing growth in the market share for this
type of unit at the expense of single-packaged ‘window’ units. The same applies to single-duct
RACs (16% versus 18%).
3.4. Operation of the market
Both the present and future impact of RACs must be assessed globally and locally, because the
various national electricity distribution utilities and the various customers have different needs.
RAC penetration by user sector
The number of households equipped with central air-conditioners is negligible in Europe, such that
RACs are the dominant system in households with electric air-conditioning (though these number
no more than 2% of all households). ‘Penetration’ is defined as the number of households with one
air-conditioner. Penetration rates in the EU show that domestic air-conditioning is far less common
than in other industrial countries (e.g. 70% of households in Japan; 55% in the USA), although the
rate in the non-residential sector (offices, hotels etc.) is significantly higher, since different
construction techniques are used and different levels of comfort are required.
Main trade marks
The total European market is equivalent in size to the domestic market of just one of the big US or
Japanese manufacturers. Japan alone has an internal market of 8 million units per year, 5 times the
size of the EU market!
According to the information gathered for this study, it appears that about 10 manufacturers operate
on an EU-wide level. This is not a result of large-scale production but reflects the fact that a number
of countries have a significant market and that there is a transfer of technology from countries with a
large market (e.g. Japan, the USA) to their European branches. Of the big 10 manufacturers
mentioned, at least seven conform to this pattern. This does not mean that the local companies have
no technical autonomy, but it does explain the structure of the market.
The 10 main trade marks have been ranked in order of market share, as estimated from national
26
market-share data (Table 3.10). They have been determined from the sum of the total number of
RACs sold under each trade mark in each of the five countries (Greece, France, Italy, Spain and
Portugal) for which data were available. The ‘100%’ shown in the table in fact represents 62% of
the national markets of these five countries, the total number of RACs sold across all brands being
1 131 171.
Table 3.10. Dominance of main trade marks sold in Greece, France, Italy, Spain and Portugal (1996).
TRADE MARKS Number of Percentage among
RACs sold in 1996 top 10
A 146 900 20.8
B 108 905 15.4
C 97 737 13.9
D 74 229 10.5
E 62 790 8.9
F 60 420 8.6
G 41 680 5.9
H 40 314 5.7
I 38 644 5.5
J 33 425 4.7
Total 705 044 100%
The 23 other trade marks specifically mentioned in the surveys sold a total of 320 154 units in 1996
in these five countries.
Distribution structure and customer demand
One factor determining RAC distribution by type is the increasing ease of installation and use.
‘Window’ units and multi-splits still require a fitter. Splits and, increasingly, single-ducts are
becoming more widely available in department stores and DIY retailers. Splits are becoming partly
movable in response to customer demand.
The domestic market for RACs is considered to be an ‘impulse’ market. This is true even if the
RAC is installed professionally. The growing proportion of directly purchased RACs leads logically
to an increase in the number of single-duct units discharging through partly opened windows,
despite manufacturers’ recommendations. The volume of RAC sales correlates with outdoor
temperature ( if there is a long period of hot weather, i.e. more than two weeks, sales figures
increase rapidly).
Multi-split units are distributed and installed by heating and cooling plumbers because their
installation usually requires specialised knowledge of air-conditioning and ventilation technologies.
In northern European countries, almost all window RACs are distributed by heating and cooling
plumbers and electricians, while in some southern Member States they are also available directly
from shops.
As in other parts of the world, room air-conditioning in Europe is needed particularly in buildings of
poor quality. Over the last few years, increasing numbers of lofts in old houses have been converted
into flats in order to gain new living space. These flats can become very hot in summer and must be
cooled, especially when they are badly insulated. The costs associated with professional installation
27
are not negligible. We have estimated installation costs for split systems, the dominant segment of
the market, in various EU countries and these range from 15% to 30% of the purchase price.
Since not all appliances are installed by a professional installer, the most important piece of
information that is widely available (the nominal EER) should be attached to the appliance rather
than be given only to the installer. It is true that the thermal quality of the building has a very
important impact (as important as the nominal EER) on the amount of electricity consumed by an
RAC.
Past market evolution
It is important to consider recent growth trends in the RAC market (Figure 3.1).
Figure 3.1. Total sales of RACs in the EU in recent years (1990 - 1996).
EVOLUTION OF TOTAL RAC SALES
1600000
1400000
1200000
Other E.U*
U.K
1000000
PORTUGAL
SPAIN
800000 ITALY
GREECE
GERMANY
600000 FRANCE
AUSTRIA
400000
200000
0
1990 1991 1992 1993 1994 1995 1996
Figure 3.1 reveals a significant decrease in sales of all RACs in 1993, regardless of the type of unit.
In particular, total sales in Spain declined by 42% from 1992 to 1993. A less marked decline
occurred in other countries during the same period, with annual sales decreases of between 5% and
10%. This major sales dip could have resulted from the economic recession of 1993, with an overall
reduction in the purchase of ‘comfort appliances’, together with a cooler summer than in 1994 and
1995.
Evolution of the southern European market
In order to explore the possible impact of various policy measures, detailed country-by-country
forecasts of the EU RAC market to the year 2020 will be made and will take into account the
increase in the number of households (population change, change in household size) and in
commercial customers and the increase in RAC penetration rate (ownership levels).
In fact, in some cases the same analysis may be performed by climatic zone for the reasons
discussed in section 2.4. For example, construction techniques and climatic conditions in Spain,
28
which is a large potential market, mean that there are three distinct areas with differing RAC
penetration rates, which should be treated separately: the north coast, like Portugal, will always have
limited penetration; the interior plateau, with its continental climate, is likely to experience high
growth; while the Mediterranean coast is between the two.
The Italian market is showing the first signs of saturation, but it still has further growth potential
owing to transfer from the lowest segment of the market to the higher. In Portugal, meanwhile,
according to estimates from CCE and the manufacturers’ association, the number of RACs sold
there will increase by a factor of five within 24 years. According to recent market studies performed
by IDAE, Spain has a high potential for RAC sales growth, with more than 5 million households
having thermal and climatic characteristics that may generate a demand for this kind of equipment.
Moreover, manufacturers are of the opinion that the Spanish market will double in the next 10
years, after which it will stabilise. The forecast annual growth rates are 1% for Italy and 7% for the
most promising countries: Greece, Portugal and Spain.
29
4. ENERGY EFFICIENCY ON THE EUROPEAN MARKET
4.1. Reference lines and models
RAC performance distributions
As an initial step, the empirical distribution of the EER (energy-efficiency ratio) was studied for the
19 categories defined in section 2.5 and for the total market (Figure 4.1).
RAC energy performance by category
The percentage difference between the maximum (or minimum) EER and the average (EERaver) is
shown for each RAC category in Figure 4.2.
Figure 4.1. Distribution of EER/EERaver for all
RACs on the EU market (1996).
Figure 4.2. Distribution of EER/EERaver per type
40
35
30
25
models (%)
20
15
10
5
0
91-100
101-110
111-120
121-130
131-140
141-150
151-160
161-170
61-70
71-80
81-90
60
EER/EERaver (%)
Average nominal EER values for each category are given in Table 4.1.
Table 4.1. Minimum, average and maximum performance on the European market
RAC category Cycle type Heat transfer EER
230V mono-phase cooling mode fluid min ave Max
MULTI-SPLIT Cooling only Air 1.91 2.70 3.74
Reverse Air 2.08 2.53 2.94
Air 1.54 2.53 3.56
SPLIT Cooling only Water 2.70 2.75 2.88
Reverse Air 1.45 2.48 3.45
Air 1.88 2.38 2.77
PACKAGED Cooling only Water 2.11 3.32 5.42
Air 1.93 2.32 2.84
Reverse Water 2.26 3.20 5.31
Air 1.35 2.07 3.09
SINGLE-DUCT Cooling only Water 2.10 2.33 3.62
30
400V three-phase cooling mode Fluid min ave max
MULTI-SPLIT Cooling only Air 1.91 2.66 3.32
Reverse Air 2.10 2.34 2.55
Cooling only Air 1.59 2.40 3.25
SPLIT Reverse Air 1.70 2.46 3.20
Air 1.79 2.38 2.97
PACKAGED Cooling only Water 3.08 3.55 4.39
Air 1.79 2.44 2.97
Reverse Water 2.42 3.67 4.33
a
Including some certified values and some manufacturer-reported values; values for single-duct RACs include specific
uncertainties, as discussed in section 2.6.
Within each general category the water-cooled systems have systematically higher EERs, but the
water consumption is not negligible. The lowest EER values are found in the SD (single-duct)
category. In most cases (except in three categories that are almost non-existent on the market:
PA/E/Wt, MS/R/At, SP/C/A, see figure 4.2), the user will find high-performing and poorly
performing models in any one category.
However, if the total range of performance is large, it should be said that most models concentrate
around the average. Some 66% of the RACs in the database are in the range of 90–110% of the
average EER. Some 90% are in the range of 80–120% of the average EER. It seems that only a
small percentage of appliances are either very efficient (4%) or very inefficient (6%).
Dependence of EER performance on cooling capacity – reference lines
Linear regressions of the EER against cooling capacity (the only performance parameter currently
made available to purchasers) were computed for each RAC category (one of which is shown in
Figure 4.3). The results, which produced more or less flat regressions, showed that there was no
significant relationship between RAC EER and cooling capacity. The reference lines that can be
defined in terms of cooling capacity can only be horizontal lines.
Figure 4.3. EER vs cooling capacity (split, 230 V, cooling-only, air-cooled air-conditioners)
7
6
5
EER (Pc/Pe)
4
3
2
1
0
0 2 4 6 8 10 12
Total cooling capacity - Pc (kW)
31
Cluster theory and the selection of reference models
For reasons of economy the techno-economic analysis requires the identification of some base-case
RACs to represent the typical design and cost-performance characteristics of an entire class of air-
conditioners. The following four steps have been implemented to identify such base-case models in
this study:
• definition of RAC categories (see section 2.3)
• statistical clustering inside categories (for large ones)
• whenever possible, elimination of redundant clusters or categories (e.g. if there are too few
models in the given cluster for it to be statistically representative)
• selection of representative models in the remaining clusters.
The process first entailed the selection of the 10 most significant of the 19 RAC categories,
followed by the identification of 20 different clusters. Since cluster theory is comparatively new, the
general principles are reviewed later. Its application has only been appropriate in this case because
the available data are of a high quality, which is not always the case in market studies. These data
allowed the focused technical study which is reported in Chapter 7.
Noise levels
It should be mentioned that testing of RAC noise levels shows that there is an enormous range in
acoustic performance though there is a clear relationship with cooling capacity, i.e. smaller RACs
are generally far quieter.
4.2. Previous market-transformation efforts within EU Member States
Utilities such as EdF (France) and ENEL (Italy), national energy agencies, ministries, consumer
associations etc. have in the past made attempts to advise the wider public about better equipment
on the market in order to help manufacturers promote energy efficiency, for instance. As an example
of these local or national efforts, which until now have been temporary and not coordinated across
the unified market, one of the most ambitious campaigns, currently under way, is the ‘EdF directory
of energy-efficient reversible air-conditioners’, discussed here.
An example of a utility-led energy-efficient RAC promotional campaign
EdF aims to promote reversible air-conditioners as the principal heating source in new French
households and is directing its promotional efforts in the following ways:
• the branding of some global commercial offers (not just the appliances) by EdF under the name
of PROMOTELEC, which also serves to promote some non-competitive electricity uses
• the provision of financial incentives under the VIVRELEC campaign to thousands of new
houses if they attain thermal performance levels that exceed the current mandatory building
regulation levels (which are expected to remain stable)
• the affiliation of reversible air-conditioner manufacturers with the business association GIE
Climatisation et Développement, which carries out collective advertising aimed at RAC market
development, in exchange for which affiliates sign a formal agreement with EdF to submit
products for inclusion in EdF’s new ‘directory’
• the establishment and promotion of the ‘EdF directory of energy-efficient reversible air-
conditioners’ (available from February 1999), which will indicate the appliances on the market
32
that are efficient enough to be used in the PROMOTELEC/VIVRELEC schemes previously
described.
Participating manufacturers must supply Eurovent with more test-point data than the minimum
required to enter Eurovent (which are currently the T3 test results in the heating mode and the T1
test results in the air-conditioning mode). Supplying this data to Eurovent means that it can be
independently verified. EdF also obliges products promoted within the directory to attain a
minimum energy efficiency value under each test condition, as defined in Table 4.3. All values used
are Eurovent certified values, subject to independent testing.
Table 4.3. Threshold energy performance values for inclusion of RACs in the ‘EdF directory of energy-
efficient reversible air-conditioners’.
Mode External unit Internal unit Minimum COP
Dry-bulb Wet-bulb Dry-bulb Wet-bulb or EER
temperature temperature temperature temperature
(°C) (°C) (°C) (°C)
Cooling mode 27 19 35 24 2.4
Heating mode 7 6 20 12 2.7
–7 –8 20 12 1.6
Abbreviations: COP = coefficient of performance; EER = energy efficiency ratio.
The management of the scheme will be conducted by Eurovent Certification, so that the French
directory of energy-efficient products will coexist with the European directory of certified products.
The testing expenses will be covered by the manufacturers wanting to enter the scheme, and the
management costs will be covered by EdF. The new directory will be used by the PROMOTELEC
field promoters to advise customers about existing efficient models.
Compared with previous temporary national initiatives, the establishment of a permanent European
framework is likely to be more durable and could take advantage of the existing public familiarity
with the A–G energy efficiency ratings currently used to label a number of other types of appliance.
An example of a national scheme to promote energy-efficient RACs through building thermal
regulations
As described in section 2.4, thermal insulation, which is often introduced into building codes to
limit heating requirements, very often has a positive influence on cooling needs; however, in other
cases increased insulation could create summer discomfort and lead to the purchase of a new
appliance.
European national building codes usually deal only with winter energy consumption and are not
intended to influence energy consumed in room air-conditioning. It appears that only Portugal has
an alternative approach, that of imposing limiting values on both summer and winter energy needs.
The idea is that the building code will result directly in reduction of energy consumption in summer
and indirectly through a reduced need for appliance installation because of acceptable comfort
levels. Furthermore, the regulations require the installation of a central system when the cooling
load is above 25 kW.
4.3. The Eurovent Certification programme
A transnational performance-certification programme is managed by Eurovent Certification, which
33
is a business association created for this purpose, not a notified body. Equipment with Eurovent
Certification has been selected independently (not by the manufacturer) and then tested according to
international standards in a replicable way (the same testing equipment is used for all tested
appliances). Models to be tested are selected randomly by Eurovent. Specific units are picked
directly from the production lines. Though testing is conducted independently of the manufacturers,
they have to cover the expenses of both the management of the programme and the testing.
Organisation of the programme
The objective of participating manufacturers is to have their appliances’ performance tested and
reported in the annual directory, which is circulated to around 20 000 consultants and installers, and
to be allowed to use the Eurovent label (Figure 4.4). For a cost of less than 0.2 % of their total
production costs manufacturers can, depending on the number of models, have all their models
listed in the directory.
About 10% of all models on the European market are actually tested every year. The fear of a test
result that contradicts their own information has led a number of manufacturers to readjust the
indication of EER in their commercial documentation. It should be noted that the values claimed in
the participants’ literature should not differ from the actual Eurovent laboratory test results by more
than 8% on the EER (combining error in measurement and error in sampling).
Effectively, errors are in fact lower than this value. The average margins of error are 5% in cooling
capacity and 1% in electrical capacity, giving a possible total margin of about 6% on the EER.
Combined with a measurement error of the same order of magnitude, this could mean in an extreme
situation a 12% error margin, rather more than the accepted allowable error of 8% and much higher
than what is generally found. In practice, when data from various laboratories are examined, the
differences are found to be less than 3% for cooling capacity and 1% for power input. This could
mean an effective 4% error margin on the values stated in the catalogue (far less than the values just
reported under the standard).
Marking of appliances
With this definition of acceptable errors, one can rely on the values given in the Eurovent directory
for studying the european market. The equipment tested in that way and included in the directory is
marked with the following label:
Figure 4.4 Label stamped on Eurovent certified appliances.
34
A high percentage of the European RAC market is already included in the Eurovent scheme (for
instance, 99% of our database, and probably 80-90% of the total market). The 10 most important
manufacturers listed in section 3.5 all participate in the scheme. At present single-duct units have
not been included in the Eurovent Certification programme.
4.4. The European Commission’s efforts to raise RAC energy efficiency
The Directive on Efficient Appliances defines a framework for policy actions. The European
Commission may combine these informative measures with other measures such as minimum
performance values and voluntary agreements to accelerate the penetration rate of more efficient
appliances.
The rating system (using letters A–G) already used to label other types of appliance has proved to be
educational and may be appropriate for RACs. The commitment of CECED, the dominant
association in the area of household appliances, has increased the impact of these measures. Under
this E&E umbrella, specific voluntary agreements have been signed with the Commission. No such
result has been obtained yet for RACs.
A contract was awarded to TNO (the Netherlands) (Contract SAVE no. XVII/4.1031/Z/95-055) on
‘Energy labelling of room air-conditioners’, the results of which have been available for the current
study. The TNO study team proposed to compare every single appliance’s EER, as certified by
Eurovent, with the average EER for all Eurovent appliances in the same category.
Meanwhile, as a partial result of the present study, the Commission has sent a mandate to CEN in
order to improve the accuracy of EN 814, particularly with respect to cooling capacity and other
specific RAC features. In particular, the aim is to identify measurement methods that:
• cover all existing types of RACs
• allow for testing at full and partial loads
• allow for testing in a number of different climatic conditions
• consider reversible function (heating mode)
• develop adequate measurement methods for single-duct units
• improve accuracy and replicability for small-capacity equipment.
35
5. MARKET TRANSFORMATION OUTSIDE THE EU
5.1. Minimum energy efficiency standards and labelling schemes in non-European
countries
Table 5.1 summarises known energy-performance test protocols, mandatory and voluntary energy-
rating labelling programmes, minimum efficiency standards and voluntary agreements applicable to
RACs in various countries around the world as well as the EU. Detailed reports have been made on
each. The applicable testing standards in these countries have been reviewed as well, and these
indicate a high degree of convergence with the relevant ISO standards (the parent standards of
EN 814).
