Top-down evaluation methods of energy savings Summary report
Bruno Lapillonne Didier Bosseboeuf Stefan Thomas
ENERDATA ADEME Wuppertal Institute
EIE_06_128 EMEEES
March 2009
The Project in brief The objective of this project is to assist the European Commission in developing harmonised evaluation methods. It aims to design methods to evaluate the measures implemented to achieve the 9% energy savings target set out in the EU Directive (2006/32/EC) (ESD) on energy end-use efficiency and energy services. The assistance by the project and its partners is delivered through practical advice, technical support and results. It includes the development of concrete methods for the evaluation of single programmes, services and measures (mostly bottom-up), as well as schemes for monitoring the overall impact of all measures implemented in a Member State (combination of bottom-up and top-down). Consortium The project is co-ordinated by the Wuppertal Institute. The 21 project partners are: Project Partner
Country
Wuppertal Institute for Climate, Environment and Energy (WI)
DE
Agence de l’Environnement et de la Maitrise de l’Energie (ADEME)
FR
SenterNovem
NL
Energy research Centre of the Netherlands (ECN)
NL
Enerdata sas
FR
Fraunhofer-Institut für System- und Innovationsforschung (FhG-ISI)
DE
SRC International A/S (SRCI)
DK
Politecnico di Milano, Dipartimento di Energetica, eERG
IT
AGH University of Science and Technology (AGH-UST)
PL
Österreichische Energieagentur – Austrian Energy Agency (A.E.A.)
AT
Ekodoma
LV
Istituto di Studi per l’Integrazione dei Sistemi (ISIS)
IT
Swedish Energy Agency (STEM)
SE
Association pour la Recherche et le Développement des Méthodes et Processus Industriels (ARMINES)
FR
Electricité de France (EdF)
FR
Enova SF
NO
Motiva Oy
FI
Department for Environment, Food and Rural Affairs (DEFRA)
UK
ISR – University of Coimbra (ISR-UC)
PT
DONG Energy (DONG)
DK
Centre for Renewable Energy Sources (CRES)
EL
Contact Dr. Stefan Thomas, Dr. Ralf Schüle Wuppertal Institute for Climate, Environment and Energy Döppersberg 19 42103 Wuppertal, Germany
Tel.: +49 (0)202-2492-110 Fax.: +49 (0)202-2492-250 Email:
[email protected] URL: www.evaluate-energy-savings.eu www.wupperinst.org
The sole responsibility for the content of this publication lies with the authors. It does not necessarily reflect the opinion of the European Communities. The European Commission is not responsible for any use that may be made of the information contained therein.
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
Top-down evaluation methods of energy savings (Work package 5) Bruno LAPILLONNE Didier BOSSEBOEUF Stefan THOMAS
Contents 1 Introduction .................................................................................................. 4 ODYSSEE indicators .......................................................................................... 5 2 Adjustments considered to assess “measure-related energy savings” 7 2.1 Possible adjustments ................................................................................. 7 2.2 Methodologies proposed for the various adjustments.............................. 10 3 Case studies............................................................................................... 12 4 Conclusion from the case studies ........................................................... 16 4.1 Issues and problems related to the correction of total energy savings .... 17 4.1.1 Correction of market prices ................................................................... 17 4.1.2 Correction of autonomous technological trends .................................... 19 4.1.3 Correction of autonomous trends for countries with energy savings..... 21 4.2 Issues related to data limitations .............................................................. 22 4.2.1 Data gaps .............................................................................................. 22 4.2.2 Unequal level of disaggregation among countries ................................ 23 4.2.3 Difficulty to measure energy savings for some end-uses...................... 23 5 Conclusions on the possibilities of top-down calculation methods for the ESD ............................................................................................................ 24 5.1 What kind of energy savings are the subject of calculation?.................... 24 5.2 Conclusions on the applicability of ‘pure’ top-down methods using ODYSSEE indicators and a regression analysis .............................................. 27 5.2.1 Additional energy savings ..................................................................... 27 5.2.2 All energy savings ................................................................................. 28 5.2.3 Applicable top-down calculation methods ............................................. 29 5.3 What to do for indicators that do not allow a ‘pure’ top-down calculation?... ............................................................................................................. 30 5.4 A potential way forward ............................................................................ 32 6 References ................................................................................................. 35 Annex 1 : Availability of data for EMEES top-down case studies.............. 36
ENERDATA and ADEME with Wuppertal Institute
3
WP5 summary report
1 Introduction The objective of the EMEEES project is to assist the European Commission in the elaboration of methods for monitoring and evaluation of the energy savings achieved by the Member States (MS) in the framework of the EU Directive on energy end-use efficiency and energy services (ESD). This assistance includes the development of concrete methods for cost-effective top-down calculations to be used for harmonised reporting by the MS. This report summarises the work done to develop top-down methods for the evaluation of the energy savings to be included in the national indicative energy savings target of 9%/year by 2016, as required by the Energy Service Directive (ESD). A harmonised model of top-down calculation methods should be developed for the ESD reporting. Harmonisation should give a reasonable freedom for the MS, while the results reported can be compared. Therefore, the methods and the 14 case studies developed by the EMEEES project are a starting point, but are by no means excluding the use of own methods and further methods for other sectors and end uses by the MS. However, harmonisation should be ensured by key elements covered in this report: a general structure for documentation of top-down energy savings, the selection of reference trends as well as potential other correction factors (e.g., for changes in market prices of energy), and a dynamic approach to ensure improvement over time. These must also be consistent with the bottom-up methods that constitute the other type of methods that can be used for calculating ESD energy savings. The development of this harmonised model is a learning process, and the methods should be improved in the future as more experiences for MS will become available and lessons can be learned. Top-down methods refer to methods relying on statistical indicators defined by sector and/or type of end-use calculated from national averages. The use of top-down methods to evaluate energy savings means that “the amount of energy savings or energy efficiency progress are calculated using national or aggregated sectoral levels of energy savings as the starting point”1. Top-down methods rely on “energy efficiency indicators”, such as the indicators developed at the EU level for the last 15 years within the ODYSSEE project2.
1 2
4
Source: ESD Directive ODYSSEE initially covered the EU-15 countries and Norway; it has now be extended to EU-27 countries, plus Norway and Croatia.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
ODYSSEE indicators ODYSSEE indicators include on the one hand indicators expressed in terms of energy units (e.g. toe, GJ, 3 kWh) to monitor trends in energy efficiency or to compare the energy efficiency “performance” of countries (“adjusted indicators”) and on the other hand indicators of diffusions. Energy efficiency indicators are of four different types: • Energy intensities (economic ratios in monetary units), that relate the energy used in the economy or a sector to macro-economic variables (e.g. GDP, value added); they are expressed in toe/ at constant price (e.g. toe/2000). • Unit ( or specific) energy consumption (physical indicators) by sub-sector or end-use, that relates the annual energy consumption to physical indicators (e.g. toe per ton of steel, per car or per dwelling, kWh/ refrigerator, l/100km for vehicles) • Index of energy efficiency progress (ODEX) , that provides a synthetic evaluation of energy efficiency improvement by main sector and for the economy as a whole. ODEX “aggregates” the trends in unit consumption by sub-sector or end-use into one index by sector on the basis of the weight of each subsector/end-use in the total energy consumption of the sector • If the development over time of an indicator of the previous three types or of an indicator of diffusion is 4 in the direction of reduced energy consumption, “Total” energy savings (Mtoe, PJ) by sector or enduse, and by main end-use sector (industry, households, transport) can be calculated by comparison of 5 the value of the indicator in year X with the value of the indicator in a base year . Within ODYSSEE, such energy savings are also presented as a type of energy efficiency indicator. 6
7
Indicators of diffusion aim at monitoring the diffusion of energy efficient technologies and practices as 8 well as of end-use renewables , so as to complement the existing energy efficiency indicators with indicators that are easier to monitor and are more rapidly updated than energy efficiency indicators.
The ESD refers explicitly to ODYSSEE indicators, in particular the composite ODEX indicator developed in the frame of the ODYSSEE project: “In developing the top-down calculation method used in this harmonised calculation model, the Committee shall base its work, to the extent possible, on existing methodologies such as the ODEX model”. The text of the Directive specifies further that “adjustments (need) to be made for extraneous factors, such as degree-days, structural changes, product mix, etc. to derive a measure that gives a fair indication of energy efficiency improvement”. Most of the adjustments explicitly mentioned in the Directive are already done in ODYSSEE9. However the main issue analysed in the EMEES project relates to additional adjustments that may have been included under “etc”
3 4
5
6 7 8
All indicators are also available in terms of CO2. . Total energy savings are in some cases (cf. chapter 5) a good measure for “all” energy savings, which are those resulting from all technical, organisational or behavioural actions taken at the end-use level to improve energy efficiency whatever their driving factor (energy services, polices or market forces). Energy savings of a given appliance (e.g. refrigerators) or end-use are derived from the reduction in the average unit energy consumption of the appliance (kWh/year) ; a reduction of this specific consumption of 100 kWh will result in total savings equal to 100 GWh (assuming a stock of refrigerators of 1 million units). For instance: number of efficient lamps sold, % of label A or A++ in the sales of electrical appliances. For instance: % of passenger transport by public modes, % of transport of goods by rail. For instance: % of solar water heaters.
ENERDATA and ADEME with Wuppertal Institute
5
WP5 summary report
On this issue there are conflicting interpretations by the different member countries, which can be summarised as follows:
From one viewpoint that points at the exact wording of the ESD’s Articles and Annexes, only the adjustments for extraneous factors explicitly mentioned in the Annex IV should be made; this viewpoint assumes that ESD energy savings are equal (or very close) to the total top-down savings as calculated with ODYSSEE indicators.
Another viewpoint points at the intention of the ESD to contribute to additional energy savings compared to autonomous trends (which are, in turn, leading to additional energy intensity reductions in final consumption, ESD recital 10). From this viewpoint, additional adjustments (etc..) or corrections should be made so as to only measure savings linked to explicit energy efficiency improvement measures; to justify this viewpoint, reference is made to the following statement in the Directive saying that “the overall national indicative energy savings target of 9 % for the ninth year of application of this Directive, to be reached by way of energy services and other energy efficiency improvement measures”. Before that, the Directive specifies the meaning of energy efficiency improvement measures as “all actions that normally lead to verifiable and measurable or estimable energy efficiency improvement”. With this second interpretation, ESD energy savings are calculated by removing from total energy savings, the energy savings linked to all “other factors” than energy efficiency improvement measures, in particular market energy price and autonomous trend.
Independently on the final decision to be taken on this interpretation of the definition of the ESD savings, the work carried out within the framework of the EMEEES project on top-down methods (Work package 5) aimed at looking at the feasibility of methods to do additional corrections, as this will be presented below.
9
6
A first report has presented the existing ODYSSEE indicators and reviewed what adjustments are already made; it also discussed the data availability. “Definition of the process to develop harmonised top-down evaluation methods (Work package 5), Deliverable D7, December 2007.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
2 Adjustments considered to assess “measure-related energy savings” 2.1 Possible adjustments Many ODYSSEE indicators, as explained above, can provide a measurement of the gross or total energy savings, whatever their origin: e.g. energy efficiency improvement measures, but also autonomous technical progress, market forces (prices). The question is then how to only measure in the total energy savings the savings that are linked to energy efficiency improvement measures? In other words, which other factors not linked to energy services and other (energy efficiency improvement) facilitating measures10 should be corrected and how indicators should be cleaned from these factors? What other factors may be included in the top-down evaluation of energy savings? They are of two types: •
•
10
11
12
13
Factors that contribute to energy savings but are not linked to policies:
autonomous energy savings11, due to technological progress, that takes place even in the absence of any EEI measure;
price-induced energy efficiency progress, linked for instance to the increase in the price of crude oil and imported natural gas; the impact of increase in taxes however has to be taken into account as a horizontal energy efficiency measure and needs to be separately monitored 12.
Factors that may still be embedded in the top-down evaluation of energy savings, which on the opposite may minimise their values as they correspond to negative savings: e.g. direct rebound13 effects (cars and households mainly), economic rebound, energy price decreases, that may have the same effects as direct rebound effect.
An (energy efficiency improvement) facilitating measure is a term used within the EMEEES project to include energy services, energy efficiency programmes, other energy efficiency policies, and all other types of measures that support market actors in implementing technical, organisational, or behavioural actions to improve energy efficiency. The autonomous progress depends on the rhythm of level of economic growth: high growth implies faster turnover of energy equipment, thus more possibilities for savings. The debate, however, arises, when it comes to the increase of general mineral oil taxes. Although, they are not explicitly directed towards energy savings, but rather to raise new revenues for the government, there is always a ”green” justification”. Such tax increases have often a substantial impact on energy efficiency. The reading of the ESD tends to indicate that such measures are to be excluded but this is certainly further a matter of debate within the Committee. Direct rebound effect corresponds to the fact that part of the avoided expenditures resulting from the energy savings is used to increase comfort levels (e.g. higher indoor temperature after implementing insulation measures, increase use of an equipment following the installation of an energy efficient equipment (e.g. compact fluorescent lamps, efficient car, solar water heater).