Table 5.1. RAC energy-performance programmes around the world.
Country Energy Energy label Minimum efficiency standards
protocol Mandatory Voluntary Mandatory Voluntary
Australia a Yes No a No
Yes No
Canada Yes Yes No Yes No
China Yes No a a No
No Yes
Taiwan Yes No No Yes No
EU a a a a a
Yes No No No No
Hong Kong Yes a Yes No No
No
Indonesia Yes a a No No
No No
Israel Yes a No No No
No
Japan a a No No Yes
Yes Yes
Mexico Yes Yes No Yes No
New Zealand a a Yes a No
Yes No No
Philippines Yes Yes No Yes No
Russia Yes No No Yes No
South Korea a Yes No Yes Yes
Yes
Thailand a a Yes a No
Yes No No
USA Yes Yes Yes Yes No
a
Under review.
5.2. Present situation in the USA
Minimum efficiency standards in the USA
Under the terms of the 1987 National Appliance Energy Conservation Act (NAECA) the USA
introduced mandatory minimum energy efficiency standards (MEES) for RACs which became
effective on 1 January 1990. These were followed by MEES for central air-conditioners in 1992 and
for split and single-packaged systems in 1993.
In 1994 the US Department of Energy (USDOE) proposed a revised standard for RACs based on a
36
fresh techno-economic analysis. This was circulated for stakeholder review and eventually led to an
agreement being reached between industry and energy-efficiency advocates for the next round of
standards. The new standards were published in October 1997 and will enter into force on 1 January
2000 (Table 5.2). The new standards not only include tougher energy efficiency ratio thresholds but
have added a new RAC category for units installed in casement windows (narrow windows opening
on vertical hinges). The USDOE is currently reappraising the existing standards for central air-
conditioner units.3
Table 5.2. US minimum energy efficiency standards for RACs, defined in 1997 and applicable from 1 January
4
2000.
Product class and cooling capacity (Btu/h) [kW] EER (Btu/Wh) [W/W]
Without reverse cycle and with louvered sides
20 000 [5.86 kW] 8.5 [2.49]
Without reverse cycle and without louvered sides
20 000 [5.86 kW] 8.5 [2.49]
With reverse cycle and with louvered sides
7.1 kW 2.45 (3%) 2.62 (3%) 2.59 (5%)
Abbreviations: CC = cooling capacity, EERc = EER in cooling mode, COPh = COP in heating mode.
Table 5.4 The 1997 Japanese target values for cooling-only air-conditioners of ≤27 kW capacity.
Category and cooling capacity Cooling EER 1992 reference EER
Integrated units
≤4 kW 2.45 (10%) 2.23
4 kW 7.1 kW 2.45 (5%) 2.34
Weighted average for all categories 2.93 (6%)
Abbreviations: CC = cooling capacity; EER = energy-efficiency ratio.
The new ‘Top Runner’ energy efficiency thresholds announced in March 1999 and applicable from
7
Egan K, du Pont P. Efficiency standards and labeling in Asia: briefing paper [draft]. International Institute for Energy
Efficiency, Bangkok, March 1998.
8
Rosenquist R. ‘Technical aspects of window air conditioners.’ Presented at the Institute for International Education’s
forum on Developing Effective Energy Efficiency Product Labeling Programs, Washington, DC, 26 March 1997.
9
Tanabe K. Standards and labeling activities in Japan. The Energy Conservation Center, Japan; presented at the Forum
on Asia Regional Cooperation on Energy Efficiency Standards and Labeling, International Institute of Energy
Conservation, Thailand, 14–16 July 1997.
41
either 2004 or 2007 are shown in Tables 5.6 and 5.7.
42
Table 5.6: Revised Japanese efficiency targets for heat pump air conditioners up to 27 kW, applicable from
2004
Category Energy efficiency target
Cooling capacity ranges Cooling and heating
average
(EERc + COPh)/2 (W/W)
Split systems CC ≤ 2.7 kW 5.27
2.7 kW < CC ≤ 3.3 kW 4.88
3.3 kW < CC ≤ 4.2 kW 3.63
4.2 kW < CC ≤ 7.25 kW 3.15
7.25 kW < CC ≤ 27 kW 3.10
Window systems 0 kW < CC ≤ 27 kW 2.85
Multi-split systems CC ≤ 4.25 kW 4.12
4.25 kW < CC ≤ 7.25 kW 3.23
7.25 kW < CC ≤ 27 kW 3.08
Single-duct systems 0 kW < CC ≤ 27 kW 3.02
Other systems CC ≤ 3.3 kW 3.94
3.3 kW < CC ≤ 3.9 kW 3.19
3.9 kW < CC ≤ 7.25 kW 3.13
7.25 kW < CC ≤ 27 kW 3.04
CC = cooling capacity
Table 5.7: Revised Japanese efficiency targets for cooling only air conditioners up to 27 kW, applicable from 2007
Category Energy efficiency target
Cooling capacity ranges EERc
(W/W)
Split systems CC ≤ 3.3 kW 3.64
3.3 kW < CC ≤ 4.1 kW 3.08
4.1 kW < CC ≤ 7.1 kW 2.90
7.1 kW < CC ≤ 27 kW 2.81
Window systems 0 kW < CC ≤ 27 kW 2.66
Multi-split systems CC ≤ 7.0 kW 3.23
7.0 kW < CC ≤ 27 kW 2.47
Single-duct systems CC ≤ 7.1 kW 2.72
7.1 kW < CC ≤ 27 kW 2.71
Other systems CC ≤ 7.1 kW 2.88
7.1 kW < CC ≤ 27 kW 2.84
CC = cooling capacity
43
5.4. The European market situation compared with other OECD countries
It is pertinent to this study to compare energy-efficiency levels of RACs sold in the EU with those
sold in other regions of the world.
Single-packaged RACs
Table 5.5 summarises the percentage and number of single-packaged (window/wall) models in the
Eurovent database that would satisfy the various MEES regulations around the world. In the case of
Australia, which has not implemented MEES, the table indicates the Eurovent models that are above
the average efficiency of the Australian market.
Table 5.5. EU window/wall air-conditioners that would satisfy other regions’ MEES and targets.
MEES or target thresholds Percentage of No. of units No. of units
units passing passing eligible
US 2000, louvered, cooling-only 0.0 0 16
US 2000, cooling-only, unlouvered 81.0 68 84
US 2000, reversible, louvered 25.0 1 4
US 2000, reversible, unlouvered 84.1 53 63
Philippines 2002 MEES, all 66.5 111 167
Taiwanese 1996 MEES, all 88.0 147 167
Japanese 1997 targets, cooling-only 85.0 85 100
Japanese 1997 targets, reversible 89.6 60 67
Korean 1997 targets, all 52.7 88 167
Korean 1991 MEES, all 74.3 124 167
Chinese 1989 MEES, all 87.4 146 167
A
Australian average, cooling-only, fixed 78.0 78 100
A
Australian average, reversible, fixed 86.6 58 67
A
Australia has not implemented MEES: the values in the table indicate the percentage of EU
models which exceed the average EER of Australian RACs.
Split systems
Table 5.6 summarises the percentage and number of split-packaged RACs in the Eurovent database
that would satisfy the various MEES regulations around the world. In the case of Australia, which
has not implemented MEES, the table indicates the Eurovent models that are above the average
efficiency of the Australian market. The NAFTA central air-conditioner MEES are not included as
they are calculated using a seasonal EER value that is not directly comparable with the European or
ISO 5151 test procedure.
Table 5.6. EU split-packaged air-conditioners under other MEES and targets
MEES or target thresholds Percentage of No. of units No. of units
units passing passing eligible
Japanese 2007 targets, cooling-only 2.3% 13 575
Japanese 2004 targets, reversible 9.7% 52 535
Japanese 1997 targets, cooling-only 35.3 203 575
Japanese 1997 targets, reversible 86.9 465 535
Korean 1997 targets, all 20.5 227 1110
Korean 1991 MEES, all 48.7 541 1110
44
Chinese 1989 MEES, all 76.6 850 1110
A
Australian average, cooling-only, fixed 43.8 252 575
A
Australian average, reversible, fixed 87.9 470 535
A
Australia has not implemented MEES: the values in the table indicate the percentage of EU
models which exceed the average EER of Australian RACs.
Figure 5.5 shows the Eurovent EER data against cooling capacity data for the split systems, with the
associated international MEES thresholds.
Figure 5.5. Energy-efficiency ratios for split-packaged RACs in the Eurovent database and international
MEES thresholds.
Split, cooling only, air
5.2
Split, cooling only, water
4.8
Split, reversible, air
4.4
Japanese 1997 targets, cooling
4.0 only
Japanese 1997 targets,
EER (W/W)
3.6 reversible
Korean targets, all
3.2
Korean MEES, all
2.8
China MEES, all
2.4
Australia average, cooling,
2.0 fixed
Australia average, reversible,
1.6 fixed
Japanese 2007 targets, cooling
1.2 only
1 2 3 4 5 6 7 8 9 10 11 12 Japanese 2004 targets,
Cooling capacity (kW) reversible
Conclusions from international comparison
The principal conclusion of this comparative work is that there has been considerable legislative
activity to improve air-conditioner efficiency and that a large proportion of RACs currently
available for sale in the EU would not satisfy efficiency requirements in many other countries
around the world. This suggests that there is significant scope to improve RAC energy performance
in the EU and that to do so requires not technological innovation but merely implementation of
well-established higher-efficiency design options. European or national efforts reported previously
in this report appear very limited when compared with the schemes being applied in any country or
group of countries considered here.
45
6. PROJECTIONS TO YEARS 2010 AND 2020 (BAU SCENARIO)
6.1 Computation of energy consumption of appliances
The main problem for computing consumption is that there is very little information available on
the actual use in the field of room air-conditioners in the residential sector within the European
Union. No public results of measurement campaigns (which are very helpful, even if they refer only
to a few number of households) are available.
Computer models to simulate the performance of RACs
A de-coupled approach (building/system) has been used in this study. In the first instance, cooling
needs (sensible + latent) have been computed using COMFIE, a dynamic multi-zone building
thermal simulation program, after which the electricity consumption has been derived from the
simulated cooling needs and from the known performance characteristics of one of the RACs
investigated. This means that it is assumed that the specific cooling needs are independent of the
cooling system
Methods that take into account variations in RAC energy performance related to outside
temperature and humidity on the basis of T1 conditions only were required, since T2 testing
conditions are not widely used. It was decided to develop some simplified methods based on more
complex computer models. The two modular heat-pump computer models that were extensively
examined, evaluated and used are the American ORNL Heat Pump Design Model – Mark V
(Version 95d) and the French Modelisation Modulaire Windows (MoMo Win) (Version 1.5). The
two software tools were developed using the same modular philosophy, e.g. the various parts of a
heat pump (compressor, evaporator, condenser, flow control devices and connecting lines) are
characterised independently and then linked in order to describe the overall performance of the heat
pump. However, the information that needs to be input into each software tool to characterise a heat
pump is almost the same. The major difference between the two models is the manner in which the
compressor performance is modelled.
From the individual RACs to the stock of RACs
Due to the variations in RAC use, it was important to simulate the dynamics of RAC operation in
function of time (day, months) and to check the final impact on electricity companies’ peak demand.
Another piece of software by Armines has been used for simulating load aggregation and final
consumption.
The software MURELEC by Inestene has been used to analyse the expansion and renewal of RAC
stocks. The purpose of MURELEC (Modèle d’Utilisation Rationnelle de l’Energie Electrique –
model of rational use of electricity energy) is to evaluate trends in specific electricity consumption
and to measure the impacts of eventual actions of demand-side management on energy efficiency.
Data for all these steps are numerous. They have been obtained through ‘national questionnaires’,
from existing climatic databases and from Eurostat.
Part-load performance
The results of testing according to EN 814 was the basis of our study. For our study it was equally
important to establish a link between such nominal performance and part-load (or non nominal
temperature) performance. We have computed a number of functions F, each of them representative
of a cluster of appliances, which link the electrical power demand to the cooling capacity and to the
values of the indoor and outdoor temperatures (T) and humidities (w).
PE = F(PC , TOUTDOOR , w OUTDOOR , TINDOOR , w INDOOR )
46
The energy efficiency ratio under these given conditions can be calculated for each hour of each day
of the week.
Moreover two coefficients of degradation because of fouling and on-off cycles of the air
conditioners were introduced10.
The degradation coefficient for cycling is defined as below.
EERcyc
(1 − )
Cd= EERss
1− F
with Cd=0.25 and EER cyc(cycle) < EERss(steady state).
Condenser and evaporator fouling is a substantial cause of performance degradation of air
conditioners. The impact on the EER is to be determined on the whole life cycle of the machines
(10-13 years). According to an article in HVAC&R research11, a blockage of 56% of the face area of
the condenser of a rooftop air conditioner results in a reduction of 18% of EER. In the same way,
EDF12 noticed that the performance of a tested split could decrease by more than 30% with a
reduction of the nominal air flow of the condenser by 50%. That is the reason why it is estimated
that, as an average value, the EER is reduced down to 20 % because of air conditioners faults
For each base-case simulation, the total energy consumption is given when summing the power
demands (see fig 6.1). Two coefficients ((5/7)*365/12 and (2/7)*365/12) are applied in order to
weight the week and weekend days.
PE,TOTAL = ∑ PE
HOURS
Fig 6.1 Behaviour of the four studied models
The equivalent number of hours at the nominal rating conditions is then calculated by the following
formula:
10
O’Neal D.L., Katipamula S., Performance degradation during on-off cycling of single-speed air conditioners and heat
pumps: model development and analysis’, 1991,. ASHRAE Transactions part B, p. 316-323
11
M. Breuker&J. E.Braun, Common faults and their impacts for rooftop air conditioners, July 1998, HVAC&R
Research, p.303-318
12
J.M. Taldir, Le fonctionnement des climatiseurs individuels split et windows dans les conditions des DOM, Août
1996, DER-EDF
47
PE,TOTAL
n EQUIVALENT =
PE,RATING
RATING is referring to T1 conditions
Presentation of results
As the number of technical options studied and the number of models are numerous, it is not
realistic to perform the technical-economical analysis on each of the base-case simulations.
Consequently, unitary results have to be aggregated in terms of zones and sectors. An EU-average
has then been considered, knowing that the optimum is in fact varying from one country to another,
and from one sector to another. The average equivalent number of hours is given by:
n EQUIVALENT,AVERAGE = ∑n EQUIVALENT
SECTORS , ZONES
In the previous equation, it was decided to weight by the market share of the considered sector and
zone.
The number of running hours under nominal conditions (Table 6.2 )will allow us to calculate the
payback of different technical solutions thanks to a life cycle analysis.
Table 6.2: number of hours of operation resulting from computation
Number of hours at constant EER
Commercial Office Household Hotel Weighted average
Austria Salzburg 177 193 74 235 153
Austria Vienna 134 147 55 176 116
France Carpentras 1414 1307 547 595 1028
France Limoges 790 726 212 314 544
France Trappes 752 625 156 262 468
Germany Middle 431 383 168 236 264
Germany North 199 187 87 115 129
Greece Athens 984 891 741 1530 888
Greece Theso 859 729 480 1175 742
Italy Cagliari 1265 993 822 898 1057
Italy Milano 1017 727 615 726 819
Italy Napoli 1366 966 833 1097 1104
Portugal Lisbon 1226 931 611 413 851
Spain Murcia 2157 1402 1049 1870 1494
Spain Oviedo 678 300 143 382 338
UK London 230 276 94 331 247
We have then defined an average (weighted by the penetration rates in each climatic zone) number
of hours per sector.
Table 6.3 Weighted number of hours for the calculation of energy consumption
W eighted number of hours
Commercial Office Household Hotel
1019 803 519 768
All sectors
773
Detailed values for RAC use in households are specially important for our study. They take the form
of an equivalent number of hours of figure 6.2.
48
Figure 6.2 variations in operating time
Number of hours of air conditioning in the residential sector
94 87
168
156
74 55
212
143
547 615
611
1049 822 833 480
741
The case of water-cooled appliances
When they are used with water coolant the appliances have about the same operating time. If the
water is lost water (either drinkable or not) the EER remains even more constant than with air
cooled appliances, so it remains far better than with air cooling and applying the same number of
hours may seem first too tough and that these appliances behave better than computed.
However, a number of the water cooled appliances are not used autonomously. We may assume that
in the case of closed loops, a 20K penalty is typically originated by the heat exchangers. A rough
estimate of EER loss (based on EER sensitivity described previously for air cooled units) gives to a
2 %/K performance loss in that case, so a 0,4 performance drop. This performance penalty is taken
into account by applying the same operation duration to water cooled appliances which has been
computed for air cooled appliances..
On total, having in mind that a few percent only of european appliances are in that category and that
there are situations both ways, it is not a big source of uncertainty to apply to all appliances the
operating duration computed for air cooled appliance.
6.2 Stock and market in 1990, 1996, 2010 and 2020
A basecase scenario (Business As Usual) has been defined in order to analyse the technical and
economical potential of single or combined measures in several alternative energy efficiency
scenarios. The year 2010 has been chosen for projection due to the Kyoto deadline. A longer term
indication is given for year 2020. For the past, study has been mostly done for 1996 but also for
1990 for consistency with Kyoto conventions.
Evolution of the market
As already described, the extrapolations werebased on 1996 figures and national evolution trends.
Seven types of appliances were treated separately, for which the starting point is geiven in Table
6.4.
49
13
Table 6.4: RAC types and 1996 market
RAC categories % of models % of sales
MS,A: Multi-split, Air cooled 7.8 7.0
SP,A: Split, Air cooled 77.3 68.7
SP,W: Split, Water cooled 0.3 0.3
PA,A: Packaged, Air cooled 4.0 2.2
PA,W: Packaged, Water cooled 7.1 3.8
SD,A: Single-duct, Air cooled 2.0 13.8
SD,W: Single-duct, Water cooled 0.6 4.2
TOTAL 100 100
Our assumptions, based on CECED recomendations, correspond to approximately half of the
potential ownership foreseen by Eurovent for 2020, a conservative hypothesis. We have also taken
into account the market trends of the four equipment as observed on the italian market, the most
mature one, shown on Table 6.6. CECED has indicated that this evolution, and namely the decrease
in Single Ducts market shares, would become general in all countries.