ENERDATA and ADEME with Wuppertal Institute
7
WP5 summary report
Economic rebound effects are difficult to assess and may not always have an impact on energy use14. They will not be considered, as a number of analyses have shown that they are usually only a few percent of the energy savings15. Finally, the Directive restricts the eligible measures to recent measures 16. This means that the impact of older measures, implemented before 1995, should in principle not be taken into account17. In the same way, the question of how to deal with non energy related measure that have an impact on energy use can be raised and whether they should be left out (e.g. expenditures in transport infrastructures, which depending on the mode may increase or decrease the energy consumption). Figure 1 summarises the different effects that might, in theory, be accounted for to derive energy savings as defined in ESD from the total energy savings calculated in top-down evaluation. The effects of some factors may be removed or added (e.g. hidden structure effect, price effect, even the autonomous progress 18), depending on their actual impact on the consumption. Figure 1: Summary of effects that may be cleaned from total top down energy savings
14
15
16
17
18
19
8
19
Economic rebound effect induced by energy efficiency improvement promotion measures means that the avoided energy expenditures is spent on non-energy products or services, that may lead to indirect additional energy consumption (e.g. spend the money saved on an air ticket). It was estimated in the order of 5 %, for instance, according to a study of the Wuppertal Institut (Irrek/Thomas 2006). This is probably smaller than the uncertainty margin of the autonomous and price-induced energy savings. Savings eligible for the ESD are only facilitating measures from 1995 on (in some cases, from 1991 on)”, referred to as “earlier actions”. However, total energy savings in top-down evaluation contain by nature the saving effect of all old measures. In the same way, the effect of how non energy related policy measures that have an impact on energy use may be left out (e.g. expenditures in transport infrastructures. Before the agreement with car manufacturers, the actual autonomous trend for new cars was towards an increase of the car consumption due to more powerful cars and catalytic converters. Price effect refers to the market price component of the price and not to its tax component that is part of the facilitating measures.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
The approach followed in EMEEES was very pragmatic and relied on what is feasible in terms of data availability and/or actually corrected in existing top down method. For this reason, only two effects were considered for possible corrections: • Autonomous progress trend, • Market price increase. The final decision, as to the exact boundaries of the eligible energy savings in the topdown approach, and therefore of the necessary corrections, is not part of EMEEES and will eventually come from the European Commission, possibly by considering the results of EMEEES, and endorsed by the ESD Art. 16 - Committee. Table 1 discusses the pros and cons of the correction of the effects of market energy price variations, based on the discussion and arguments raised in the various national seminars and EMEEES meetings. Table 2 discusses the pros and cons of the correction of the calculated energy savings by excluding the energy savings due to the autonomous technological trend for end-use equipment. Table 1: Pros and Cons of market energy price correction
ENERDATA and ADEME with Wuppertal Institute
9
WP5 summary report
Table 2: Pros and Cons of correction of autonomous technological trend
2.2 Methodologies proposed for the various adjustments The impact of autonomous trend and price effect has been measured through econometric regression20: the variation in the indicator of energy savings21 is explained by different variables, one of which is time, to capture the trend, and another one the energy price. Using such regression analysis allows to evaluate energy savings compared to an autonomous trend, even if the trend of the underlying indicator does not allow to calculate “total” energy savings. Simple econometric methods were used to quantify the impact of prices and trends, on purpose, taking into account several criteria: • the need of transparency and of harmonisation among countries, • the easiness of implementation and of their understanding as such methods would ultimately need to be applied by the countries; • finally, the data limitations, in particular for additional explanatory variables (e.g. price/tax on cars, cost of equipment) and the uncertainty of the data handled.
Even such simple methods raised a lot of questions for their concrete implementation. The typical regression equation considered was follows: ln ES = a + b T + c ln P + d ln A + K with : ln : logarithm; ES: energy saving indicator; b: trend; T: time; P : energy price; c : price elasticity22 ; A: macro economic variable (e.g. GDP) to capture the impact of business cycles; d : elasticity to GDP; K: constant coefficient 20
21
22
10
This econometric evaluation is either made for both factors (trend and price) simultaneously or independently. Indicator of energy savings refers to the indicator, the variation of which is used to calculate the energy savings. Price elasticity may be differentiated between upward and downward price elasticity.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
The estimate of the regression coefficient is made over a period ending before the effects of facilitating measures will have to be assessed (e.g. before 1995). Then using the coefficient, the impact of the different effects can be removed over the period on which the ESD savings will be calculated (i.e. 2008-2016) (Figure 2). The price effect can be separated into two components, ex-tax energy price (market component) and energy tax (policy component), using the same price elasticity . Figure 2: Method to remove other factors from total top-down energy savings (example)
The harmonisation of the method may mean two things: • Use the same equation for each country and calibrate the coefficients for each country • Use one equation for all countries and pool the data for all countries, so as to have a single set of coefficients valid for all countries. The second approach seemed preferable, as the same trend and the same relation to price will be used for all EU countries, which seems logical, given the similarity of price level and behaviours. In practice, the second approach turned out difficult to implement and did not give good results, as explained later.
ENERDATA and ADEME with Wuppertal Institute
11
WP5 summary report
3 Case studies The methodology was validated on 15 case studies that were classified according to the statistical indicator used to calculate the energy savings, as energy savings are derived from the indicators variations (Table 3): • Market diffusion indicator • Specific energy consumption of an equipment • Unit energy consumption of a sub-sector • Total energy consumption. Table 3: Classification of top-down methods in EMEEES
The meaning of the trend was different depending on the indicator: it was close to an autonomous technological trend for end-use equipment but had a much broader meaning for market diffusion indicator or for unit energy consumption of a sub-sector. In addition, for this last category of indicator, another issue was raised due to the fact that, for several end-uses, the indicator is increasing, meaning that “total” energy savings cannot be measured with top-down indicators, even if they do exist (case of electricity uses in services for instance). Table 4 summarises the issues raised by the correction of price and trend for the various types of indicators.