Table 6.6: Evolution of Italian RAC sales to 2020 according to ANIE/COAER
Type of RAC 19 2020 market 2020 stock
96
m
ar
ke
t
Single Duct 9. 2.64% 3.30%
59
%
Split 82 91.14% 90.09%
.6
8
%
Multi Splits 3. 1.66% 4.61%
11
%
Packag. Units 4. 4.56% 2.00%
63
%
Total 10 100% 100%
0
%
The following table indicates this extrapolation.
Table 6.7 Projected share of RAC sales (% of total market)
Market Market Market Market Market
1990 1996 2000 2010 2020
13
"Models" ratio between water and air cooled units was kept for "market" figures
50
Multisplits 4% 7% 14% 14% 14%
Splits 60% 69% 66% 71% 77%
Packaged 19% 6% 7% 6% 4%
Single Ducts 17% 18% 13% 9% 5%
TOTAL 100% 100% 100% 100% 100%
Finally used with a stock model, this leads to the results of Table 6.8.
Table 6.8 Share of RAC types in 2020 stock (% of whole stock)
Single Duct Window Split Multi-Split
Italy 3.30% 2.00% 90.09% 4.61%
Spain 4.49% 6.23% 70.97% 18.31%
Portugal 0.65% 2.55% 81.37% 15.43%
Germany 23.13% 9.58% 52.80% 14.49%
Greece 0.00% 0.39% 91.77% 7.83%
France 8.35% 4.48% 68.86% 18.31%
UK 7.32% 1.99% 71.16% 19.53%
Austria 4.92% 14.56% 51.86% 28.65%
Other EU ** 3.74% 15.04% 51.30% 29.92%
Results on stocks in use and market
The number of units in use in the various countries increases greatly over simulated years. Tables
6.9 to 6.11 give the splitting according to categories, sectors and countries.
Table 6.9 Stock by type and yearly growth
Stock Stock %/year Stock Stock 2020 %1996/
1990 1996 90/96 2010 2020
Multisplits 115516 981313 42.8% 3 567 177 5 597 448 7.5%
Splits 687184 4517047 36.9% 14 627 802 22 953 264 7.0%
Packaged 205227 688477 22.4% 1 270 715 1 993 946 4.5%
Single 164435 1215945 39.6% 1 556 982 2 443 143 3.0%
Ducts
TOTAL 1172362 7402662 36.0% 21 022 676 32 987 801 6.4%
Very conservative assumptions have been made about future market growth.
Table 6.10: Estimated stock of appliances by country and yearly growth
1990 1996 2020 1990/ 1996/
1996 2020
Austria 8600 79000 849144 44.7% 10.4%
51
France 368700 1259100 6353694 22.7% 7.0%
Germany 144100 526100 4664828 24.1% 9.5%
Greece 76000 744830 3039588 46.3% 6.0%
Italy 198900 2111740 3084364 48.3% 1.6%
Portugal 136670 322820 1300000 15.4% 6.0%
Spain ? 1369000 7800051 ? 7.5%
UK 153112 674412 3097236 28.0% 6.6%
Others 86280 315660 2798897 24.1% 9.5%
EU 1172362 7402662 33777613 36.0% 6.5%
The total number of units in the European Union is multiplied by 18 in the period covered by the
Kyoto protocol. Some countries like Spain follow a quicker evolution. Associated electricity
consumption will grow in a similar manner.
6.3 Electricity consumption
The base case computed here is the BAU scenario previously defined.
Description of performance data used
In order for the model to calculate the accurate consumption related to the exact stock of appliances
of each of the four equipment, the shares of Split, Multi-Split, Single-pckaged and Single Duct have
been input into MURELEC. They are combined with the energy consumption values derived from
the Chapter 4 (average size per type).
Table 6.12: Electrical consumption for 1996 per type
Category Description Pc Pe EER
MS1 Multi-split, 230 V, cooling-only, air-cooled 5.23 1.98 2.70
MS2 Multi-split, 230 V, reverse, air-cooled 5.42 2.15 2.53
MS3 Multi-split, 400 V, cooling-only, air-cooled 7.00 2.71 2.66
MS4 Multi-split, 400 V, reverse, air-cooled 6.48 2.77 2.34
MS All air-cooled multi-splits 5.56 2.16 2.63
S1 Split, 230 V, cooling-only, air-cooled 4.87 1.98 2.53
S2 Split, 230 V, reverse, air-cooled 4.93 2.04 2.48
S3 Split, 230 V, cooling-only, water-cooled 7.66 3.19 2.40
S4 Split, 400 V, cooling-only, air-cooled 8.24 3.35 2.46
S5 Split, 400 V, reverse, air-cooled 4.53 1.64 2.75
AirS All air-cooled splits 5.28 2.16 2.48
WaterS All water-cooled splits 4.53 1.64 2.75
P1 Single-packaged, 230 V, cooling-only, air-cooled 4.92 1.55 3.32
P2 Single-packaged, 230 V, reverse, air-cooled 3.66 1.18 3.20
P3 Single-packaged, 230 V, cooling-only, water-cooled 4.36 1.90 2.38
P4 Single-packaged, 230 V, reverse, water-cooled 10.41 4.36 2.44
P5 Single-packaged, 400 V, cooling-only, air-cooled 10.27 2.90 3.55
52
P6 Single-packaged, 400 V, reverse, air-cooled 5.79 2.54 2.32
P7 Single-packaged, 400 V, cooling-only, water-cooled 10.38 4.43 2.38
P8 Single-packaged, 400 V, reverse, water-cooled 11.23 3.15 3.67
AirP All air-cooled packaged units 7.20 3.08 2.38
WaterP All water-cooled packaged units 5.40 1.65 3.32
AirSD Single-duct, 230 V, cooling-only, air-cooled 1.70 0.81 2.07
WaterSD Single-duct, 230 V, cooling-only, water-cooled 2.09 0.88 2.33
Results on consumption
Electricity consumption due to RAC is given in figure 6.3 and Tables 6.13 and 6.14 according to
sectors and countries.
Table 6.13: Electricity consumption by economic sector (GWh/year)
Sector 1990 1996 2010 2020
Households 387.2 2445.5 7483.3 11375.6
Offices 538.0 3398.1 8544.7 13321.2
Small business 735.8 4647.1 9553.7 14339.5
Hotels 266.9 1685.4 3577.4 5393.8
Total 1927.9 12176.2 29159.1 44430.2
Figure 6.3 Growth of energy consumption
E v o lu tio n o f R A C c o n s u p tio n p e r s e c to r - B A U s c e n a r io
45000
40000
35000
30000
25000 h o te ls
in GWh
s m a ll b u s in e s s
o ffic e s
20000 h o u s e h o ld s
15000
10000
5000
0
1996 2000 2005 2010 2015 2020
Table 6.14: Energy consumption of RACs by country (GWh/year)
1990 1996 2010 2020
Austria 68.6 121.3 235.0 364.5
France 331.6 1782.1 5517.2 8975.5
Germany 155.9 672.4 1914.0 3197.3
53
Greece 208.8 1006.6 2281.3 3478.6
Ialy 761.0 4494.1 5743.6 7033.9
Portugal 162.4 713.8 1806.8 2552.2
Spain not av 2496.4 9366.4 15146.6
UK 120.0 446.0 1135.7 1783.8
Other EU 119.6 443.5 1159.1 1897.7
Total EU 1927.9 12176.2 29159.1 44430.2
Contribution to Electricity peak Demand
Electricity demand due to RAC is computed on the basis of the dynamics of the RAC units. Part of
the peak demand is coincident with the grid peak.
Figure 6. 4 Load curve trend
Load fo r a n a v e r a g e w e e k d a y o f a u g u s t- B A U s c e n a r io
45
40
35
30
in GW
25 1996
20 2020
15
10
5
0
11
13
15
17
19
21
23
1
3
5
7
9
Table 6.15: Load peak per economic sector
Unit: GW 1996 2010 2020
Households 4.91 15.29 23.76
Offices 1.35 2.45 3.51
small business 2.03 4.19 6.38
Hotels 2.06 5.53 9.55
Total 10.34 27.46 43.21
Table 6.16: Load peak per country
Unit: GW 1996 2010 2020
Austria 0.01 0.20 0.40
France 0.86 2.81 4.58
Germany 0.24 1.43 2.51
Greece 1.72 4.23 6.74
Ialy 4.25 5.52 6.94
54
Portugal 0.46 1.26 1.84
Spain 2.62 11.01 18.04
UK 0.04 0.14 0.65
Other E.U 0.14 0.85 1.50
Total E.U 10.34 27.46 43.21
6.4 – Environmental impact
Environmental impacts of RAC take place in the atmosphere: acid pollution, ozone depletion, green
house gases emission, but some specific effects can also be mentioned (solid waste, water use).
Global Warming
There are direct effects of RAC on global warming (refrigerants release in atmosphere) and indirect
effects (energy consumption over their life time). R22 is the most commonly used RAC refrigerant;
however, as this fluid has an ozone-depletion and a global warming potential its production will be
prohibited in developed countries. Some RACs already use alternative refrigerants such as R290,
R407C, R-134a and R-410a (these refrigerants are more or less compatible with the running
parameters of a traditional R22 unit). In fact, a higher energy efficiency and a more environmentally
benign refrigerant will both result in a lower contribution to global warming. TEWI (total equivalent
warming impact) is the integrated index used to measure the global-warming impact of all gaseous
emissions, including those from direct and indirect sources. TEWI can be easily computed from
known figures.
However, our study group was not in charge of the issue of change of refrigerants. The direct
contribution of RAC to global warming will drop independantly from our action. The only unsolved
problem is to know if technical potential will remain the same after the refrigerants change, and if
the manufacturers can accomodate the sped of change generated by the various policies of the EU.
We shall concentrate here on our possible range of actions: energy consumption changes resulting in
a lower indirect CO2 release. Since there is a European bubble, one could only consider the average
CO2 content of the European kWh, set to 440 gCO2/kWh, the OECD average. In fact, the exact
figure for CO2 content per kWh were avilabel for each country, this was taken into account.
Water use
In a similar way, the use of drinkable (or not drinkable) water by water-cooled RACs can be easily
computed from known data. All these data should be used in an environmental impact assessment.
Since few people have access to industrial water or to a private well, water-cooled RAC use mostly
drinkable water or need a close loop on a cooling tower. This data not being available, it was
estimated that 20% use drinkable water, 60% use a closed loop and 20% some other solution, like
their private well. The use of waste water for heat rejection is foreseen in research projects (Milano).
In the case of closed loop a 20K penalty is originated by the heat exchangers. This has been already
mentionned and estimated. It is taken into account by applying to the water cooled units the same
operating duration as to the other types.
Something should be said about the case of drinkable water use. The figures in kWh and EER taken
from 6 representative models allow the estimation of the water used per year, and per cooling kW as
37 m3. On the basis of a conservative estimate of the market penetration (20 %) the total drinkable
water use for the purpose was at least 40 000 000 m3 for year 1996 in Europe, a figure that will be
multiplied by 4 by year 2020. Such figures correspond to the use of water of a medium size city
55
(100000 to 500000 inhabitants) and treatment costs over 100 Meuro. For these appliances, the bill
for water will be of the same size as for electricity. A way for warning the buyer should be found.
Regional atmospheric pollution
Atmospheric pollution from power plants is composed of dust, NOx and SO2 which have a regional
impact and CO2 with a world wide impact. We have admitted here that European regulation on acid
pollution was well designed and applied and that a specific effort on acid pollution was not part of
our objectives. It is different for CO2; the Kyoto protocol is recent and its full implementation in
Europe not yet achieved; furthermore there is a European bubble and the trends or measures
considered here can gave directly positive or negative consequences on the achievement of the
European objectives.
56
Results on CO2 emissions
The emission of CO2 in Europe due to RAC have been summarised in Table 6.17.
Table 6.17: Emissions by country
Unit: tonnes CO2 1990 1996 2010 2020
Austria 5 186 9 170 17 765 27 554
France 17 226 92 578 286 612 466 267
Germany 25 358 109 369 311 321 520 055
Greece 151 686 731 261 1 657 288 2 527 087
Ialy 389 162 2 298 202 2 937 174 3 597 010
Portugal 94 768 416 537 1 054 355 1 489 331
Spain not av (may be 1 171 211 4 394 340 7 106 179
300 000)
UK 76 197 283 201 721 146 1 132 676
Other E.U 19 453 72 137 188 533 308 669
Total E.U 779 037 5 183 665 11 568 532 17 174 827
The emission of CO2 in Europe due to RAC is multiplied by 11 in the time of the Kyoto protocol.
The additional emission due to this type of appliance is about 10 out of the 3000 tonnes of the initial
emissions by fuels in the EU. The direct effect of the growth of this market on total EU emissions is
+0,33% to be compared with the -8% target.
57
7. TECHNICAL–ECONOMICAL STUDY OF OPTIONS
7.1 Technical definitions: the four real appliances used, their physical modelling and its validation
Measures for improving RAC energy efficiency are of three main types:
improvement of component performance (better compressor, fan etc.)
optimised design (optimised system temperatures, pressures etc.)
improved operational control.
Our study is based on the EN 814 standard previously described, which doesn’t yet cover all
conditions. Until its extension as a result of mandate M/274, it is necessary to use a computer
model, in our case ORNL Mark V, to extend testing results. The use of improved components and
optimised design will show results at any load and therefore their benefits will be reflected both
under the EN 814 test and in our ORNL computations. For instance, a better compressor will give a
better EER. Split units can use three types of compressor: (i) conventional reciprocating
compressors (one piston in a chamber), (ii) rotary compressors (with a ‘rolling’ compressible space
in a circular chamber) and (iii) in the upper power range, scroll compressors (two spirals creating
continuous compression at high efficiency).
Scroll compressors are capable of operating at different speeds and are becoming more common
among RAC models. However, not all the benefits they bring are measurable when the present EN
814 test protocol is used, which leads to a control problem: how should inverter-driven variable-
speed compressors with significantly higher part-load performance be dealt with? Inverters run
under an adjustable frequency (from 40% to 120% of the nominal value), resulting in varying
compressor speeds (and, as a result, fan speed).
Representation of appliances by four specific units
As already described, the appliances studied were divided into seven types:
Table 7.1: RAC selected for technical analysis
RAC types RAC model studied Share of sales (%)
01 MS,A: Multi-split, air-cooled - 7.0
02 PA,A: Packaged, air-cooled - 2.2
03 PA,W: Packaged, water-cooled - 3.8
04 SD,A: Single-duct, air-cooled C 13.8
05 SD,W: Single-duct, water-cooled - 4.2
06 SP,A: Split, air-cooled A, B, D 68.7
07 SP,W: Split, water-cooled - 0.3
TOTAL 82.5% 100.0
From the seven categories, four specific models close to the centre of some representative ‘clusters’
were selected for technical analysis: three very different models from the largest category (splits,
69% of the market) and one from the second largest category (single ducts, 6% of stock, 14% of the
market). No multi-split model was chosen because they are similar to splits, nor were any models
from the less numerous packaged unit categories.
58
Table 7.2. The four models investigated.
Unit Type (according to 4-part clustering process)
A S5: Split, 380 V, reverse, air-cooled
B S1: Split, 220 V, cooling-only, air-cooled
C SD1: Single-duct, 220 V, cooling-only, air-cooled
D S4: Split, 380 V, cooling-only, air-cooled
All the models studied in detail were air-cooled. In fact, water-cooled appliances are not really
autonomous – they rely on communal or private water (with associated costs and environmental
impacts) or they use a secondary heat exchanger in the outside air (with consequent investment,
temperature and EER penalties). This is described above.
Performance data used
Two types of questionnaire, one for air-cooled and one for water-cooled systems, were compiled in
order to collect the technical data and establish the parameters required for the modelling of the
examined air-conditioners. The questionnaires were submitted to the European associations of room
air-conditioner manufacturers, which then undertook the task of contacting the manufacturers of the
selected representative units (or of other similar units in case it was not possible to obtain the data
for the specific model) and providing us with the required data. The manufacturers provided the
working party with complete data for four units and thus only these units were studied. 14
At first the EER of each model (in its present state) was calculated and compared with the one
measured under standard test conditions as submitted to us by the manufacturers. The results are
illustrated in Table 7.3:
Table 7.3 Comparison of calculated and manufacturers’ submitted EER
Unit Measured EER Calculated EER
A 2.66 2.72
B 2.43 2.48
C 1.90 1.92
D 2.60 2.75
It can be seen that for three of the simulated units (A, B and C) the calculated EER is in very good
agreement with the measured values (2%, 2% and 1% difference, respectively), while the difference
between the calculated and measured EER value of unit D is higher (6%). This difference is due to
the fact that not all the required information was available for this unit and thus default typical
values were used in this case. In conclusion, the modelling of the units is successful and thus the
ORNL software can be used confidently to simulate the performance of the examined units after the
application of the proposed improvements and to evaluate their efficiency.
7.2 Options and technical results
Next, the impact of each of the proposed improvements to the units’ overall performance was
simulated by calculating each unit’s EER after the application of the intervention and under standard
14
It must be noted that data for three other units were submitted, but as these data sets were not complete and a lot of
critical information was missing it was not possible to simulate the performance of these units as well.
59
test conditions. In order to do this, the required modifications were input into the software for each
scenario. However, it was not possible to simulate directly all the scenarios, because of the
software’s limitations, and details from existing literature were also used in some cases.
General results
Various technical options and their effect on EER are shown in Table 7.4.