12
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
Table 4: Issues with top-down methods in EMEEES
Six case studies have been selected to test in a comprehensive way the methodology so as to cover the different sectors and the different types of statistical indicators used to measure energy savings; they are the following by sector. The second column of the table indicates, whether the development over time of the indicator allows the calculation of “total” energy savings compared to the value of the indicator in previous years. It should be noted that this can change over time as the direction of development of an indicator changes. Table 5: Comprehensive top-down case studies in EMEEES
Sector and indicators
Residential sector : Solar thermal collectors (market diffusion indicator) Specific uses of electricity (unit energy consumption indicator) Transport sector: New cars (specific energy consumption indicator) Modal shift in goods transport (market diffusion indicator) Industry sector: Industrial thermal energy use (excl. electricity) (unit energy consumption indicator);
ENERDATA and ADEME with Wuppertal Institute
indicator allows the calculation of “total” energy savings compared to the value of the indicator in previous years yes no in many countries
yes no in most countries
no in most countries
13
WP5 summary report
Tertiary sector: Electricity end- uses (excl. thermal uses) (unit energy consumption indicator);
no in most countries
The other case studies have been considered in a less comprehensive way to identify if additional issues would come up: Table 6: Less comprehensive top-down case studies in EMEEES
Sector and indicators
Residential sector : Specific white goods (refrigerators) (specific energy consumption indicator) Building shell and heating (unit energy consumption indicator) Transport sector: Cars, bus and trucks (specific energy consumption indicator) Modal shift in passenger transport (market diffusion indicator) Industry sector: Industrial electricity use (unit energy consumption indicator) Industrial CHP (market diffusion indicator) Tertiary sector: Building shell and heating (unit energy consumption indicator) General policy instruments: Energy taxation (energy consumption indicator)
indicator allows the calculation of “total” energy savings compared to the value of the indicator in previous years yes
yes
yes no in most countries
no in most countries no in many countries
yes
not relevant (calculation is on price elasticity)
Originally, developing a second method for a general policy instrument, targeted information campaigns, was intended. However, the effect of such campaigns is part of packages of all policies, energy services, and other facilitating measures influencing the development of the indicators analysed. It turned out that it is impossible to
14
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
separate their effect, also because energy savings from such campaigns are usually too small to be detected by a top-down indicator. All case studies were carried out in a standardised way in 9 steps, as follows: 1. Identification of relevant indicators from ODYSSEE to measure total energy savings for the sector or end-use/equipment; 2. Review of all possible variables that could be considered to correct total energy savings, independently of the data availability; 3. Review of available data sources; 4. Selection of variables to model energy savings, taking into account the data constraint; 5. Analysis of indicator trend for each EU country and grouping of countries in homogeneous groups; selection of representative countries to be studied in the case studies to adapt and validate the methodology; 6. Identification of existing energy efficiency policies (including early actions?), from MURE data base and other sources; 7. Econometric analysis for each country case study: Selection of the regression period (accounting for policy if any); Selection of relevant explanatory variables, both from statistical (statistical test) and economic point of views (meaningfulness); Quantification of parameters; 8. Calculation of ESD savings as “total” savings minus savings linked to market energy price and trends; 9. Conclusions and issues for replication to all countries
The presentation of the case studies is given in a separate report23.
23
WP5-Case studies, Enerdata, April 2008.
ENERDATA and ADEME with Wuppertal Institute
15
WP5 summary report
4 Conclusion from the case studies A general conclusion of the econometric analysis carried out to measure the trends and price effects is that the results obtained were not very robust, and the price elasticity or trend were often not significant from a statistical or economic point of view, as discussed below. One of the reason was the fact that the data series used were often too short24, either because longer time series are missing (case of some new member countries), or because the regression had to be stopped when facilitating measures were implemented 25. In addition, the case studies and the presentations in the national seminars have raised two types of problems or questions, related, on the one hand, to the corrections made to get the ESD energy savings and, on the other hand, to data limitations. With respect to the corrections of total savings, three questions came up: •
How to correct for the price effect, or in, other words what value to use for the price elasticity, taking into account the insignificant results of most econometric regressions?
•
Which reference trend to use for the autonomous trend of a specific equipment (transport mode, electrical appliance), as different historical trends may be taken as a reference, depending on which historical period is considered? Should a national trend be used or the same trend for all countries? For instance, for the specific consumption of new cars, there is the trend before the ACEA/JAMA/KAMA agreement with the European Commission in 1995 and the trend after 1995, which is very different among EU countries.
•
For some countries, for which “total” energy savings can be measured (i.e. the indicator’s development is going in the right direction), but cannot be linked easily to well dated measures, should there be a correction for an autonomous trend or should all the savings from 2008 onwards considered without correction26?
With respect to the data limitations, several problems were encountered:
24
25
26
16
Even if present evaluations of price elasticities are not significant, calculation may be more relevant if the most recent years with high prices were included (e.g. 2005-2007)… but only if no policy has been implemented over these years. To measure a trend that is cleaned from the impact of facilitating measure the regression had to be made over a period without measures. For instance if a facilitating measure was implemented in 2000, the regression had to be made up to year 2000 (e.g. 1990-2000). This is the case of market diffusion indicators that are increasing (e.g. solar heaters, cogeneration, share of public transport) or unit consumption indicators that are decreasing (unit electricity consumption per employee). For instance, the share of rail and water is increasing in the UK as a result of a mix of factors: the question is whether all the energy savings calculated from this modal shift should be included in the ESD savings.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
•
Data were not available for some countries to calculate the required indicator27;
•
Only aggregated data were available in some countries, when disaggregated data would enable to remove the effect of hidden structural factors and enable a better assessment of energy savings 28;
•
For some end-uses or sub-sectors, only an aggregated indicator can be used to calculate savings, which often is going in the wrong direction (increasing when energy savings should make it decrease and vice versa), which means that no apparent energy saving can be measured from the variation of the indicator: this is the case of the unit electricity consumption indicator in most sectors and countries, and for modal shift in transport29.
4.1 Issues and problems related to the correction of total energy savings 4.1.1 Correction of market prices The question here is what value should be used for the price elasticity, taking into account that for most countries and case studies the values obtained were either not statistically significant or not relevant from an economic viewpoint 30? • National data if relevant (for the few countries where the values obtained from statistical regression were significant); • Or harmonised values, the same for all countries, by sector/end-use Table 7 summarises the pros and cons of these two options. The proposal of the EMEEES project is to select the second alternative so as to better harmonise the corrections made. Table 7: Use of national price elasticity vs harmonised values ?