Table 7.4. Average gain per option (averaged over the four models)
Option Technical improvement Average increase in EER (%)
1(a) Increase frontal coil area (15%) 4%
1(b) Increase frontal coil area (30%) 8%
1(c) Increase frontal coil area (45%) 11%
2(a) Add one refrigerant tube 10%
2(b) Add two refrigerant tubes 16%
3(a) Increase fin density (10%) 10%
3(b) Increase fin density (20%) 16%
4 Add subcooler to condenser coil 1%
5 Improve fin design (modify fin pattern) 11%
6 Improve tube design 8%
7(a) Use of a high efficiency fan motor 1%
7(b) Use of an electronically commutated motor 2%
8(a) Improve compressor efficiency (5%) 3%
8(b) Improve compressor efficiency (10%) 5%
8(c) Improve compressor efficiency (15%) 8%
9 Use of R410a with optimised system 5%?
10 Use of variable speed compressors 12%
11 Use of electronic expansion valves 5%?
12 Use of (fuzzy) controls 4%?
Specific problems of variable speed technology (inverters)
The use of variable speed (inverter-type) compressors instead of single frequency ones could lead,
according to some, to seasonal cooling energy savings ranging from 10% to 40%. A review of the
literature on the energy savings due to the use of inverter technology in room air-conditioners has
been made but suggests that there is not enough evidence to completely support such high figures.
Table 7.5. Studies of inverter use and reported energy savings
Reference Study type Studied mode Power (kW) Energy saving (%)
[SENS85] Experimentation Heating 4.4 29
[MILL88] Experimentation Cooling/heating 8.8 15–20
[MACA88] Simulation Heating 15
[HORI85] Experimentation Cooling/heating 3.0 15
[TORI87] Experimentation Heating 6.9 7-15*
60
[SENS89] Simulation Cooling 14.5 30
[SHIM85] Simulation Heating 6.3 20–40
[PARK88] Experimentation Cooling 3-5 17–18*
[LBNL96] – Cooling 10**
* Gains in terms of EER (or COP) and not in energy consumption.
** The report states that ‘a conservative value’ has been taken into account.
The difference in physical behaviour of ‘inverter’-type RACs lies in their variable speed operation,
for which EN 814 testing is not enough, and so experts could discuss further the reproducibility of
results. Our estimation of the energy efficiency of inverter-type RACs is based on the report of one
single manufacturer on the differences between nominal behaviour (as in EN 814) and practical
behaviour. Despite the magnitude of the possible gains, this lack of reproducibility should prevent
the Commission from immediately promulgating the benefits to be gained by using this type of
appliance. In the medium term, the outcome of the mandate M/274 to CEN and CENELEC will
provide an experimental basis enabling the evaluation of variable speed equipment.
As mentioned above, manufacturers claim that inverters result in energy savings of between 20%
and 40% compared with a traditional unit. Inverters are reported to allow a quicker (as a result of
the available 120% frequency) and stricter (because the unit generally runs 100% of the time)
control of the indoor set temperature. This may produce a significant gain in comfort but not
necessarily a gain in energy savings. Reliability could also be affected.
A description of the main sources of energy gain achieved with an inverter follows here. By
lowering the compressor frequency, the flow rate of the refrigerant fluid decreases, thereby
decreasing the high pressure level. With the resulting lower compression ratio, the electrical power
input to the compressor is reduced. As the total cooling capacity is lower, the unit will run longer
than a conventional RAC in cycle mode, but with a greater EER. A secondary source of energy gain
in inverters is that since they generally run 100% of the time, they avoid energy losses resulting
from restarting that are found with conventional compressors.Most of these benefits could also be
obtained by using certain kinds of multi-speed drives, which have not been tested on RACs but have
on other appliances like refrigerators.
On the other hand, there is concern about the continuous operation of the fan and possible electric
conversion losses, which should not be disregarded in the extension of EN814 testing.
The increase in energy performance (EER) of an inverter-driven air-conditioner at part load is
shown in Figure 7.1.
15
Figure 7.1. Benefits reported by one manufacturer for inverter RACs .
15
The figure is taken from the TNO (1998) interim report.
61
7 Ambient temperature 10 dgr.C
Ambient temperature 18 dgr.C
C. 6
O. Ambient temperature 31 dgr.C
P. 5
(c
oo 4
lin
g) 3
2
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
CAPACITY ratio
To gain an idea of the order of magnitude of potential gains and in order not to delay progress, data
from one manufacturer was analysed and gave 10–12% as a preliminary estimate of weighted gains
by comparing the non-nominal behaviour of the ORNL simulated units and the non-nominal
manufacturer-reported values.
Potential gains have been evaluated in six simulations. For each of them, two figures are given,
according to the value chosen for the coefficient of degradation due to on-off cycles applied to non-
inverter units. The lower estimate (minimum) is obtained with Cd=0 (no performance degradation
due to cycles; all the gains are from constant adjustment of LP and HP levels). The upper estimate
(maximum) is obtained with Cd=0.25.
Table 7.6: Gains with inverters over a full cooling season, computed here
Potential gain in terms of
average EER (%)
Country Sector Min Max
France Small businesses 15 36
Offices 12 29
Households –3 16
Italy Small businesses 15 36
Offices 15 41
Households 9 28
We continued the study assuming an estimated 10–12% gain. Further research, as well as the results
of mandate M/274, will soon bring new evidence. Any remaining uncertainty arises from the
imperfect definition of the quality of control.
Energy consumption of single-duct appliances
The question of whether energy consumption of single-duct RACs can be estimated with the same
computational rules as splits, multi-splits and window units has been raised by the manufacturers’
associations.
By using a hole in the window, the gap in a window that is partly open or a hole in the wall, single-
ducts are likely to be easier to install by consumers themselves. In some cases single-ducts are not
comparable in use to split units for visual reasons (e.g. in historic centres) or for mobility reasons
(mobility of residents, mobility within one house). Other technical solutions exist in each case to
effect cooling (multi-split or central air-conditioners in historic centres; mobile splits) but single-
ducts are less expensive to purchase and easier to install and to move (outside manpower would be
62
needed to obtain the same result with the other systems). So the real question is to compute their
energy consumption.
Their efficiency is generally lower in test conditions. Many experts think that even larger over-
consumption occurs in field situations because of air infiltration. As no published evidence was
available, a computer simulation was made.
The difference in physical behaviour of single-ducts lies in their integration with the room
ventilation. Both units are inside the room and the condenser heats air which is extracted from the
room and rejected, creating an additional ventilation. The air extracted by the RAC is replaced by
the same quantity of outside air both in test and in real situations.
Single-ducts create low pressure in the cooled room, causing infiltration:
from outside (α)*(air flow rate treated)
or from another room of the building. (1- α)*(air flow rate treated)
This infiltration is high. For a 2 350 W unit, the air flow rate treated is typically 350m3/h, as opposed to 50–
100 m3/h for natural ventilation. So, part of the time (when the ventilator operates) cooling loads are
increased:
Total cooling loads = Cooling loads needed to maintain set temperature in the room + Cooling loads due
to additional air leakage from outside + Cooling loads due to additional air leakage from another room
We have computed the field energy consumption of single-ducts on the basis of the simple physical
model of ventilation just described. We have found that, depending on climatic conditions, the
behaviour of single-ducts is in the range of –12% to +7% from the value computed by applying to
them the computational rules defined for the other RACs. Performance is better in northern
climates, where the single-duct provides additional ventilation, and worse in southern countries like
Italy (the bottom of the range here). Computations have not been done for Spain and Greece, for
which over-consumption could reach 15%.
Some physical uncertainties also remain concerning the reproducibility of test results, though the
mandate M/274 to CEN and CENELEC and some collective effort on the part of manufacturers will
solve these in the medium term. This shouldn’t prevent us from using manufacturers’ reported data
for a while yet.
Summary of technical combinations
Evidently the potential savings decrease as more technological options are added since less energy is
available to be saved (or less EER is available to be improved) due to the impact of the previously
added option(s). The most important combinations of technical options were completely re-
simulated, so that no assumption could be made when considering combined EER increases.The
range of possible combinations is the as follows:
Table 7.7 Gains for sets of improvements
TECHNICAL Options EFFECT ON EFFECT ON EFFECT ON EFFECT ON EER AVERAGE
FEATURE considered EER OF A EER OF B EER OF C OF D INCREASE (%)
EXISTING APP. - 2.72 2.48 1.92 2.75
HEAT 1b, 2b, 3b, 3.80 3.69 2.93 3.88 46
EXCHANGERS 5,6
COMPRESSOR 8c 2.94 2.68 2.04 2.97 8
HE+ 1b, 2b, 3b, 3.94 3.79 2.99 4.01 50
5,6, 8c
COMP
63
ALL 1c, 2b, 3b, 3.97 3.81 3.00 4.04 51
SCREENED 5,6, 7, 8c
OPT.
INVERTER 10 3.05 2.78 2.15 3.08 12%
TECH ALONE
HE + COMP + 1b, 2b, 3b, 4.12 3.93 3.09 4.19 56%
INVERTER 5,6, 8c, 10
ALL + 1c, 2b, 3b, 4.14 3.95 3.10 4.21 57%
INVERTER 5,6, 7, 8c, 10
Best on market - 3.20 3.56 3.09 3.25 +37%
in class
(+20%) (+46%) (+63%) (+25%)
In the case of inverters the improvement reflected here by a better nominal EER is in fact an
improvement on seasonal EER which has been applied here in a rough manner, in order to
summarise information.
7.3 Economic calculations for the screening of cost-effective measures
The economic analysis was performed estimating the net benefits of the technological options of the
various models to consumers. This needs a careful study of costs and overcosts and the use of
criteria like net present value of investment or minimum life-cycle cost.
Study of costs and overcosts
RAC models base-case values are given in Table 7.8 for three of the four appliances studied. The
costs are without margins.
Table 7.8 Starting point for the three models optimisation
Total cost Pc Pe EER Energy Cons. Specific
Energy Cons.
2
(EURO) (kW) (kW) (kWh/y) (kWh/y/m )
Model B 485 5.25 2.16 2.43 1 102 21
Model C 361 1.22 0.64 1.9 327 27
Model B-2 520 5.12 2.14 2.39 1 093 21
The individual costs of changes given by the manufacturers’ survey are summarised below in the
form of a range. For the upper limit conservative options have been taken, such as the fact of taking
into account the overcosts in the thermodynamic parts and not the undercosts in the electrical parts
(motors, compressors etc.) that are generated by a smaller electric consumption for the same
nominal cooling capacity and the higher possible margins in improvement.
In fact the normal choice is to keep the cooling capacity constant when you lower the size of the
compressor and motor and increase the heat exchangers. People buy an appliance on the basis of its
cooling capacity, not its electrical consumption! However, one manufacturer reported that the
industry’s choice would be to share 50/50 the overcapacity generated, in order to keep the same
compressor. A conservative estimate was used for the undercost: 20% of variable cost only in
compressor cost while it can reach 50% for large performance steps but may be nil for small
changes remaining in the same ‘frame’.
In the same way manufacturers and traders are not to request the same relative margin (as
64
percentage) for higher-value goods as for lower-value goods; hence a possible marginal value of
20% instead of a regular 40% regular margin. The combination of the two favourable assumptions
gives the lower value of the overcost range that a consumer may encounter.
65
Table 7.9 Overcosts for model B
Single Definition Manufact. EER New New Downs Consumer
options overcost increase undercost overcost
(No.) (EURO) For option EER Pe (EURO) (EURO)
1b Frontal area 46.5 0.08 2.60 2.02 3.8 51/65
HE
2b Depth coil 36 0.16 2.82 1.86 8.4 33/50
HE
5 HE Fin 2 0.11 2.67 1.96 7.2 –6/3
design
6 HE Tube 4 0.08 2.62 2.00 4.8 –1/6
design
8b COMP. Effic. 35 0.05 2.55 2.06 2.2 39/49
9 Alt refriger 100 0.1 2.67 1.96 0 120/140
10 Inverter 47 0.10 2.55 2.06 0 56-66
The manufacturing cost increases are assumed to be passed completely to consumers through price
increases. Two levels of price were assumed: a 40% (base-case) and a 20% (lower value) mark-up
over manufacturing cost increases, to include distribution and sales costs. The average
manufacturing mark-up is 30%; thus a 40% total mark-up implies a 10% mark-up for distribution
and sales. The lowest manufacturing mark-up is 20%, thus a total mark-up of this amount implies
that there is no increase in mark-up for distribution and sales.
Selection of optimum package on four studied appliances (residential sector)
The optimum combination of technological options has been calculated. First the individual options
are sorted according to the payback period (or the ratio of NPV/investment) with the higher-return
options first. Second the energy savings are calculated for the combined options. Finally the net
present value (NPV) is calculated for the combined options. In order to see the impact of adding
each subsequent option, the net present value of adding a specific option is calculated. This is
known as the marginal net present value of adding a given option. Since the options are added in
order of their potential economic contribution, options are added until their marginal net present
value is zero or negative. This determines the optimum design and is also the point at which the
total net present value of the combined options is maximum.
The NPV and life cycle analysis are equivalent. This is due to the fact that the life cycle cost is a
constant value minus the NPV of the improvements; thus the maximum NPV gives the minimum
life cycle cost. The constant value used in the life cycle cost is the annualized cost of the RAC
without improvements. The calculation was initially performed for the residential sector and then
extended to office, shop and hotel sectors.
EU electricity price: 0.127 EURO/kWh
Average operational time: 510 hours/year
Average RAC lifetime: 12 years for all appliances, except 10 years for SD
Annuality factor: 8.39 or 7.36 (discount rate of 6% for 12 or 10 years)
Mark-up to consumer: +40% (base-case) and +20% (lower value)
NPV is positive up to option combination (5+6+2b) for models B and C.
66
Results are given here for model B:
Table 7.10 Economic optimisation of model B
Option Cost EER New EER New Pe New en. Lower Lower Upper
combination increase increase Cons. payback NPV NPV
(EURO) for option (kWh/year) (years) (EURO) (EURO)
5 2 0.11 2.70 1.95 993 0.20 113.60 121.2
(5)+6 4 0.071 2.89 1.82 927 0.67 64.74 70.4
(5+6)+2b 36 0.131 3.27 1.61 819 3.70 64.03 79.7
(5+6+2b)+1b 46.5 0.055 3.45 1.52 777 12.00 -19.55 -6.5
(5+6+2b+1b) 35 0.032 3.56 1.48 753 16.20 -23.61 -14.4
+8b
(5+6+2b+1b+ 47 0.030 3.66 1.43 731 23.59 -42.38 -30.9
8b) +10
(5+6+2b+1 100 0.057 3.87 1.36 691 27.92 -97.91 -74.0
b+8b+10)+
9
The energy saving for options (5+6+2b), compared with the base-case annual energy consumption
of 1 102 kWh, is 277 kWh (25%) for model B. For model C, options (5+6+2b) give a saving of 82
kWh (25%) compared with the base-case annual energy consumption of 327 kWh. If one moves
from the lower NPV to the upper NPV (lower margin on overcosts, downsizing of electrical parts),
the limit between cost-effective improvements and non-cost-effective ones is not altered.
A different approach to the same figures may be done in the spirit of the Kyoto Protocol. Since the
citizens of European countries are ‘buying’ CO2 emission reductions, how much can they obtain
from new RACs at no cost (maintained LCC)? This leads to the search for the level of performance
at which the total marginal NPV is zero. For model B, it is far beyond the optimal 25%, in an
uncertain zone of our study, but is around 50%.16
Sensitivity of selection
NPV calculations for other sectors were also performed, using the following values:
EU electricity price: 0.1095 EURO/kWh
Average operational times:
Sector Hotel Office Shop
Hours/year 751 815 1 026
Results for model B and model C show the same trend as in the residential sector.
The energy savings for options (5+6+2b), compared with the base-case annual energy consumption,
is 408 kWh/year in hotels, 442 kWh/year in offices and 557 kWh/year in shops (about 25%) for
model B. For model C, options (5+6+2b) give a saving of 121 kWh/year in hotels, 131 kWh/year in
offices and 165 kWh/year in shops. Following are the results for model B.
16
By extrapolation of non-cost-effective energy gains, until they balance the profit generated by the cost-effective ones.
67
Table 7 11: NPV analysis results for Model B – other sectors (base case)
HOTEL OFFICE SHOP
Option NPV (EURO) Payback NPV (EURO) Payback NPV (EURO) Payback
(years) (years) (years)
5 144.99 0.16 157.58 0.15 199.10 0.12
(5)+6 83.70 0.53 91.31 0.49 116.40 0.39
(5+6)+2b 94.88 2.91 107.26 2.68 148.08 2.13
(5+6+2b)+1 -7.26 9.45 -2.34 8.71 13.91 6.92
(5+6+2b+1)+8b -16.76 12.76 -14.02 11.76 -4.96 9.34
(5+6+2b+1+8b)+10 -36.07 18.58 -33.54 17.12 -25.18 13.60
(5+6+2b+1+8b+10)+9 -86.56 21.99 -82.01 20.26 -67.00 16.10
The process was extended to two electricity price scenarios, corresponding to two future paths that
this variable may follow. If external effects of energy are included, it means that there is a ‘societal’
cost in kWh a little bit higher that the cost paid by the consumer. In other words electricity is
assumed to become more expensive due to environmental pressure. This effect was studied for a
20% increase, as in the ‘Renewable greenbook’ of the Comission. If competition in the electricity
sector generates significant price decreases and no environmental taxation, it could lead to the other
situation: –20% on electricity cost. This was also computed as a ‘competitive’ scenario.
For model B the lower electricity prices limit the acceptable options to (5+6+2b) in all sectors. This
is the optimum combination. Whereas with 20% higher electricity the optimum combination
changes: for the residential sector it is the same combination of options (5+6+2b), but for the hotel
and office sectors the optimum is (5+6+2b+1) and finally, for shops, with the longest operational
hours, the optimum is (5+6+2b+1+8b). The average optimum among sectors is (5+6+2b+1).
Thus 20% higher electricity prices would permit the addition of option 1 to the optimum
combination.
For model C 20% lower electricity prices leave the optimum combination at (5+6+2b) in all sectors
except the residential sector, where adding option 2b to (5+6) results in a slightly negative NPV. On
the other hand, with 20% higher prices the results are consistent in all sectors: the optimum being
(5+6+2b). Evidently the next possible option, option 9, is quite costly for model C and even with the
20% increase in electricity prices the marginal NPV is quite negative.
68
Summary of economic results
The tendency is very general: the 25% improvement obtained through options (5+6+2b) is always
cost-effective to every user. The further improvement to 36% (the technical potential) is cost-
effective for some users, but not for residential ones. A graphical presentation of the results is given
below with two different sets of financial assumptions for the single-duct unit studied.
The figures are very similar for the two dominant types of air-cooled appliances (splits and single-
ducts). The small remaining part of the market is primarily made up of multi-splits (7%), which
obviously tend to follow the pattern for splits. The study represents 89.5% of the market. Packaged
units (2.2%), which are often manufactured in the US, were subject to an LBL study similar to ours
and submitted to an equivalent efficiency increase. We now only have to consider water-cooled
units (8.3%). It is widely known that it is less costly to increase the heat exchange area with water as
a coolant than with air, while the other options have identical costs and benefits. The 25% cost-
effective performance increase is an underestimation of the potential in that specific segment.
Figure 7.4. Example of improvements on the single-duct appliance
BASE +25% +36% +48%
The use of inverter technology at the costs presently indicated is not cost-effective if only financial
factors are considered, disregarding the effect on comfort. Specifically, it results in a 12% energy
efficiency improvement and leads to a total 48% savings potential when combined with the all other
proven technical options.
69
8. POLICY ACTIONS REALISABLE BY THE YEAR 2010
8.1 Methods for the study of scenarios
The uncertainties of global market trends, relative market share etc are high for RACs because they
are a product just entering into an S-curve in growth. We made very conservative assumptions about
all such factors. Other uncertainties arise from changes in refrigerant regulations. Economic impacts
on company growth patterns have been performed using an American program called USA-LBL.
Special attention was paid to the impact on small and medium-sized enterprises, as well as possible
consequences on employment, the impact on consumers, changes in sales, changes in cooling
systems, and the impact on electricity utilities. Manufacturers were interviewed and economic
calculations made on the basis of their responses. Customer phone interviews were also carried out
in two countries, using a questionnaire formulated by the group members.
Change of refrigerant
It is often stated that the changes in refrigerant needed to comply with the Kyoto protocol should
obviate the need for any improvement in the EER of RACS within that timescale. There is evidence
of the difficulty some manufacturers are finding in complying with the phasing out of R22. But
there is also evidence pointing the other way, not just in the academic world but in the actual
marketplace. For instance, the 26 split models in the Eurovent database already using R407C
instead of R22 display the same performance as the hundreds of other splits (2.43 compared with
2.48).
In any case, a starting point should be the assumption that there will be additional costs as a result of
refrigerant compliance in the coming years, and then it should be possible to evaluate the
relationship between refrigerant and energy efficiency and how that impinges on cost.
First, the nature of the two questions is not the same: increasing energy efficiency as envisaged in
this report pays for itself. There are additional initial expenses to consumers but then benefits accrue
to them later. It is not a real additional expense, given the longer view. Environmental progress may
cause additional expenses to society as a whole. After the initial change, the goal is that the
increased energy efficiency should pay for itself. The present study need only check that the cost-
effective measures investigated here remain cost-effective even with a change in refrigerants.
Secondly, let us note that many changes can occur at the same time in the technical world: new
materials, new refrigerants, new lubricants, new designs, new colours etc. As soon as the
introduction of one specific change can be extended over a few years, it becomes just one of many
small variations and its merits should be studied for themselves – hence our will to have a
progressive introduction of measures, even if they are very cost-effective. Coordination of the
timing of energy measures and environmental measures should be sought.
Thirdly, the essential point: do the changes considered here remain cost-effective with new
refrigerants (i.e. not R22)? To answer this question a simulation was repeated using the most
interesting group of options found in Chapter 7 but using different fluids, in order to discover
whether the manufacturers can expect to obtain similar results by using more or less the same means
as described in this study when the new fluids are compulsory.
For that purpose, the performance of the examined units was calculated using R-410a refrigerant
instead of R-22. Then the performance of the units was calculated still using R-410a but also the
optimal combination (5+6+2b). The simulation was not performed for unit C as the data regarding
its compressor’s performance with the new fluid were not available. Also, it must noted that the
derived results regarding the effect of using R-410a on the units’ performance might not be
70
accurate. Thus, the EER change has been computed only when the technical measures were applied,
as given in the following table:
Table 8.1 effect of new refrigerants on change considered in this report
UNIT EER INCREASE (WITH R-410a FLUID) (%)
A +25
B +29
D +23
Aver. +25
These results show that the percentage increase in EER is exactly the same regardless of the fluid
that is used. Thus, it can be said the proposed improvements remain just as cost-effective if an
alternative refrigerant is used instead of R-22.
Impacts on European companies
Manufacturers were contacted in order to gather data on current production situations and costs, as
well as their experts’ views on expected changes yielded by the introduction of product changes.
Only the most promising technological improvements were submitted for the manufacturers’
consideration.
Questionnaires formulated by the EERAC study partners were posted to manufacturers to obtain
their responses. To ensure the highest possible market coverage, the questionnaires were distributed
and collected with the help of the two manufacturing associations, Eurovent and CECED, who were
fully supportive. As most of the information collected is considered confidential by the
manufacturing industry, data are presented here in an anonymous and aggregated form.
The companies which answered the questionnaire had an overall turnover in 1997 of 721 million
EURO and an overall market share of 61.8%. The breakdown of the average unit production costs is
the following:
Machinery and other costs 12.0
Energy 1.5
Non-energy raw materials 58.0
Labour 13.5
Distribution, marketing 12.0
Royalties, know-how 3.0
TOTAL 100.0
It is important to mention that these figures are the weighted averages of the cost structures of the
companies, which individually vary a lot: this is particularly true for the machinery costs, which
range from 2% to 20%, and the total production costs, which range from 300 to 800 ECU.
The proposed introduction of technological innovations triggers different reactions among the
manufacturers: 3 out of 5 believe that consumers will respond to these changes with indifference, as
energy efficiency considerations do not yet represent a discriminatory factor. Therefore, any
prospective price increase might result in a loss of personal market share, which will cause a partial
or complete cost transfer to prices.
The majority of manufacturers believe that a higher cost will force the market to shrink and will
71
inevitably induce stronger competition from aggressive and technologically advanced industries
overseas. It is worth underlining, however, that one Japan-based manufacturer, in contrast to this
widespread negative perception, asserts that despite a market contraction the company’s image and
competitiveness will be positively affected by an upgrade in efficiency.
Strangely enough, when the manufacturers’ market growth forecasts are broken down by type of
unit, the figures clash somehow with the overall pessimistic view. In the case of technological
innovation, two manufacturers foresee a considerable increase in the number of units produced, two
do not anticipate any substantial change, while only one expects a significant decline in production.
In conclusion, the possible introduction of technological innovations seems to elicit differing
responses from the manufacturers, though they concede that most cost increases will be passed on to
consumers. Consequently, the market will shrink but its structure – on the supply side – should not
undergo any major change unless the dreaded expansion of aggressive competitors from South-East
Asia and Japan occurs.
More detailed and quantitative information was given on forecast cost and structure changes as a
result of technological improvements. Perceptions of the overall effects on companies’ accounts are
hardly equivocal: two manufacturers forecast an increased turnover (5% average), one expects an
unchanged turnover and the other two predict a turnover decrease (ranging from 5% to 20%). On the
other hand gross profits would seem to suffer more negatively: three manufacturers foresee a
decrease (10–15%) while two believe there will be no change.
The effects on employment levels appear to be negative as well and there is a fear that there would
be a downsizing trend: three manufacturers expect lower employment (a decrease of between 5%
and 20%), one manufacturer believes it will remain stable while the other manufacturer forecasts a
rise of between 0 and 5%.
Finally, all manufacturers appear to be sensitive to the role of energy efficiency issues in marketing,
claiming that they are a key ingredient in advertising campaigns, training programmes, price
policies and overall corporate image.
Simulation of the effects of the technological intervention
The USA-LBL model for the appliance industry was used to evaluate the impact of the above
scenario on European manufacturers. The key input parameters included: a consumer discount rate
of 6%; equity cost of capital of 4%; typical manufacturers’ mark-up of 31%; ratio of highest to
lowest mark-up of 2.04; typical firm size as percentage of total shipments of 13%; industry
shipment of 909 090 units; shipments by product class of 227 000; and industry income tax rate of
40%. Industry price elasticity was estimated at between 20% and 28%. The other cost data were
obtained from the questionnaires sent to the manufacturers. These parameters constitute the input
data set for the basic simulation.
The base-case company was considered to be a medium-sized company having, before the
technological changes, 90 000 shipments, an average factory gate price of 581 EURO per unit,
revenues of 50.3 million EURO, a net income of 2.91 million EURO and a return on equity (ROE)
of 10.9%.
With a rather low elasticity and with substantial benefits to consumers, the unit shipments are
predicted to increase by as much as 15% in the combined option case (5+6+2b). This is the design
option with 25% energy savings. Revenues increase accordingly from 7% to 18% (around 9 million
EURO). On the other hand net income and return on equity are much more stable: the former
increasing very slightly and the latter declining about 1%.
For the sensitivity analysis the most important design option (5+6+2b) was used as the reference.
72
The input parameters and costs were modified by +20% and –20% to determine the change in the
output variables of shipments, price, revenue and net income and ROE, always referring to the case
of design option (5+6+2b). The sensitivity analysis was also performed for the base-case and all the
other individual design options; however, the results are not very different.
It is difficult to extrapolate the results obtained for one typical firm to the whole economic sector.
The way the results areextrapolated (assuming that the sector is made up of a number (N) of similar
firms) is consistent with the assumption that all firms are obliged at time t (the year 2000 for
instance) to move their average EER by 25% upwards. This means that not only the average is
changed by 25% but that almost all the equipment is changed by 25%. Practically N is around 10 in
the present situation; the sector thus is represented by 10 times our typical firm.
Figure 8.1. Market transformation as predicted by the LBL model
% Frequency σ = 10 % s = 10 %
+ 25 %
Present Optimal
average EER EER
The policy option considered initially in the LBL modelling is an ideal one, and so more realistic
scenarios and policy options were then used, adding more time and greater flexibility to the
modelling procedure. We know that this adaptation process is highly non-linear, and the impacts of
real scenarios and policy options on manufacturers will in fact be far smaller than are computed
with a linear interpolation, as is done here.
The overall employment picture is highly positive: in the manufacturing and trading sectors there
would be a 10% increase in number of jobs, i.e. around 600 jobs, although some job losses (or non-
creation of new jobs) would occur among utilities. However, employment levels are so different in
utilities than in manufacturing that this would result in a relatively limited alteration to overall
numbers (e.g. 500 jobs created instead of 600).
Methodology of the consumer survey
Consumers were interviewed in two countries, Italy and Spain, which are considered as the primary
sources of demand for the European RAC market. An external company17 specialising in market
research carried out this part of the work in November 1998.
The interviews were based on an ad hoc questionnaire with 21 questions covering the following
main areas of information:
• the degree of satisfaction among consumers with the various performance characteristics of
17
ASM-Analisi e Strategie di Mercato is the Italian subsidiary of the GfK group; GfK is a market research company
active all over Europe.
73
installed air-conditioning systems;
• the factors motivating their purchasing decisions;
• the consumer response to a new ‘improved’ product;
• RAC purchasing potential in the future (short and long-term) in households without an air-
conditioning system.
In the chosen sample we tried to balance the relatively low penetration of RACs in households with
the need to ensure the statistical validity of the results from owners of air-conditioners. Therefore,
samples of 2 000 (Italy) and 3 000 (Spain) households – called the ‘screening sample’ – were
chosen from about 5 000 total contacts.
Through telephone interviews a ‘valid output sample’ of 300 questionnaires per country out of the
total contacts was chosen in such a way to comprise 70% RACs owners (210 households/country)
and 30% households without room air-conditioning (90 households/country). 600 valid
questionnaires were finally collected.
A draft questionnaire was initially validated through 30 pilot interviews in both Italy and Spain (20
households with an RAC and 10 without). These results were used to design the final questionnaire,
which was then submitted to the above-mentioned ‘screening sample’. The duration of the interview
was about 15 minutes for non-owners and 20 minutes for owners of air-conditioners.
Results of the consumer survey
The market penetration of domestic air-conditioners in Italy is approximately 10%. Of the RACs
owned by the interviewed sample, 40% are fixed split and 31% of the single-duct type. On average,
air-conditioning units are used for approximately 2.8 months a year and for 6 hours a day, which
confirms the results obtained by the technical study. The overall functioning performance of
installed units is considered good (73% either ‘very good’ or ‘good’), and other results were also
similar to those found by the technical study: appliance noise (51% ‘very good’ or ‘good’) and time
needed to reach a comfortable temperature (44% ‘very short’ or ‘short’).
Fourteen percent of the Italian owners plan to change their installed air-conditioning system, and
36% of non-owners say they may want to buy one in the next few years. The RAC type they intend
to buy is mostly fixed split or multi-split. The reasons behind this choice are essentially cooling
performance (45%) and a preference for a modern and technologically advanced appliance (21%).
The interest in possible energy savings is quite high, with 75% of the respondents who would
change or buy an air-conditioning system saying they would prefer a new, improved type with
consequent savings on their electricity bill.
The penetration of domestic air-conditioners in Spain is approximately 5%. The most widespread
air-conditioning system is the fixed split (40%), followed by the packaged type (27%). On average,
air-conditioning units are used for approximately 3.2 months a year and for 5 hours a day. Nearly
half (45%) the interviewed sample of owners had installed their RAC between 1995 and 1997. The
overall functioning performance of installed air-conditioners is considered positive (81% ‘very
good’ or ‘good’); and the majority are happy with the noise level (65% ‘very good’ or ‘good’) and
the time needed to reach a comfortable temperature (65% ‘very short’ or ‘short’).
Sixteen percent of the Spanish owners plan to change their present system, and 19% of non- owners
say they may want to buy a new one in the next few years. Consumers who intend to change their
air-conditioner are mostly owners of units that are more than four years old, and the overall
preference is for multi-split and fixed-split systems. The factors influencing their future choice are
different from those of the past, the most important being the wish to buy a modern, technologically
advanced system (28%) and the system’s cooling performance and price (both at 17%). The interest
74
in an improved energy-saving air-conditioner is quite high (63%). The numbers of those saying they
would purchase an appliance that, even if more expensive, would guarantee savings on the
electricity bill increases with the increase in the amount of energy to be saved, but this is a non-
linear trend, starting at a very low level (11.6%) and reaching a maximum of 46%.
8.2 Behavioural changes, controls, comfort conditions and thermal regulations
The behaviour of EU consumers is likely to evolve in the direction of more air-conditioning. In
some ways technical progress can allow a focus on certain improvements – better control, smaller
temperature variations etc. – but this tendency is negligible when set against the overall global
growth in the market for RACs, which is determined by:
• the increasing demand for comfort, which is beyond our scope of control;
• the quality of buildings, whether existing or new (and the influence of thermal regulations);
• the quality of appliances, on which we can have an influence through various policy actions;
• the decisions of those involved (architects, installers etc.).
Variable speed and comfort
Better control and smaller temperature variations generally means innovations in variable power
operation, and this trend will develop in the future. The current ‘inverter’-type models represent but
the first step in that direction and are presently sold more for their extra comfort than for their
energy efficiency. Intermediate options exist in the form of multi-speed ventilators and compressors.
Making inverters, with their lower energy consumption in the field, more attractive to the consumer
would be a clear indication of a policy in favour of progress, bringing both increased comfort and
energy efficiency. There is little evidence, however, from the manufacturers’ associations (that is,
until the revision of EN 814) that even this step is being taken.
Thermal regulations
Present thermal regulations in most countries relate only to space heating. To our knowledge, only
Portugal has developed a specific limit, which is to be applied in all building projects, on the total
cooling energy demand. An additional piece of regulation requires that central air-conditioning,
instead of RACs, is to be used in buildings with more than a 25 kW cooling load. The optimisation
of a building project’s space heating efficiency entails high insulation and low (although hygienic)
air renewal. These tendencies are often in the direction of low air-conditioning consumption but not
always. Furthermore, optimisation of solar energy inputs is completely missing in ‘winter’ building
codes. There is also a large potential for the development of summer energy efficiency in building
codes.
Architects and installers can play an important role in the correct decision-making about air-
conditioning because they advise on the match between the unit and the zone and on the importance
of good performance. Barriers to energy load reduction can be found in many countries in that the
architect in a construction market is paid on the basis of a certain percentage of the air-conditioners
installed, not according to square meters. They are interested in playing ‘safely’ and in siting the
plant to avoid any risk of discomfort. Other types of markets, namely those expressed in terms of
‘energy service’ (cooling a surface at a certain price), generate the opposite trend among the new
operators or facility managers.
Availability of data for subsidiary measures
So it is recommended that municipalities, central administrations etc. of southern European Member
States, as well as energy agencies, consumers’ associations etc., to introduce regulations or at least
75
advisory schemes on the size of systems, operating temperatures and especialy energy consumption
calculations. All the necessary facts for establishing a common rule for the easy calculation of such
energy consumptions have been gathered in the present study. If an ‘installed appliance’ scheme for
RACs is proposed in the countries represented in this report, all the technical elements needed are
available here. The role of maintenance has been specifically indicated as a possibility for acheiving
significant gains.
8.3 Minimum energy efficiency standards
We have demonstrated in chapter 7 that the improvements which presently allow some RAC models
on the European market to have a better EER than others correspond to the magnitude of the gains
in cost-effectiveness for the consumer and affordability for manufacturers (25% of the range
computed in the study). It is not enough just to bring this information to the consumer because of
the conditions in which he (she) makes the decision (and anyway in many cases the consumer is not
the decision-maker).