27
28 29
30
Annex 1 describes the situation of the data availability. In addition, in industry energy consumption and activity data are not available by industrial branch for industries not included in the ETS, and therefore covered by the Directive. This is the case for instance of the energy consumption by sub-sector in the service sector. Typically, when there are energy savings, unit electricity consumption should decrease and modal shift of public transport should increase. The price elasticities were often not economically significant (>0 value whereas should be - 0.8%/year for diesel Figure 3: The different options for the definition of an autonomous technological trend: new diesel cars in France
20
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
The implications on the assessment of ESD savings after correction for an autonomous trend will depend of course on the actual country’s trend as well as the reference trend used (national, EU average, average of 3 slowest EU trends): •
•
For countries, with a national trend slower than the EU average trend, the ESD savings will be reduced if an EU average trend is used instead of the national trend. This is the case of Germany, as illustrated in Figure 4 with a fictive estimation of future savings: the use of an EU average trend instead of national trend would divide, in that example, ESD savings by a factor of 2; if the default value for the reference trend is average of the 3 slowest trends, the ESD savings would be multiplied by 2. For countries, with a national trend more rapid than the EU trend, using the EU average trends would on the opposite reduce the correction and thus increase the ESD savings (e.g. case of new diesel case in France).
Figure 4: Correction for autonomous trend: sensitivity study
31
4.1.3 Correction of autonomous trends for countries with energy savings In the case studies, energy savings have been measured in some countries that could be linked to well identified policies. The question that came up is whether all these energy savings should be accounted for in the ESD savings or whether a correction for an autonomous trend should be done for these countries32. This issue came up for case studies based on market diffusion indicators (e.g. diffusion of solar water heaters33, modal shift to public transport34), as well as for indicators of unit consumption 35. 31
32
33 34 35
This is a simulation of possible developments in Germany to show the impact of different corrections for autonomous trend. In reality, the calculation will only be made once the real trends are known. Previously, the discussion was different and related to the correction and measurement of an autonomous technological trend for a given equipment. Case for instance of. Cyprus and Germany Case of the transport of goods in UK Case of the electricity consumption in services in Sweden (kWh/employee).
ENERDATA and ADEME with Wuppertal Institute
21
WP5 summary report
Such a correction would enable to remove energy savings that may be just linked to other policies (e.g. transport) or to hidden effects that have not been corrected (e.g. saturation in services). However, as most countries do not show energy savings in these sectors/end-uses, countries with good results should be credited. In EMEEES, we consider that this correction should not be made. This could even underestimate energy savings compared to the autonomous trend, which could be increasing in examples such as the one presented in the following figure. Figure 5: Corrections for autonomous trend for market share or unit consumption indicators
4.2 Issues related to data limitations 4.2.1 Data gaps For some case studies, data are not available for part of the countries on the indicator used to measure total energy savings. This is for instance the case of the household electricity consumption of electrical appliances and lighting (i.e. excluding thermal use), which is not available for 11 countries. Another example is the specific consumption of new cars that is not available for 9 countries. Annex 1 lists the data availability for the various case studies. To cope with this issue, the proposal of EMEEES is the following: •
•
22
For specific energy consumption indicators (e.g., appliances, cars) and maybe for diffusion indicators, countries may try to reconstruct past indicator values but this may cause a considerable cost; if not, they will try to get the value for the year 2008 and apply EU default correction factors for market energy prices and autonomous technical progress. For unit energy consumption indicators (by sector or sub-sector), the method of correction is applicable if default values are used for the reference trend: the correction can then be applied, as soon as a country starts to develop the indicator.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
4.2.2 Unequal level of disaggregation among countries For some case studies, different approaches can be applied depending on the availability of data: a detailed approach to clean “hidden structure effects” or an aggregate approach if detailed data are not available. However, the two approaches do not give the same trend for the indicator and thus the amount of energy savings. Therefore the question is whether the same approach should be applied to all countries, with the risk to harmonise to the situation of countries with the poorest data, or whether the most detailed level of disaggregation should be used whenever possible? To see the impact of working with aggregate or disaggregate indicators, an evaluation of energy savings was done in the case of the electricity uses in the service sector in Sweden. Trends calculated over the period 1985-1994 give different savings over 1995-2004 with aggregate approach or with the detailed approach by sub-sector the difference (about 1000 kWh/employee in 2004 and about 3 TWh in terms of savings) is due to hidden structural effects (shift to sub-sectors with lower electricity consumption per employee (Figure 6) Figure 6 : Impact of the level of aggregation: case of services
4.2.3 Difficulty to measure energy savings for some end-uses For several case studies, generally based on indicators of unit consumption for electricity, the indicators are increasing in most countries instead of decreasing, which means that “total” energy savings cannot be calculated from the reduction of the unit consumption (cf. tables 5 and 6). This does not necessarily mean that there is no energy saving; it rather reflects the fact that the energy savings are hidden among other factors that drive the indicators upwards (e.g. new end-uses, such as air conditioning or IT appliances). In other words, the lack of more detailed data makes it difficult to measure energy savings. In the case studies, we have proposed to estimate the savings, not from a reduction of the indicator, but a slow-down in the rate of increase of the indicator. This point deserves further decision from the Committees on how to deal with this issue. It also shows the limit of top-down methods to measure energy savings in some end-uses and the need to improve the data collection in these sectors.