Given this, there is certainly room for the introduction of minimum efficiency standards. For the
progressive adaptation of such standards, it is necessary to establish some intermediate targets such
that a certain percentage of existing appliances is already over the new limit, this percentage being
sufficiently distributed between countries and manufacturers. We have chosen the average of the
1996 market as a suitable limit for an initial Minimum Energy Efficiency Standard (MEES) that
could be applied in a few years (somewhere between 2000 and 2002).
Figure 8.2. First MEES (for 2000 or 2003)
% Frequency
MEES 2003 Optimal
= average EER
Taking the present market average as a limit, not directly the minimum LCC value, is an
intermediate step such that each producing country and each manufacturer already has satisfactory
appliances available. The average lines defined in chapter 4 and used here show slopes which are
not significant in statistical terms. For the representation of the average it is simpler to use only an
EER value per class as a threshold.
One can expect a recalibration of the European market after this first MEES and a second step can
be introduced, for instance, between 2003 and 2010 at the first level +10 %.
Figure 8.3. Possible second step for MEES (2010 level is 2003 level + 10%).
76
% Frequency
MEES MEES Optimal
2003 2010
EER
Indicating both future steps in MEES is a way of creating a specific R&D effort in increasing
efficiency of RACs, to which national and European funds could contribute.
First possibility: separate MEES per categories
Table 8.2.
MEES on EER First MEES (present Present best value Second MEES (+10%)
average)
MS,A: Multi-split, air-cooled 2.63 3.74 2.89
PA,A: Packaged, air-cooled 2.38 2.97 2.62
PA,W: P, Water-cooled 3.32 5.42 3.55
SD,A: Single-duct, air-cooled 18 3.09 2.28
1.80
SD,W: Single-duct, water-cooled 2.36 3.62 2.60
SP,A: Split, air-cooled 2.48 3.56 2.73
SP,W: Split, water-cooled 2.75 2.88 3.03
The analysis of the consequences is made laterr, after saying a few words about another possibility.
Second possibility: a global MEES with two thresholds (water, air)
Table 8.3
MEES on EER First MEES Category Present best Second MEES
(global av.) average value (+10%)
MS,A: Multi-split, air-cooled 2.48 2.63 3.74 2.89
PA,A: Packaged, air-cooled 2.48 2.38 2.97 2.62
PA,W: Packaged, water-cooled 3.22 3.32 5.42 3.55
SD,A: Single-duct, air-cooled 2.48 2.07 3.09 2.28
SD,W: Single-duct, water-cooled 3.22 2.36 3.62 2.60
18
Due to the experimental uncertainty and to the potential existence of a size effect, the actual average of 2.07 has been
replaced by a lower value which avoids all problems; this change, necessary in the short term due to the ‘youth’
problems of EN 814, will not be necessary for longer-term measures that will take place after the results of the EU
mandate to CEN on testing accuracy.
77
SP,A: Split, air-cooled 2.48 2.48 3.56 2.73
SP,W: Split, water-cooled 3.22 2.75 2.88 3.03
We see immediately that this scheme is more difficult to put in practice because it has a more
marked effect on some segments of the market. We can say that this second possibility corresponds
more to the goal (energy saving) but is difficult to implement by legislation. Only the voluntary
commitment of the industry can solve the internal problems of the industry.
Discussion of consequences
With the two successive MEES levels, the percentages of models already satisfying the standards in
the most important category of RACS (splits) are as follows:
Table 8.4a. Level of performance of production from various countries.
ORIGIN OF PRODUCT Satisfying SPLITS Satisfying SPLITS
(first MEES) (second MEES)
BELGIUM 43.85% 21.93%
FRANCE 50.80% 16.04%
GREECE 62.50% 25.00%
ISRAEL 62.50% 6.25%
ITALY 65.06% 24.52%
JAPAN 61.98% 26.45%
PORTUGAL 100.00% 40.00%
SPAIN 52.38% 19.30%
TAIWAN 92.31% 69.23%
UNITED KINGDOM 22.32% 5.06%
AVERAGE 51.2% 18.8%
All categories of equipment have a good proportion of satisfactory appliances:
Table 8.4b
st
Category Satisfactory for 1 MEES
MS, A 46.8%
PA,A 53.7%
PA,W 52.8%
SD,A 50.0%
SD,W 23.1%
SP,A 51.6%
SP,W 50.0%
ALL 51.2%
To a large extent most manufacturers are already meeting the first MEES:
78
Table 8.5. Level of performance of production from all manufacturers.
Level of Satisfying manufacturers Satisfying manufacturers
compliance
(first MEES) (second MEES)
80-100% 21.4% 0%
60-80% 23.8% 9.5%
40-60% 33.3% 7.1%
20-40% 11.9% 40.0%
0-20% 9.5% 52.4%
International comparison
The comparison has been made with the MEES and targets of goverments in the largest producing
countries. The proposed first and second MEES occur just in the middle of the lines applied by
these competitors. The first MEES is clearly needed to avoid the possibility of a number of South
East Asian countries, which already export a lot to Europe, being tempted to sell here the products
that they can no longer place on their national market due to its low performance.
Figure 8.4. International comparison.
3.2
Split, cooling only, air
2.7 Split, reversible, air
Japanese targets, cooling only
EER
Japanese targets, reversible
Korean targets, all
Korean MEES, all
2.2 China MEES, all
proposal MEES EU 2003
proposal MEES EU 2010
1.7
1.2
0.0 2.4 2.7 3.3 3.7 4.5 5.0 5.5 6.4 7.1 7.7 10.3
Cooling capacity [kW]
8.4 Information and labelling on appliances
Energy labelling of RACs is part of directive 75/92. It is certainly also seen as an important market
transformation tool but cannot create as significant a shift in purchasing behaviour (unless supported
by additional measures) as a strict minimum efficiency standard can. In the many situations in which
the installer decides on the appliance, the label loses part of its effect. However, it can help
manufacturers and consumers go further, encouraging research, promoting very efficient models etc.
79
(people looking for a quality indicator will be sensitive to the ‘A’ indication).
Basic features of the potential system
From a technical perspective, a number of difficult issues in the design of a room air-conditioner
energy label must be met:
(a) What product categorisation structure should be adopted and what maximum cooling
capacity threshold should be applied? How to present the categorisation to the public?
(b) How should comparative energy performance be indicated?
(c) Should only the EER be highlighted; which grading algorithms should be applied if a
gradings approach is adopted?
(d) How should annual energy consumption be treated?
(e) Should operating costs be supplied and if so in what form?
(f) How should other performance aspects such as cooling capacity be treated?
(g) Should part-load performance be addressed?
(h) Should heating performance also be rated for reversible units?
(i) How much information of a given complexity can the purchaser be expected to absorb?
Discussion of these questions follows from previous technical findings.
(a) Product categorisation in four technical groups (SP, MS, PA, SD) with two cooling media
(air/water) gives seven types (one potential type being non-existent, MS water-cooled). The
limit of 12 kW cooling capacity has been retained. The four technical groups should be very
clearly indicated to the public because the performance range is very different and they
shouldn’t put the best of the best category (EER=3.74) and the best of the worst category
(EER=2.97) together in the same grouping. For this reason, we have prepared translations of
the English words that could allow national consumers to understand the difference between
product categories. If the consumer can differentiate them sufficiently, it’s better to make
seven categories of grading. If to the consumer these appliances are all just ‘air-
conditioners’, one single grading should then be applied to them.
(b) The indication to be given on the label: the climatic influence on consumption in kWh is
such that only the EER can be used for grading within each of the seven categories; the
understanding of EER is difficult enough, so a grading system is necessary. The EER value
will be better understood if translated in a A-G grading system, for which the EU consumer
is now ready.
(c) Which basis, EER or kWh? The translation of EER in kWh being about the same for each
type of RAC (with two exceptions) across Europe, the A-G grading can be based on EER or
kWh, so should be based on EER because it’s simpler. The two exceptions just mentionned
are: single-ducts, for which a correction could be introduced for air infiltration, but if they
remain a separate category that is clearly indicated to the consumer, this correction has no
influence on grading; inverters, for which an improved EER compared with standard testing
should be recognised (they may be found in all categories in theory) as soon as more
experimental evidence is available. The EER values used should be of good quality, hence
the certification logo, when available.
(d) Should we abandon indications of consumption? The difference between categories is such
80
in terms of consumed kWh/m2 due to their performance that a conventional figure of
absolute consumption calculated with our results should be also indicated on the label for
people that can understand this indication. It would be very conventional (sizing, climate)
because a better indication of energy consumption could only be national, given the wide
range of equivalent numbers of hours of operation. Detailed kWh indications are
recommended for the potential ‘installed appliance’ note.
(e) Same answer as previous point (with a national and sectoral average price of electricity).
(f) Since cooling capacity is not ambiguous, it can be on the label; translation in number of m2
should remain on the additional note, with national values, because sizing the system is not
an EU responsibility.
(g) Until the results of the mandate M/274 to CEN/CENELEC are available and due to the lack
of technical data provided by manufacturers, it is impossible to take into account part load
performance.
(h) Heating performance is for the time being a secondary aspect of RACs in most European
countries. We have shown also that there is no contradiction between performance for
heating and performance for cooling; so only the indication of ‘reversible’ or ‘cooling only’
need appear on the label.
(i) Other additional information that a customer may want to have is the noise level. Although
not compulsory it is highly recommended that this indication be given in a systematic
manner on the label (but this may be made possible by other means, e.g. voluntary
agreements).
Figure 8.5: General layout of possible label.
T e c h n ic a l g r o u p :
S p lit, M u lti-s p lit, W in d o w ,
S in g le -d u c t
S p lit
A /C
E n e r g y c o n s u m p tio n :
A v e ra g e n b o f h o u rs * (1 0 0 W /m ²) /
EER
In d ic a tiv e e n e rg y c o n s u m p tio n
k W h /(y e a r.m ²)
A c tu a l e n e r g y c o n s u m p tio n w ill d e p e n d o n
h o w th e a p p lia n ce is u s e d a n d o n c lim a te C o o lin g c a p a c ity :
a t T 1 c o n d itio n s
C o o lin g c a p a c ity [k W ]
C o o lin g o n ly / R e v e r s ib le C o o lin g o n ly
A ir c o o le d / W a te r c o o le d A ir c o o le d
H e a tin g m o d e (o r n o t)
C o o lin g m e d ia
A n y c e r ty fy in g lo g o
E u ro v e n t o r o th e r
81
First possibility: definition of gradings with seven separate scales
The average plots of EER/Pcooling show slopes which are not significant in statistical terms. In the
context of the label it is recommended to use only an EER value per class as a threshold. For
determining the levels, some principles should first be defined.
If we assume a parallel introduction in a few years of a MEES corresponding to the average of the
present European market, it is logical to apply a ‘G’ grading to all models under this average line.
The accuracy of 6% of EN 814 and the range of practical values on the market are consistent with
the use of classes in increments of 10%.
Numerous labelling schemes which have been studied by the group support the adoption of a
grading system that is based on equal 10%-wide classes. The following grading system would then
apply:
Table 8.6: Possible gradings of RACs and EER (air-cooled).
Limit (air) EER/EERav Split Multi-split Packaged Single-duct
A starts over 150 3.72 3.95 3.57 3.11
B starts over 140–150 3.47 3.68 3.33 2.90
C starts over 130–140 3.22 3.42 3.09 2.69
D starts over 120–130 2.98 3.16 2.86 2.48
E starts over 110–120 2.73 2.89 2.62 2.28
F starts over 100–10 2.48 2.63 2.38 2.07
G starts below 100 2.48 2.63 2.38 2.07
Table 8.7. Possible gradings of RACs and EER (air-cooled).
Limit (water) EER/EERav Split Packaged Single Duct
A starts over 150 4.13 4.98 3.54
B starts over 140–150 3.85 4.65 3.30
C starts over 130–140 3.58 4.32 3.07
D starts over 120–130 3.30 3.98 2.83
E starts over 110–120 3.03 3.65 2.60
F starts over 100 -110 2.75 3.32 2.36
G starts below 100 2.75 3.32 2.36
The models on the market compare in the following way with this grading:
Table 8.8
Grading (air) EER/EERav Split Multi-split Packaged Single-duct
A 150 0% 0% 0% 0%
B 140–150 0.1% 0% 0% 0%
C 130–140 1.0% 1.3% 0% 0%
D 120–130 3.6% 5.1% 3.7% 1.7%
82
E 110–120 9.1% 12.0% 11.0% 31.0%
F 100 -110 34.2% 26.6% 36.6% 15.5%
G 100 52.0% 55.0% 48.8% 51.7%
Table 8.9
Grading (water) EER/EERav Split Packaged Single-duct
A 150 0% 1.4% 0%
B 140–150 0% 1.4% 7.7%
C 130–140 0% 0% 0%
D 120–130 0% 7.6% 0%
E 110–120 0% 13.9% 15.4%
F 100 -110 16.7% 26.4% 0%
G 100 83.3% 49.3% 76.9%
There is a sufficient representativity in most categories. Category D corresponds to the minimum
LCC determined previously. If the gains due to variable speed are considered by the revised EN 814
following the EU mandate, they bring one or two grades upwards and we have ‘A’ models in all
categories.
Second possibility: single grading of all appliances
If we start by answering that we want to have only one category called ‘air-conditioners’, the
grading scale will be different:
Table 8.10
Limit (air) EER/EERav Limit % satisfying
A starts over 150 3.72 1.1
B starts over 140–150 3.47 1.3
C starts over 130–140 3.22 2.2
D starts over 120–130 2.98 3.5
E starts over 110–120 2.73 8.6
F starts over 100–110 2.48 27.1
G starts below 100 2.48 56.2
Although easier to understand for the consumer (‘an ‘A’ appliance is an ‘A’ appliance’), this single
grading is not very easy for the legislator because it requires different efforts from different
manufacturers to reach a ‘A’ and results in unbalanced information on water-cooled appliances.
This last problem could be avoided by limiting the benefit of European labelling to air-cooled
appliances or by adding a warning on water use on water-cooled appliances. The first problem could
be solved by a voluntary agreement among manufacturers to be ready to enforce together the new
rule and develop research.
Additional measures and impacts
The impacts of the proposed labelling
83
- on manufacturers
- on consumers
- on utilities
- on the environmental balance of the EU
- on CO2 emissions
- on the energy situation
are difficult to estimate as they are all effected by the vagaries of human behaviour. One cannot
expect that all consumers will change their choice from a present G/H-graded appliance to a D-class
one. The label will be most effective in helping to restructure the market at the time of
implementation of the first MEES.
8.5 Voluntary agreement of manufacturers, performance certification
Voluntary agreements should allow a low cost acceleration of the market transformation process.
Any negociation of a voluntary agreement is likely to use other policy measures as a reference, for
example either the concept of a market weighted EER by a certain year or directly through an
industry enforced MEES. The weighted performance approach allows manufacturers the possibility
of sharing the burdens of research, retooling, etc between them, although a minimum performance
level approach is easier to enforce. It's better to combine both approaches.
A voluntary agreement in principle may also aid market transformation because it can use other
information circuits. For instance when the final users of the equipment are not the decision makers,
it is more efficient that the manufacturers oblige fitters to install products having a certain
performance level, than to expect that they would voluntarily select a more expensive but well
graded appliance.
Possible bases for voluntary agreements in terms of MEES
Industry can reach a compromise with the Commission by removing from sale inefficient
equipment, as was the case with CECED’s voluntary agreement for washing machines. If one
wishes to achieve the same results as the proposed second level MEES by a voluntary agreement,
the class F and G appliances (as defined in the discussion on labelling) should be removed from the
market. If one wants to achieve the fuller objective of bringing the market to its techno-economic
optimum through such an agreement, then the class D and E appliances should also be removed
from sale. In the situation where the industry is strong enough to enforce the more challenging
agreement one would also propose that a single grading of all appliances is chosen for energy
labelling because of the greater clarity for the consumer. The strength of the industry commitment
would help to solve the internal problems of technical change and allow this better solution.
Possible bases for voluntary agreements in terms of WEER
In terms of average performance, if we want to achieve by a voluntary agreement the economic
optimum (+25%) in the installed stock, the WEER (Weighted EER) at market level should change
10 to 13 years before. The sooner the better; however, as it can be seen that even if the voluntary
agreement results in an immediate 100% uptake of the best technology, the target for the stock
cannot be reached before 2010. As a result three scanarios have been made, having targets of
bringing the market to its optimum point by 2010, 2015 or 2020 respectively. Check points at 2005,
2010, etc. would allow the Comission to monitor the progress made towards the optimum.
Table 8.11 Weighted EER corresponding to achieving the techno-economic optimum target by 2020 where
84
the market shares of the various RAC types used for overall averaging are their shares in 1996.
Year 1996 2000 Check point Check point Check point Target 2020
2010 2015
(=1996) 2005
Splits (air) 2.48 2.48 2.79 3.10 3.10 3.10
Splits (water) 2.75 2.75 3.10 3.44 3.44 3.44
Multi-splits 2.63 2.63 2.96 3.29 3.29 3.29
Packaged (water) 3.32 3.32 3.74 4.15 4.15 4.15
Packaged (air) 2.38 2.38 2.68 2.98 2.98 2.98
Single ducts (air) 2.07 2.07 2.33 2.59 2.59 2.59
Single ducts (water) 2.36 2.36 2.66 2.95 2.95 2.95
All RACs 2.44 2.44 2.75 3.06 3.06 3.06
Table 8.12 Weighted EER corresponding to achieving the techno-economic optimum target by 2015 where
the market shares of the various RAC types used for overall averaging are their shares in 1996.
Year 1996 Check point Check point Check point Target 2015
2010
2000 2005
Splits (air) 2.48 2.79 3.10 3.10 3.10
Splits (water) 2.75 3.10 3.44 3.44 3.44
Multi-splits 2.63 2.96 3.29 3.29 3.29
Packaged (water) 3.32 3.74 4.15 4.15 4.15
Packaged (air) 2.38 2.68 2.98 2.98 2.98
Single ducts (air) 2.07 2.33 2.59 2.59 2.59
Single ducts (water) 2.36 2.66 2.95 2.95 2.95
All RACs 2.44 2.75 3.06 3.06 3.06
Table 8.13 Weighted EER corresponding to achieving the techno-economic optimum target by 2010 where
the market shares of the various RAC types used for overall averaging are their shares in 1996.