ENERDATA and ADEME with Wuppertal Institute
23
WP5 summary report
5 Conclusions on the possibilities of top-down calculation methods for the ESD 5.1 What kind of energy savings are the subject of calculation? The preceding chapters presented the analysis of Top-Down evaluation methods in the EMEEES project. This analysis started from the review of existing ODYSSEE indicators. The starting hypothesis was that it is possible first to calculate “total” energy savings by comparison of the value of the indicator in year t with the value of the indicator in a previous year. This summary report has then looked at the way these total energy savings could be corrected for an autonomous trend and for effects of changes in market energy prices. The idea was to calculate energy savings due to energy efficiency improvement measures in that way. The analysis showed that for some of the indicators, the development of their value shows a decreasing trend of specific or unit energy consumption over time, or an increasing trend of the diffusion of energy-efficient technologies or transport modes. For these indicators, “total” energy savings could be calculated in the way intended (cf. list of indicators analysed as case studies in chapter 3). For other indicators, there is an increasing trend of specific or unit energy consumption over time, or a decreasing trend of the diffusion of energy-efficient technologies or transport modes. For these, no “total” energy savings can be calculated. However, it should in principle be possible to define an autonomous trend and the effects of changes in market energy prices. However, what is really needed for top-down calculation of energy savings for the ESD? What are, therefore, the conclusions from the analysis performed here for the applicability of top-down evaluation methods for the ESD? As said above, some stakeholders and experts in the debate on ESD implementation interpret “total” energy savings figures as those relevant for the evaluation of ESD energy savings, using the definition given by the ESD for energy efficiency improvement measures. These include ‘all actions’ leading to measurable energy savings. Others, however, claim that energy savings due to autonomous trends of energy efficiency improvement should be excluded. Autonomous trends are those due to autonomous progress in energy efficiency, observing that with time, most technologies are developed further and use less energy per unit of service provided, even without energy services and other energy efficiency improvement measures specially introduced or encouraged by the Member States or market actors to improve energy efficiency. Annex IV of the ESD includes text about “measuring the realised energy savings as set out in Article 4 with a view to capturing the overall improvement in energy efficiency and to ascertaining the impact of individual measures”. Therefore, evaluation methods must be able to both measure the total (all) energy savings, and the energy savings
24
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
that are due to energy services and other energy efficiency improvement measures introduced or encouraged by the Member States, and hence are additional to autonomous progress. However, even in cases for which “total” energy savings can be calculated, what would be the trend of the indicator without any energy efficiency improvement, and what would be its autonomous trend? All or “Really total” energy savings would need to be calculated against the reference trend of the indicator without any energy efficiency improvement action. This reference trend could be stable, decreasing, or increasing, depending on the drivers of consumption and hidden structural effects. Equally, energy savings that are “additional” would need to be calculated against the autonomous trend. However, autonomous trends could also either be stable, or decreasing, or even increasing, depending on the drivers of consumption and structural effects. The two following graphs are illustrating these different types of energy savings quantities for two typical situations of the observed trends of top-down indicators. Figure 7 presents the situation with a downward trend of a specific or unit energy consumption indicator. In this situation, it is possible to calculate “total”, or better “apparent total” energy savings from the difference between a fixed indicator value (in blue, dotted line) and the actual development of the indicator (in black). But these “apparent total” energy savings do not have to be equivalent to all energy savings compared to a reference trend without any energy efficiency improvement action. The figure shows that this (hypothetical) frozen efficiency trend (in green) could be increasing, so all energy savings are actually greater than “apparent total” energy savings that are just calculated against a fixed indicator value. The figure also shows how the additional energy savings can be calculated from the difference between the autonomous trend (in red), obtained by extrapolation of the historical trend, e.g. through regression analysis as proposed in this report, and the actual development of the indicator (in black).
ENERDATA and ADEME with Wuppertal Institute
25
WP5 summary report
Figure 7: Reference trends and types of energy savings quantities for a specific energy consumption indicator with a decreasing trend
Figure 8, by contrast, presents the situation with a strong upward trend of a specific or unit energy consumption indicator. In this situation, it is not possible to calculate “apparent total” energy savings from the difference between a fixed indicator value (in blue, dotted) and the actual development of the indicator (in black). In this example, “total” energy savings would have a negative value, so there are no “apparent total” savings. But it would be possible to calculate all energy savings compared to a (hypothetical) reference trend without any energy efficiency improvement action (frozen efficiency; in green). In this case, too, the additional energy savings can be calculated from the difference between the autonomous trend (in red), obtained by extrapolation of the historical trend, e.g. through regression analysis as proposed in this report, and the actual development of the indicator (in black).
26
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
Figure 8: Reference trends and types of energy savings quantities for a specific energy consumption indicator with a increasing trend
5.2 Conclusions on the applicability of ‘pure’ top-down methods using ODYSSEE indicators and a regression analysis 5.2.1 Additional energy savings Energy savings that are additional to an autonomous trend and to energy savings due to increases in market energy prices could, in principle, be evaluated from long timeseries of indicators via a regression analysis (cf. the formula in chapter 2.2). This was our starting hypothesis: a regression over past periods, without facilitating measures in place, would deliver the baseline projection. Extending this reference trend over the period 2008-2016 would allow to calculate ESD energy savings by comparison with the actual development of the indicator. The positive features of such an approach are twofold: (1) it allows calculating additional energy savings from statistical data only and (2) it allows calculating additional energy savings even for indicators that do not allow to calculate “total” energy savings, since the unit or specific energy consumption value is increasing or the diffusion of energy-efficient technologies or transport modes is decreasing (as, e.g., in figure 8).
ENERDATA and ADEME with Wuppertal Institute
27
WP5 summary report
However, the analysis of case studies was quite inconclusive. In most cases, it was possible to identify a reference trend for some countries but not for all. An exception here is the market diffusion of solar water heating, which can be assumed to be practically entirely due to facilitating measures in the past, so the baseline would be zero market penetration (cf. chapter 4.1.3). Whether this holds for the future, however, would need to be analysed. The same picture showed up for the correction for market energy prices. Therefore, the next hypothesis was to consider using EU default values for both the correction for market energy prices, and for the autonomous trend for the specific energy consumption indicators. •
For market energy prices, the EMEEES proposal is to use price elasticities between 0.1 and 0.2, and only correct for the effects of market energy price increases.
•
For the autonomous trend of specific energy consumption indicators (e.g., for cars and appliances), the proposal is to use the average trend obtained for the three countries with the slowest decrease (i.e., lowest percent change per year) in the value of the indicator. This is based on the assumption that these would be countries without (strong) national EEI measures in place. Such EU default values for the autonomous trend of specific energy consumption indicators should be harmonised with any corresponding EU default values for the percent change per year of the baseline to be used for bottom-up evaluation methods for the same type of equipment. The value achieved through the top-down analysis would be the starting point for such a harmonisation. Such a default value was developed for the average fuel consumption of cars (cf. chapter 4.1.2). Due to budget and data constraints, it was not possible within the EMEEES project to really develop EU average default or country-specific values for the autonomous development of other specific energy consumption indicators (e.g., for appliances).
For unit energy consumption or diffusion indicators, countries are usually so different that it will not be possible to define EU default values for reference trends. Country-specific trends must be defined. For some countries and indicators, this may be done using the regression analysis method, as said above. Furthermore, not all indicators are available for all or most EU Member States (cf. Annex 1).