Year 1996 Check point Check point Target 2010
2000 2005
Splits (air) 2.48 3.10 3.10 3.10
Splits (water) 2.75 3.44 3.44 3.44
Multi-splits 2.63 3.29 3.29 3.29
Packaged (water) 3.32 4.15 4.15 4.15
Packaged (air) 2.38 2.98 2.98 2.98
Single ducts (air) 2.07 2.59 2.59 2.59
Single ducts (water) 2.36 2.95 2.95 2.95
All RACs 2.44 3.06 3.06 3.06
One can judge the depth of a voluntary agreement proposed by the industry by comparing the
WEER levels proposed at certain years (typically 2000 and 2005) with the WEERs of the Tables
8.10 to 8.12. Voluntary agreement may cover also some additional aspects like advice on sizing,
offering to only supply high performance equipment to installers, maintainance offers, indications of
acoustic power on labels, etc…
85
Performance certification
Since a certification system addressing both the EER and noise is already in place for most of the
markets, there are very good prospects of having reliable performance declarations by
manufacturers. One large manufacturers’ association of Room Air Conditioners (Eurovent-
Cecomaf) has built a strong commitment among its members to energy efficiency and has installed a
structure that checks the performance of appliances on the market and thus increases positive
competition on this aspect between its members. Between 80% and 90% of the EU RAC sales are
from its members. The other association (CECED) is strongly represented among the remaining
share of EU sales although there are multiple instances of manufacturers belonging in some way to
both organisations.
Eurovent’s figures for appliances on the market have proved very useful for the present study and
for industry. CECED has a large experience in and commitment to energy efficiency and is ready to
help in the market segment where it is representated; its data have been very useful in the segment
not covered by Eurovent. In consequence it is clear that there are organisations with whom a
voluntary agreement can be negociated and signed. The framework given by CECED under the title
‘E&E’ is an example of this willingness to sign and enforce a voluntary agreement.
It has been stressed in the study that self reported values given by manufacturers can suffer from
larger experimental errors in one market segment (Single Ducts). This segment, which has taken a
growing share of EU RAC sales, is precisely the one which is not certified in Eurovent Certification,
and where measurement consistency is not realised by an adhoc business association. As a
consequence any voluntary agreement with manufacturers should in our view include:
(a) the introduction by the industry of a declared performance checking mechanism for single
duct RACs, by the same or another business association, to be created or extended on a
similar basis
(b) the consolidation by Eurovent-Certification of its performance certification scheme in the
three other segments.
Furthermore we think that the level of reliability obtained in the three RAC segments (splits, mult-
splits, and single-packaged) has to be gradually realised for the “missing” segment (single ducts).
The standard certification system in place under Eurovent Certification is based on systematic
testing of a statistical percentage of models not on “challenge testing”. A “ challenge test ” system
(checking of one manufacturer’s product at the request of another one) has been in place for a
number of years through ARI in the United States and for five years as an option at Eurovent
Certification. The level of hostility between companies created by a potential “challenge test”
process is such that people may not report such cases when they suspect them. A statistical process
is more economic and systematic, even the industry doesn’t select exactly the same process as
Eurovent Certification did.
8.6 – Actions by national bodies
All electric utilities can play an important role in supporting labelling schemes, although they face
markets, which differ significantly according to legal and climatic conditions. EU Member States
can enhance any actions initiated at the European level.
Northern Member States Utilities
For electric utilities in northern EU countries, air conditioning has emerged as an increasing market
for which they are unlikely to need to make additional supply-side capacity (kW) investments.
Provided that they can sell energy (kWh) at the correct price, air conditioners are likely to make a
86
positive impact on their business while enabling consumers to enjoy increasing comfort.
Nonetheless, these utilities may be bound to promote energy efficiency as part of their “public
service” obligations, if any.
Southern Member States Utilities
For electric utilities in southern EU countries, air conditioning is a market that could pose a risk in
some regions. In Summer peaking zones, additional supply side capacity investment in generation,
transmission and distribution may be needed asa result of growing air conditioning loads. To know
the extent of the phenomenon, one has to investigate the contribtion of RAC usage to the system
peak.
Table 8.14 Estimated contribution of RACs to southern utilities system peak power demand (data was
unavailable for Greece)
1996
Winter peak Summer Peak
Hour GW Hour GW
France 7 p.m. 63 11 a.m. 49
Italy 6 p.m. 42 11 a.m. 40.5
Portugal 11 a.m. 4.6 11 a.m. 4.2
Spain 7 p.m. 24 1 p.m. 22
2020 (estimation)
Winter peak Summer Peak
Hour GW Hour GW
France 7 p.m. 105 11 a.m. 82
Italy 6 p.m. 70 11 a.m. 68
Portugal 11 a.m. 8 11 a.m. 7
Spain 7 p.m. 40 1 p.m. 37
2020 (estimation)
Contribution of RACs to system peak
Hour GW
France 11 a.m. 0.0
Italy 11 a.m. 3.5
Portugal 11 a.m. 0.3
Spain 1 p.m. 2.5
When the tariffs (or market prices) don’t cover the additional supply-side capacity investment, air
conditioning becomes a risk for the utility because there is a disparity between the revenues and the
costs. Without entering into detail, Table 8.15 gives the estimated cost of the supply-side
investment, assuming a uniform capacity cost of 550 Euro/kW, needed in four Mediterranean
countries to accommodate the growth of peak load due to room air conditioners.
Table 8.15 Estimated utility capacity investments to accommodate RAC peak power demand
Additional Capacity (MW) Additional Investment
Greece 1500 825 M-Euro
Italy 3500 1925 M-Euro
Portugal 300 165 M-Euro
87
Spain 2500 1375 M-Euro
TOTAL 7800 4290 M-Euro
If all users were to select the most cost-effective RACs it would save considerable investment costs
for these utilities. Such peak power investment costs are more and more difficult to recover from the
consumers and the regulators. If an economic imablance of 10 % occurs in this recovery process, it
represents a loss of 430 Meuros, which poses a serious risk to southern utilities. When dividing by
the number of appliances sold every year (2 million during 10 years), this gives an order of
magnitude of what some utilities could invest to avoid this load: ~21 Euro per sold appliance.
Reversibility: the new option for utilities
There is a third group of electric utilities, including EdF in France, which are not summer peaking
and support the development of air conditioning. In a competitive space heating market, EdF has
been looking for adavantages to encourage electric space heating. EdF discovered that since summer
comfort was something valued by customers, reversible air conditioners were a market opportunity.
The study members think other utilities will follow, namely because space heating equipment is
seldom available in the south and as soon as a RAC is available, its use in winter becomes possible.
This goes beyond the scope of this study but interesting information has been gathered in annex 1 of
the present report. On the one hand reversibility may replace less efficient electric heating sources
but on the other hand it may be less efficient than non-electric alternatives or lead to increased use
of space heating in Southern countries. All this needs to be investigated.
Actions by Member States governments
Is there room for national measures to promote higher efficiency RACs, for European measures or
for a combination of both? Since there is a stronger climatic relationship driving RAC usage that for
many other appliances, we can expect stronger national measures (to compliment European
measures) than for some other appliances. It might be appropriate to optimise RAC performance at
the national level using the tools developed in the present study.
The potential for building codes to reduce RAC related power demand if extended to address
summertime building thermal performance in Mediterranean countries has already been mentioned.
In addition to any EU labelling scheme, there is scope for national “information schemes”, namely
those addressing optimum sizing ratios and equivalent number of hours of use for energy
consumption calculations, two quantities that are better defined at the national or local level.
National authorities are in a good potition to initiate training programs and/or certification schemes
for installers.
At the same time, policy measures at the EU level are necessary to give a coherent basis for national
actions: defining objective, reliable and verifiavble efficiency categories so that these cannot appear
as barriers to the internal market. Promoting energy efficiency where it is cost effective doesn’t
necessarily mean that countries with low average cooling needs shouldn’t implement given
measures. It is in the nature of air conditioning that households will have differing same needs from
country to country. If the “average” household needs 200 hours of air conditioning per year in a
given country, only households exposed to extreme climatic conditions (poor insulation, orientation
to sun, lofts, …) will install air conditioning but they are likely to use it more than “average”
households would do, for instance 500 h instead of 200 h. So the actual cost effectiveness of the
measures proposed may be less different from one country to the other than might otherwise be
thought. This is especially true when considering that the proportion of the RAC stock used in
commercial, hotel and other non-residential sectors will usually be far greater in the northern EU
88
states.
These considerations could lead to the development of national RAC installation rules, as already
described in some previous EU guides on “installed appliances”. Each Member State could apply
the common rules with its own climatic values. All the relevant values needed to develop policies at
the national level are available in this study. One potential example of national action is that it
would be easier to prohibit the use of drinkable water in water-cooled RACs, a costly and
environmentally detrimental way of rejecting the heat, at the national level than at the EU level.
A specific potential exists at the national level for the improvement of maintainance schemes.
Terms of reference could be defined with the national installers (and retailers) associations. The low
level (or complete lack) of maintainance presently evident in all countries has been taken into
account when computing energy consumption with a certain level of fouling. Significant energy
savings may be expected from a higher level of awareness of maintainance effects. How much of
the 20% drop in RAC efficiency due to fouling could be avoided by any given maintainance scheme
and for a given cost is an interesting subject for national evaluation, prior to or independently of any
common European measure.
8.7 – Summary of possible actions
A "first step" minimum efficiency standard is recommendable for immediate enforcement between
years 2000 and 2003. The proposed level is the present European average RAC performance, which
is some 25% below the cost effective optimum. Its impacts, obtained by a detailed stock simulation
are given in Table 8.16
For labelling, four separate categories can be used for single-packageds, splits, multi-splits and
finally single ducts. The EER measured according to EN 814 (at the T1 conditions) is generally
representative of in situ energy consumption performance. Since the ‘unit’ EER value is not
representative of the energy consumption for water-cooled systems, specific limits should be applied
to their apparent EER. This eventually leads us to 7 RAC categories. Single ducts have an in situ
performance that is different from other RAC types, but is in proportion such that -with the present
market share- it doesn't seem necessary to apply a corrective term to their nominal EER to transform
it into an indicator of true energy consumption. On the other hand, the positive impact of variable
speed technologies, complete 'inverters', or simply multispeed drives, that can be applied to any
class of equipment deserves a positive correction in the rated EER; this bonus should however be
delayed until the revision of EN814 has provided an objective means of quantifying the bonus.
The impact of the labelling scheme may increase after the first step and prepare the market for the
second level of MEES. However, the impacts of the first MEES and of labelling are not enough to
make the market reach the economic optimum. Stronger actions, such as the second MEES and the
involvement of manufacturers, national agencies and utilities will be necessary to bring the market
closer the optimum energy efficiency level somewhere between 2010 and 2020.
If voluntary agreements with manufacturers are used as a way to obtain better EER levels, a
certification scheme similar to the one existing for splits, multi splits and single-packaged RACs
should be introduced by the manufacturers themselves for single ducts. Targets for all segments
have been defined (over the second MEES) level, on the way to the economic optimum. The
removal from the market of class F and G appliances can be accompanied by monitoring of the
WEER, (sales-Weighted EER), for which check points have been established.
All electric utilities and agencies can have an important role. Northern utilities can promote high
efficiency reversible units; see annex 1. Southern utilities could support the European scheme
because of the savings it will induce in their investments. Also, in addition to Europe-wide
89
measures, southern countries could launch national information schemes on optimal sizing ratios
(kW cool/m2), energy consumption indications (under the form of an equivalent number of hours of
operation), advice on solar control and other ways to lower cooling loads can be effective to lower
RAC demand.
Aside from the +25% efficiency goal, what defines the target is the date at which this optimum is
reached. The energy consumption and environmental impacts associated with various policy targets
have been estimated using the market stock model for the year 2010, the deadline for the
satisfaction of the Kyoto protocol targets and for the year 2020 and are summarised in Table 8.16.
The average RAC life expectancy (10-13 years) and hence replacement rate is such that reaching the
target in 2010 or 2015 requires almost immediate action, while targetting year 2020 to reach today’s
economic optimum would allows progressive measures.
Table 8.16 Summary of targets and scenarios
Scenario Policy measures TWh TWh CO2 CO2 Annual Annual Avoided
saved saved saved in saved in gain in consum- risk in
(labelling is an per year per year 2010 - 2020 Manuf. er 2020
accompanying in 2010 in 2020 % of M-tonnes Revenue benefit (southern
measure in all cases) 19
(M-Euro (M-Euro utilities)
1990
in 2010) in 2010)
BAU No measure - - - - - - -
"First Step" 1st level MEES in 0.6 1.6 60% 0.7 12 72 88
2003 +labelling
"Target MEES in 2003 and 2.8 8.6 120% 3.4 57 336 411
2020" 2007 +labelling
"Target MEES in 2000 and 3.3 9.4 100% 3.7 67 396 485
2015" 2005 +labelling
"Target 2nd level MEES in 4.4 10.3 140% 4 90 528 646
2010" 2000 +labelling
High energy savings are possible, as well as significant CO2 emissions reductions, at no (in fact
negative) cost, such that all parties (manufacturers, consumers and utilities) would find a benefit in
the marketing of efficient RACs. At the point of zero cost (i.e. for higher energy efficiencies that
give the same consumer Life Cycle Cost as present) the CO2 emissions reductions would be twice
as high as those at the economic optimum, although the results are less certain for very high
efficiency levels. Additional research is required to improve confidence in the analysis at higher
efficiency levels.
19
Around 1 Mt CO2 emitted in 1990 due to RAC on a total of fuel related emissions of 3000 MtCO2; the percentage is
expressed in the terms of the Kyoto protocol, i.e. related to the total 1990 emissions (like the 8% target), but only the
RAC emissions. In terms of the total emission the values are between 0,02 and 0,07%.
90
REFERENCES
General references
EUROVENT, 1996, 1997, 1998: Eurovent Directory of Certified products
TNO, 1997: Energy labelling of room air Conditioners, Interim report for DGXVII
TNO, 1998 Common final report to Contracts no. XVII/4.1031/Z/95-055 "Energy labelling of room air
Conditioners" and no. XVII/4.1031/Z/96-024 "Energy labelling of domestic air to air heat pumps", under the EC -
SAVE programme, also financially supported by: NOVEM, the Dutch organisation for energy and environment.and
NUTEK, the Swedish national board for technical development
International standards
ISO 5151:1994 ‘Non-ducted air conditioners and heat pumps – Testing and rating for
performance’
ISO 13253:1995 ‘Ducted air-conditioners and air-to-air heat pumps – Testing and rating for
performance’
European standards
The European standard EN 814-1, 2, 3 provides the basis for RAC performance results used in this
study (and EN-255-1, 2, 3, 4 for heating performance).
EN 814-1 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Cooling mode – Part 1: Terms, definitions and designation’
EN 814-2 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Cooling mode – Part 2: Testing and requirements for marking’
EN 814-3 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Cooling mode – Part 3: Requirements’
EN 255-1 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Heating mode – Part 1: Terms, definitions and designation’
EN 255-2 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Heating mode – Part 2: Testing and requirements for marking (space heating)
EN 255-3 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Heating mode – Part 3: Testing and requirements for marking’ (space heating
and water heating)
EN 255-4 (1997) ‘Air conditioners and heat pumps with electrically driven compressors –
Heating mode – Part 4: Requirements’ (space heating and water heating)
Other standards on RAC security and sound level
IEC 60335-240(1995)‘Safety of household and similar electrical appliances – Part 2: Particular
requirements for electrical heat pumps, air-conditioners and dehumidifiers’
ISO/FDIS 13261-1 ‘Sound power rating of air-conditioning and air-source heat pump equipment
– Part 1: Non-ducted’
ISO/FDIS 13261-2 ‘Sound power rating of air-conditioning and air-source heat pump equipment
– Part 2: Non-ducted’
91
Literature review on energy savings due to RAC inverter technology
Reference Studied mode Power Energy savings
KW
[SENS85] Experimentation heating 4.4 29 %
[MILL88] Experimentation cooling/heati 8.8 15 - 20 %
ng
[MACA88] Simulation heating 15 %
[HORI85] Experimentation cooling/heati 3 15 %
ng
[TORI87] Experimentation heating 6.9 7-15 %***
[SENS89] Simulation cooling 14.5 30 %
[SHIM85] Simulation heating 6.3 20 - 40 %
[PARK88] Experimentation cooling 3-5 17 - 18 %***
[LBNL96] cooling 10 % *****
***** = the report states that ‘a conservative value’ has been taken into account
*** = gains in terms of EER (or COP) and not on energy consumption
[HORI85] 'Seasonal Efficiencies of Residential Heat Pump Air Conditioners With Inverters-Driven
Compressors', Hori M., Akamine I., Sakai T., ASHRAE Transactions 1985, pp. 1585-1595,
1985
[LBNL96] 'Potential impact of alternative efficiency levels for room air', LBNL, prepared for the US
DOE, 1996
[MACA88] 'Optimal comfort control for variable-speed heat pumps', MacArthur J., Grald E., ASHRAE
Transactions 1988 part 2, pp. 1283-1297, 1985
[MILL88] 'Laboratory Examination and Seasonal Analyses of the Dynamis Losses for a Continuously
Variable-Speed Heat Pump', Miller W.A., ASHRAE Transactions 1988, pp.1246-1268, 1988
[PARK88] 'The Scroll Effect', Parkes R., Refrigeration, Air Conditioning and Heat Recovery, pp.38-43,
avril 1998
[SENS85] 'Annual Energy-Saving Effect of Capacity-Modulated Air Conditioner Equipped with
Inverter-Driven Scroll Compressor', Senshu T., Arai A., Oguni K., Harada F., ASHRAE
Transactions 1985 part2B, pp. 1569-1584, 1985
[SENS89] 'Research and development on packaged heat pump air-conditioners in Japan', Sensh T.,
Terada H., 1988
[SHIM85] 'Inverter Control Systems in the Residential Heat Pump Air Conditioner', Shimma Y.,
Tateuchi T., Sugiura H, ASHRAE Transaction, pp. 1541-1553, 1985
[TORI86] 'Prediction Using Equivalent Ratio in Estimating of the Performance of Heat Pump Air
Conditioner', Torikoshi K., Uemura S., Yajima R., Fujiwara M., I.I.F. - I.I.R. - Commissions
B1, B2, E1, E2, pp.233-240, 1986
92
APPENDIX 1- Information on reversible air conditioners (EdF)
93
Testing standards, terminology
Terminology
This terminology is in accordance with the European standard EN255-1. This standard specifies the definitions of the air
condensing and water condensing air conditioners and air to air, water to air, air to water and water to water liquid
chilling packaged and heat pumps.