5.2.2 All energy savings In theory, if all structural effects influencing an indicator are corrected for, all energy savings can be calculated from the difference between the indicator value in the base year and the current value of the indicator in the measurement year (e.g., 2016). All
28
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
energy savings would then be the same as the “apparent total” energy savings found in practice. However, as the analysis of case studies has shown (cf. chapter 3 and Lapillonne/Desbrosses 2009), the indicator is only going into the ‘right’ direction (i.e. showing “apparent total” energy savings) for about 60 % of all the 14 indicators and countries analysed in EMEEES. The reason must be that there are still some structural effects not yet removed due to lack of data. Figure 8 in chapter 5.1 presents an example of such a situation. Therefore, in practice, it may only be possible for some specific energy consumption indicators to assume that “apparent total” energy savings are a good approximation for all energy savings. For all other types of indicators, this would lead to no savings at all or to inconsistent and arbitrary measures of energy savings between Member States.
5.2.3 Applicable top-down calculation methods In conclusion, Table 9 summarises, which top-down calculation methods based on ODYSSEE indicators were analysed in EMEEES, and which of these appear applicable for a harmonised calculation system for the ESD. These are the five marked ‘yes’ in the column ‘Applicable’ in the table. Three are marked with a “sometimes”, as they may be applicable depending on the country situation. The method for global taxation cannot be done if a correction for energy market price and a calculation of the energy savings due to taxation is also done in the other cases, to avoid double counting (thus the “yes*”).
ENERDATA and ADEME with Wuppertal Institute
29
WP5 summary report
36
Table 9: Applicable TD-methods , if data available and corrections possible
As part of WP5, the data availability has been checked for each indicator and each Member States (MS) (cf. Annex 1). The column “Data MS” gives an overview of data availability37. The column ‘Robust results?’ refers to whether the regression analysis has delivered conclusive results, and to methods with specific energy consumption indicators that can be used with a default value for the autonomous trends.
5.3 What to do for indicators that do not allow a ‘pure’ top-down calculation? There are two kinds of problems that cannot be solved with a ‘pure’ top-down calculation: (1) reference trends of autonomous progress (enabling the calculation of energy savings additional to them) for countries, for which a regression analysis is inconclusive; and (2) reference trends of at least the unit energy consumption and diffusion indicators without any energy efficiency improvement action (enabling the calculation of all, i.e., really total energy savings due to all end-use (energy efficiency improvement) actions) for all countries on a country-specific basis. There appears to be only one potential way to calculate reference trends for both of these situations. This only solution appears to be a bottom-up modelling of the whole stock over time, based on surveys of population samples (EMEEES method
36
37
30
The energy savings calculated for new cars and for the improvement of the car, bus, and truck stock must not be added to each other, since the energy savings obtained for the total vehicle stock include those obtained with new cars. “all” represents all the EU-27 countries, “most” around 80% of countries, “many” >15, “EU-15” = 10 to 15, and “few” less than 5 countries. In some cases, the indicators are currently available only for EU-15 (cf. Annex 1), but potentially for all or most of the EU-27 countries. This is mentioned in brackets, e.g., ‘EU-15 (all)’.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
BU6 / TD1)38. The resulting two types of reference trends (i.e., (1) no end-use action, i.e., “frozen efficiency” and (2) baseline projection including autonomous progress) can then be compared with the actual real development of the indicator. These trends would be equivalent to the green lines – reference trend (frozen efficiency) for all energy savings – and red lines – reference trend (autonomous) for additional energy savings – in Figures 7 and 8, respectively. This would apply to unit energy consumption and diffusion indicators for both cases, but to specific energy consumption indicators only for the reference trend without any energy efficiency improvement action. For the autonomous progress of specific energy consumption indicators, we propose to use EU default values (cf. chapter 5.2). However, this potential solution has not yet been tested within EMEEES, and cannot be tested within the project due to limited resources. It is, therefore, at present only a theoretical concept that appears the only potential solution for these cases. The feasibility will depend on data availability for the modelling in a Member State wishing to apply it, which may be difficult for indicators covering many end-uses and subsectors. The applicability will also depend on the possibilities to harmonise the methods and basic assumptions for bottom-up modelling in order to make the results really comparable and thus credible for ESD monitoring purposes. Such a bottom-up modelling also requires more effort than the original regression analysis proposal for determining the reference, but this may only have to be repeated every three to five years to recalibrate the system. The National Energy Efficiency Action Plans have to be filed every three years, and so this recalibration could be repeated in line with the Plans. For the intermediate years, this is quite an easy way to calculate energy savings top down by comparing the actual real development of the indicator with the reference trend obtained from the modelling. Furthermore, if the surveys are further refined to identify, (1) which end-use actions were taken autonomously by the final users and investors, and contribute to the baseline projection, and (2) which were taken due to facilitating measures to explain the actual development of the indicator, the bottom-up modelling method becomes a real bottom-up method. The combination of the modelling results and the actual indicator value they explain then becomes an integrated bottom-up / top-down method. Results of other bottom-up evaluations for facilitating measures in the same sector and / or for specific end-use actions in the sector can be compared with these results (cf. the report from EMEEES WP 6, Boonekamp/Thomas 2009).
38
If the modelling and the surveys do not distinguish, whether the end-use (EEI) actions are taken as a consequence of facilitating measures or not, this is considered a top-down method, too (cf. report from EMEEES WP 3: “Distinction of energy efficiency improvement measures by type of appropriate evaluation method”)
ENERDATA and ADEME with Wuppertal Institute
31
WP5 summary report
5.4 A potential way forward Summarising the above conclusions and considerations, the overall proposal for the harmonised top-down measurement system could be as provided in Table 10 for the three types of top-down indicators that exist. For calculation additional energy savings, the conclusion has been that it can in some cases be done with ‘pure’ bottom-up methods. This is what has been tested within the EMEEES project. It is presented in the upper part of Table 10. For specific energy consumption indicators, this should always be possible using EU default values; for the other two types of indicators – unit energy consumption and diffusion indicators – it may be possible for a country through regression analysis. If such a regression analysis is inconclusive for a unit energy consumption or diffusion indicator, a potential way forward could be to determine the autonomous progress reference trend for additional energy savings through bottom-up modelling. The steps in the calculation that would be needed are presented in the second part in the middle of Table 10. However, it should be noted that it was not possible in EMEEES to test the feasibility of this proposal. For calculation of all energy savings, only for specific energy consumption indicators of appliances and vehicles, the reference trend can be assumed to be equal to the base year value of the indicator39. The calculation then becomes very simple, as shown in the second but last section of Table 10. For all other types of indicators, our conclusion was that, in principle, only bottom-up modelling can correct the hidden structure effects or tell what would have been the reference trend without any energy efficiency improvement action (‘frozen efficiency’). This reference trend is necessary to calculate all energy savings, including those due to autonomous trends and increases in market energy prices. The so-called ‘total’ (or ‘apparent total’) energy savings, if they can be directly calculated from an indicator compared to its value in a base year, are normally not the correct value for all energy savings, since there may be hidden structure effects (cf. chapter 5.1 above, Figures 7 and 8). The steps in the calculation of all energy savings, when using the bottom-up modelling approach for the reference trend, are presented in the lowest part of Table 10. Again, however, it was not possible to test the feasibility of this approach within EMEEES. Further analysis will be needed to examine the feasibility in practice.