Heat pumps: Appliance which takes heat at a certain temperature and restores it at a higher temperature.
NOTE: When the function of the heat pump is to supply heat (to heat rooms or water for example), it is said to operate
in heating mode while when its function is to take heat away (as cooling rooms for example), it is said to operate in
cooling mode.
Heat recovery: Use, by means of an additional heat exchanger, of the available heat provided by an appliance which is
mainly controlled in cooling mode.
Inside heat exchanger: Heat exchanger designed to transfer heat to, or away from, the inside of a building or inside hot
water system (sanitary water for example).
NOTE: In the case of an air conditioning unit operating in cooling mode, this may be the evaporator, see EN 814-1.
Outside heat exchanger: Heat exchanger designed to take heat from, or transfer heat to, the outside environment, or
any other available heat source.
NOTE: In the case of an air conditioning unit operating in cooling mode, this may be the condenser, see EN 814-1.
Calorific energy (QH): In heating mode, quantity of useful heat transfered by the appliance to the heat transfer fluid
(see 3.14), over a defined interval of time.
NOTE: If the heat is taken from the inside exchanger for defrosting, this is appropriately taken into account.
Calorific power (PH): Calorific energy divided by the defined time interval.
Effective absorbed power (PE): Mean electrical power absorbed by the appliance over the defined time interval, and
comprising:
− the power absorbed by the compressor operation, and all power absorbed by defrosting;
− the power absorbed by all the appliance's control and safety devices;
− the share of the power of devices (as fans, pumps for example) ensuring the circulation of the heat transfer fluids (see
3.14) inside the appliance.
Total absorbed power (PT): Power absorbed by all the components included in the appliance when delivered.
Performance coefficient (COP): Ratio between the calorific power and the effective power absorbed by the appliance.
Operating range: Range indicated by the manufacturer and bounded by the appliance's upper and lower operating
limits (for example: temperature, relative humidity, voltage) within which the appliance is considered to be operational
and provides specified features.
Defrosting state: In heating mode, the appliance's state corresponding to a modification or an inversion of operation
with a view to defrosting the outside heat exchanger.
Defrosting duration (tD): Duration of the appliance is in defrosting condition.
Operating cycle with defrosting: Compressor operating time between two defrosting operations, including the
defrosting duration.
Heat transfer fluid: Liquid or gas (usually water or air) used to transfer heat to or from the appliance.
94
Nominal conditions: Standardized conditions used to determine the appliance's characteristic magnitudes, in particular
calorific power, absorbed power, performance coefficient.
Acoustic power level (LW): Ten times the decimal logarithm of the ratio between an acoustic power and the reference
acoustic power, expressed in decibel. The reference acoustic power is 1 pW (10-12 W).
Testing standards
International standards
ISO 5151 (1994): « Non ducted air conditioners and heat pumps. Testing and rating for performance »
ISO 13253 (1995): « Ducted air conditioners and air to air heat pumps. Testing and rating for performance »
ISO/FDIS 13256-1: « Water source heat pumps. Testing and rating for performance. Part 1: Water to air and brine to air
heat pumps »
ISO/FDIS 13256-2: « Water source heat pumps. Testing and rating for performance. Part 2: Water to water and brine to
water heat pumps »
ANSI-ASHRAE 58 (1986): « Rating Room Air Conditioner and Packaged Terminal Air Conditioner. Heating Capacity,
Method of Testing »
ANSI-ASHRAE 116 (1995): Seasonal Efficiency of Unitary Air Conditioners and Heat Pumps: Methods of Testing »
European standards
EN 255-1 (1997): Air conditionners, liquid chilling packages and heat pumps with electrically driven compressors -
Heating mode - Part 1: Terms, definitions and designations.
EN 255-2 (1997): Air conditionners, liquid chilling packages and heat pumps with electrically driven compressors -
Heating mode - Part 2: Testing and requirements for marking for space heating units.
EN 255-3 (1997): Air conditionners, liquid chilling packages and heat pumps with electrically driven compressors -
Heating mode - Part 3: Testing and requirements for marking for sanitary hot water units.
EN 255-4 (1997): Air conditionners, liquid chilling packages and heat pumps with electrically driven compressors -
Heating mode - Part 4: Requirements for space heating and sanitary hot water units.
Description of standard EN 855-2
The European Standard EN 855 (1,2,3 and 4) specifies the terms, definitions and methods for the ratio and performance
of air and water cooled air conditionners, air/air and water/air heat pumps with electrically driven compressors.
Rating test conditions
Test conditions (T1) (T2) (T3) (T4)
Outside air/Water With defrost A7(6)/W50 A2(1,5)/W50 A15(12)/W50 A-7(-8)/W50*
Without defrost A7(6)/W50 A15(12)/W50 A7(6)/W35**
Exhaust air/Water A20(12)/W50 A20(12)/W35
Water/Water W10/W50 W10/W35 W15/W50
Water/Water B0/W50 B0/W35 B-5/W50
Outside air/Recycling air With defrost A7(6)/A20(12) A2(1,5)/A20(12) A-7(-8)/A20(12)
Without defrost A7(6)/A20(12) A15(12)/A20(12)
Exhaust air/recycling air A20(12)/A20(12)
Exhaust air/Air A20(12)/A7(6)
Outside Water/Recycling air W10/A20(12) W15/A20(12)
Internal Water/Recycling air W20/A20(12)
Note 1: All air and water temperatures are inlet temperatures in °C.
Note 2: All air temperatures in parentheses are wet bulb temperatures in °C.
Note 3: All tests are carried out with nominal flow rates indicated by the manufacturer in m3/s. Where non nominal flow
is indicated by the manufacturer, and only a range of flow rates is given, tests shall be carried out at the minimum value.
95
Note 4: Permissible external pressure difference at the evaporator and condenser shall be indicated by the manufacturer
in Pa for appliances with duct connection and for those discharging into double floor, double ceilling and double
ceilling. If the fan is not included, the internal pressure difference shall be indicated instead.
*
If (T4) is not possible, take A2(1,5)/W50
**
If (T3) is not possible, take A10(8)/W35
Basic principles of performance test
Measurement of the power supplied without defrosting
Stabilized operating condition for water/water and water/air appliances
This operating condition is considered to be reached and maintained when all the measured quantities remain constant
without the set values having to be modified. Periodic fluctuations in the measured magnitudes caused by the operation
of regulation and control means are permissible, provided the mean value of these fluctuations does not exceed the
permissible differences given in Table 4.
Stabilized operating condition for other appliances
To ensure that no defrosting occurs, this operating condition must be reached 2 hours before measurements begin. This
operating condition should be considered to be reached and maintained when all the measured magnitudes remain
constant without the set values having to be modified. Periodic fluctuations in the measured magnitudes caused by the
operation of regulation and control means are permitted, provided the mean value of these fluctuations does not exceed
the permissible differences given in Table 4.
If it is not possible to hold the stabilized condition over a period of 2 h, the power must be measured as per 4.5.3 for
defrosting appliances.
Measurement of the calorific power
To measure the power, significant magnitudes must be continuously recorded. In the case of recorders operating on a
cyclic basis, the sequence must be adjusted so that a complete record is performed every 2 min at least. The power must
be measured under stabilized conditions. The measurement duration must not be less than 30 min.
Measurement of the power supplied with defrosting
The stabilized operating condition defined in 4.5.2.2 may not be reached under certain test conditions due to the outside
heat exchanger. In this case, as the operating conditions vary constantly due to the frosting up of the evaporator, all the
essential values must be recorded and a corresponding mean value must be determined. In the case of automatic data
acquisition programs, all the measured values must be recorded and printed at sufficiently short intervals. After a
duration of 2 h plus a defrosting cycle, the measurement must be carried out over a whole number of cycles, without
being less than 2 h or greater than 24 h.
The initial period of 2 h must not be added to the 1 indicated in 4.5.2.2.
With the flow rate-volume for the heat exchanger constant, and the appliance switched on, the test conditions described
in 4.5.1 must be adjusted. The resulting inlet temperature at the inside exchanger must be held constant during the
heating phase.
A change in the outlet temperature of the inside heat exchanger caused by evaporator defrosting is thereafter
permissible.
During the defrosting phase, a difference of ± 5 K in the inlet temperature of the outside exchanger is permissible over a
period of up to 3 min, and then a difference of ± 2 K over the next 3 min. During the rest of defrosting, a difference of ±
1 K in the inlet temperature of the outside exchanger and inside exchanger is permissible. When the appliance returns to
heating phase, the differences stipulated above apply.
The increase in the permissible difference in the inlet temperature to ± 5 K takes account of the regulation possibilities
of the test installation in a non stabilized operating condition. If the heat pump trips due to the overrun, the test must be
determined using the incorporated defrosting control. The pulses of this control correspond to the starting and stopping
96
points of the defrosting and heating phases.
In the case of connected assemblies or appliances, a visual check must be carried out during the test to establish whether
water flows or not from different holes than the ones provided for this purpose.
Measurement permissible deviations
The heating capacity shall be determined within a maximum uncertainty of 5 % independent of individual uncertainties
measurement, including the uncertainty on refrigerant properties. The electrical power shall be measured within 1 %
uncertainty. The uncertainty on the COP is 6 %.
Permissible differences from the set values:
Measured magnitude Permissible differences from the Permissible differences from the
arithmetic mean of the values individual measured values relative
relative to the set values to the set values
Water or water containing glycol
− inlet temperature ± 0.2 K ± 0.5 K
− outlet temperature ± 0.3 K ± 0.6 K
− flow rate (volume) ±2% ±5%
− static pressure difference - ± 10 %
Air
− inlet temperature (dry, wet) ± 0.3 K ±1K
− flow rate (volume) ±2% ±5%
− static pressure difference - ± 10 %
Voltage ±4% ±4%
Marking
A rating plate shall be fixed on each unit with at least the following informations (in addition to informations required by
safety standards):
- Manufacturer or supplier,
- Manufacturer’s model designation and serial number,
- The COP of two significant figures and the test conditions at which it is measured (T1),
- The heating capacity in kW, with one digit after the decimal comma but no more than tree significant figures.
Performance certification
When tested by Eurovent laboratory, the difference with the obtained and claimed characteristics will be less than:
Air to air heat pump:
Cooling capacity or heating capacity: -8%
Effective power input: +8%
Air to water heat pump, water to water heat pump:
Cooling capacity or heating capacity: -5%
Effective power input: +5%
Technical analysis
Eurovent database
For the reversible models, the parameters included in Eurovent database are:
97
February 1, 1998 EUROVENT Certification
Directory of Certified Products Page XX
January 31, 1999 MANUFACTURER NAME
Participant :
Trade name :
Air conditioners up to 12 kW / Air cooled / Split / Reverse cycle
AC 1 / A / S / R
FLUID Outdoor Indoor
Model R XX Pc Pe (c) Ph Pe (h) MPS ducted Lw ducted Lw Mounting
Designation kW kW kW kW or not dBA or not dBA
Outdoor Indoor unit
Type : XXXXX
XXXXXXX XXXXXXX 3,20 1,04 3,20 0,95 230 - 1 - 50 N 63 N 43 L/S
XXXXXXX XXXXXXX 4,23 1,67 4,38 1,55 230 - 1 - 50 N 64 N 45 L/S
XXXXXXX XXXXXXX 5,23 2,22 5,96 2,00 230 - 1 - 50 N 67 N 51 L/S
XXXXXXX XXXXXXX 7,00 2,79 6,99 2,47 230 - 1 - 50 N 67 N 55 L/S
Pc: Total cooling capacity (kW)
Pe (c): Effective power input in cooling mode (kW)
Ph: Total heating capacity (kW)
Pe (h): Effective power input in heating mode (kW)
M.P.S.: Main Power Supply
Lw dbA = A-weighted sound power level (noise)
Energy Efficiency ratio
Energy Efficiency Ratio is calculated as:
EER (in cooling mode) = [Total cooling capacity (in kW)] / [Effective Power Input (in kW)]
EER (in heating mode) = [Total heating capacity (in kW)] / [Effective Power Input (in kW)]
Minimum and maximum values for EER are shown in the following tables:
RAC Category EER in cooling mode EER in heating mode
Min Max Average Min Max Average
Multisplit 2.08 2.94 2.51 1.44 3.58 2.63
Split 1.45 3.45 2.46 1.65 3.70 2.81
For example, for a non reversible air conditioner system, the minimum and maximum values for EER are:
RAC Category EER in cooling mode
Min Max Average
Multisplit 1.91 3.32 2.66
Split 1.59 3.25 2.40
The difference between performance in cooling mode for a non reversible and a reversible air conditioner is no
significant. The different supplementary components of a reversible system do not infuence the global performance.
98
Technical/economic analysis
The optimisation at the heating mode of the split system depends on:
- the climate conditions,
- the heating energy need,
- the cost of the split system,
- the electricity price.
A good optimisation reduces the electricity consumption and the global cost of the system.
For these systems, we have to designed the split for all the cooling loads of air conditionning. At the heating mode, the
optimisation depends on the climate and energy needs. In the following tables, we see the optimum capacity for:
- 3 locations: Trappes (near Paris), La Rochelle (Atlantic coast), Perpignan (South of France),
- 3 buildings (heating energy need at Trappes): Mozart (4,7 kW), Gershwin (6 kW), Vivaldi (7 kW).
Mozart: heating energy need = 4,7 kW
Heat pump
capacity at the
minimum 40% 50% 60% 70% 80% 90% 100%
temperature/
Heating needs
Trappes
Heating 4960 4647 4530 4490 4480
consumption
(kWh)
Heating cost 3016 2826 2754 2732 2725
(Francs)
Split cost (Francs) 9631 11540 13448 15357 21084
La Rochelle
Heating 3141 2920 2798 2745 2735 2732 2732
consumption
(kWh)
Heating 1892 1759 1685 1653 1647 1645 1645
cost(Francs)
Split cost (Francs) 7974 9468 10963 12457 13952 15446 16941
Perpignan
Heating 1989 1960 1943 1936 1932
consumption
(kWh)
Heating cost 1163 1146 1136 1132 1130
(Francs)
Split cost (Francs) 11209 12744 14280 15816 17351
The grey parts of this table are the optimum overall costs including investment and operating costs during 15 years and
for this capacity, the split capacity covers the cooling loads..
Gershwin: heating energy need = 6 kW
Heat pump
capacity at the
minimum 40% 50% 60% 70% 80% 90% 100%
temperature/
Heating needs
Trappes
Heating 6337 5937 5786 5739 5726
consumption
(kWh)
99
Heating cost 3854 3610 3519 3490 3482
(Francs)
Split cost (Francs) 11752 14191 16630 19069 26386
La Rochelle
Heating 3738 3583 3514 3502 3498 3497
consumption
(kWh)
Heating cost 2252 2158 2116 2109 2107 2106
(Francs)
Split cost (Francs) 11562 13475 15389 17302 19215 21128
Perpignan
Heating 2532 2494 2473 2463 2459
consumption
(kWh)
Heating cost 1481 1459 1447 1441 1438
(Francs)
Split cost (Francs) 13722 15676 17630 19585 21539
The grey parts of this table are the optimum overall costs including investment and operating costs during 15 years and
for this capacity, the split capacity covers the cooling loads..
Vivaldi: heating energy need = 7 kW
Heat pump
capacity at the
minimum 40% 50% 60% 70% 80% 90% 100%
temperature/
Heating needs
Trappes
Heating 7358 6893 3718 6662 6647
consumption
(kWh)
Heating cost 4475 4192 4085 4052 4043
(Francs)
Split cost (Francs) 13321 16152 18984 21815 30309
La Rochelle
Heating 4380 4198 4117 4103 4098 4097
consumption
(kWh)
Heating cost 2639 2529 2480 2471 2468 2468
(Francs)
Split cost (Francs) 13204 15446 17688 19930 22171 24413
Perpignan
Heating 2936 2892 2867 2857 2851
consumption
(kWh)
Heating cost 1717 1692 1677 1672 1668
(Francs)
Split cost (Francs) 15594 17860 20127 22393 24659
The grey parts of this table are the optimum overall costs including investment and operating costs during 15 years and
for this capacity, the split capacity covers the cooling loads..
In France, a well designed split system is as follow:
- H1 part (continental part of France): the split system capacity is 50% of the heating energy need at the minimum
temperature,
- H2 and H3 parts (Atlantic coast and south of France): the split system capacity is 100% of the cooling load.
100
The electrical consumption for heating and cooling for an average house for three typical different French locations is
(for a 138 m2 house):
Consumption in kWh Trappes La Rochelle Marignanne
(Paris) (Atlantic Coast) (Mediterranean Coast)
Heating 5570 2353 1409
Cooling 178 533 1804
The electrical consumption for cooling in Paris and in the South of France is respectively 4% and 40% of the energy
bill.
Conclusions
From the manufacturer database analysis, we notice a no significant difference of the performance between reversible
and no reversible system.
The reversible systems are economically interesting for customers. For a good installation, we must:
- have a well designed system. Some recommandation guides have to specify the design and the installation. The
optimisation of the split system capacity depends on location and on energy needs,
- Improve the reliability of the systems,
- Improve the performance to be competitive. EDF wants a supplementary test point at the external temperature –7°C
(test with Eurovent) with a minimum COP.
101
APPENDIX 2 - Additional statements of manufacturers’ associations
(CECED and Eurovent/Cecomaf)
102