39
32
For the solar water heaters, this is even assumed for additional energy savings.
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
Table 10: Steps in top-down calculation for different types of indicators Specific energy consumption indicators
Unit energy consumption indicators
(e.g., appliances, cars)
(e.g., electricity consumption in industry, tertiary, households; heating and process fuel consumption in households, industry and tertiary)
Diffusion indicators (e.g., solar water heaters, CHP industry, modal split in passenger and goods transport)
‘Pure’ top-down calculation of additional energy savings (due to energy efficiency improvement measures) - Approach tested in various case studies by the EMEEES project 1. Set an EU default value for the autonomous technical progress of specific energy consumption indicators (e.g., for cars and appliances), based on a regression analysis for all countries with data available, and on the average of the three countries with the slowest trend found in the analysis.
1. Do the regression analysis for each country. Identify plausible autonomous trends, if possible.
1. Do the regression analysis for each country. Identify plausible autonomous trends, if possible.
2. Set an EU default value for the price elasticity (between 0.1 and 0.2) Correct the reference trend with this elasticity, if the market price of energy is moving upwards
2. Set an EU default value for the price elasticity (between 0.1 and 0.2) Correct the reference trend with this elasticity, if the market price of energy is moving upwards
2. Set an EU default value for the price elasticity (between 0.1 and 0.2) Correct the reference trend with this elasticity, if the market price of energy is moving upwards
3. Calculate additional energy savings from the difference between the reference trend (step 1), corrected for effects of increasing energy market price (step 2), and the actual development of the indicator.
3. Calculate additional energy savings from the difference between the reference trend (step 1), corrected for effects of increasing energy market price (step 2), and the actual development of the indicator.
3. Calculate additional energy savings from the difference between the reference trend (step 1), corrected for effects of increasing energy market price (step 2), and the actual development of the indicator.
If ‘pure’ top-down calculation of additional energy savings not possible: Proposal for Top-down calculation of additional energy savings based on bottom-up modelling of reference trend - Feasibility NOT tested by the EMEEES project, needs further analysis Not applicable
1. Perform a (bottom-up) modelling of the reference trend (autonomous development). This should be harmonised in the methods and basic assumptions between countries as much as possible.
1. Perform a (bottom-up) modelling of the reference trend (autonomous development). This should be harmonised in the methods and basic assumptions between countries as much as possible.
Not applicable
2. Set an EU default value for the price elasticity (between 0.1 and 0.2) Correct the reference trend with this elasticity, if the market price of energy is moving upwards
2. Set an EU default value for the price elasticity (between 0.1 and 0.2) Correct the reference trend with this elasticity, if the market price of energy is moving upwards
Not applicable
3. Calculate additional energy savings from the difference between the reference trend (step 1), corrected for effects of increasing energy market price (step 2), and the actual development of the indicator.
3. Calculate additional energy savings from the difference between the reference trend (step 1), corrected for effects of increasing energy market price (step 2), and the actual development of the indicator.
ENERDATA and ADEME with Wuppertal Institute
33
WP5 summary report
Specific energy consumption indicators
Unit energy consumption indicators
(e.g., appliances, cars)
(e.g., electricity consumption in industry, tertiary, households; heating and process fuel consumption in households, industry and tertiary)
Diffusion indicators (e.g., solar water heaters, CHP industry, modal split in passenger and goods transport)
‘Pure’ top-down calculation of all energy savings - Approach tested in various case studies by the EMEEES project 1. Calculate all energy savings from the difference between the value of the indicator in the base year and the actual development of the indicator.
Not applicable
Not applicable, with the single exception of solar water heaters For solar water heaters: 1. Calculate all energy savings from the difference between the value of the indicator in the base year and the actual development of the indicator.
Proposal for Top-down calculation of all energy savings based on bottom-up modelling of reference trend - Feasibility NOT tested by the EMEEES project, needs further analysis 1. Perform a (bottom-up) modelling of the reference trend without any energy efficiency improvement action (frozen efficiency). This should be harmonised in the methods and basic assumptions between countries as much as possible.
1. Perform a (bottom-up) modelling of the reference trend without any energy efficiency improvement action (frozen efficiency). This should be harmonised in the methods and basic assumptions between countries as much as possible.
1. Perform a (bottom-up) modelling of the reference trend without any energy efficiency improvement action (frozen efficiency). This should be harmonised in the methods and basic assumptions between countries as much as possible.
2. Calculate all energy savings from the difference between the reference trend (step 1) and the actual development of the indicator.
2. Calculate all energy savings from the difference between the reference trend (step 1) and the actual development of the indicator.
2. Calculate all energy savings from the difference between the reference trend (step 1) and the actual development of the indicator.
34
ENERDATA
Bruno Lapillonne, Didier Bosseboeuf, Stefan Thomas
6 References ADEME-IEE, 2007: Evaluation of energy efficiency in the EU-15 indicators and measures, Paris 2007, 141 p. ADEME-IEE, 2007: Evaluation and monitoring of energy efficiency in the new EU members countries and the EU-25, Paris 2007, 106 p. P.G.M. Boonekamp and Stefan Thomas, 2009: Harmonised calculation of energy savings for the EU Directive on energy end-use efficiency and energy services based on bottom-up and top-down methods – development, assessment and application of a practical approach. ECN, Petten and Wuppertal Institute, Wuppertal Wolfgang Irrek and Stefan Thomas, 2006: Der EnergieSparFonds für Deutschland, edition der HansBöckler-Stiftung Nr. 169, Düsseldorf, 2006 (in German) Bruno Lapillonne, 2007: “Definition of the process to develop harmonised top-down evaluation methods (Work package 5), Enerdata, Grenoble, December 2007, 38 p. Bruno Lapillonne and Nathalie Desbrosses, 2009 : Top-down evaluation methods of energy savings, Case studies summary report. Enerdata, Grenoble ODYSSEE: www.odyssee-indicators.org
ENERDATA and ADEME with Wuppertal Institute
35
WP5 summary report
Annex 1 : Availability of data for EMEES top-down case studies
36
ENERDATA
