Primary Energy Factors for Electricity Production in Europe

Abstract

The European Union (EU) has committed to supporting the United Nations’ efforts in line with the Paris Agreement for addressing climate change and has set ambitious targets to reduce primary energy consumption and emissions. Similar commitments have also been set by EU-27 member states. For this purpose, it is necessary to use a primary energy factor (PEF) for converting electricity use to primary energy units and for assessing energy conservation measures. Lower PEFs reflect efficiency improvements in power generation, an increased share of renewable energy sources in the fuel mix for electricity generation, and lower transmission and distribution losses. Over the past decades, there have been intensive efforts and notable progress in the EU-27 for increasing the use of renewables in the energy mix for electricity generation. However, the EU default PEF value for electricity was not regularly updated and remained at 2.5 for several years till it was finally recalculated at 2.1 in the 2018 recast of the Energy Efficiency Directive. This paper reviews different calculation options for estimating the PEF for electricity from official annual statistics, presents the historical evolution of the calculated conversion factors, and provides simple linear correlations for projecting the PEF values that can be used to facilitate more-realistic forward-looking calculations and assess national energy efficiency, climate change, or decarbonization plans in EU-27 member states. A more detailed analysis and case studies on the impacts of this work are illustrated for Greece and Poland.

1. Introduction

The total final energy consumption in the European Union (EU-27) reached 937.9 million tons of oil equivalent (Mtoe) in 2019 [1]. This is considered a more representative year before the anomalies caused during the COVID-19 lockdowns, resulting in a notable drop of the total final energy use to 885.8 Mtoe in 2020.

The most significant energy end user is the buildings sector. Over the years, the final energy use in European buildings has been around 40% of the total final energy consumption. The final energy consumption in residential buildings practically remained constant at 248.22 Mtoe in 2019 and 248.24 Mtoe in 2020, while there was a notable drop in nonresidential buildings from 128.62 Mtoe in 2019 to 121.38 Mtoe in 2020 [1]. In particular, buildings dominate the final energy use, reaching 50.2% in Estonia and only 25.9% in Luxemburg, averaging 40.2% in EU-27 during 2019 (Figure 1).

Progressively, more emphasis is placed on the environmental impacts of the buildings sector as a result of its contribution to the greenhouse gas (GHG) emissions that are responsible for the climate crisis. The GHG emissions from the various forms of energy that are used for the operation of buildings is estimated at 805 million metric tons of carbon dioxide equivalent (MtCO₂-eq), or 24%, of the total GHG emissions [1]. This includes the direct emissions of 435 MtCO₂-eq, or 13%, of the total from the onsite combustion of fossil fuels. In addition, it includes approximately 370 MtCO₂-eq, or 11%, of the total from indirect emissions that relate to the allocated emissions of the energy industry for the use of public electricity and heat production by the buildings sector.

In the European Union, various policies, legislative instruments, and energy and climate targets are commonly articulated around primary energy consumption—for example, the commitment to improve energy efficiency and reduce the use of primary energy by 20% by 2020 [2] and the new target aiming at 39% by 2030 [3]; the building energy performance certificates and nearly zero energy buildings according to the energy performance of buildings directive (EPBD) [4]; and the energy-efficiency values defined in the Ecodesign Directive and in the Energy Label Regulation [5].

Primary energy consumption is defined as any kind of extraction of energy products from natural sources to a usable form [1]. The exploitable natural resources include coal, crude oil, natural gas, etc., while the transformation of energy from one form to another, such as electricity or heat generated by thermal power plants, is not included in primary production.

According to the Energy Efficiency Directive (EED) [6], the relevant analysis is based on calculations that are expressed in primary energy, i.e., gross inland consumption and excluding nonenergy uses. The final energy consumption or savings to be achieved by different policy measures cover all forms of energy supplied to the final end-use sectors (e.g., buildings, industry, transportation).

The conversion of final energy to the equivalent primary energy units is performed using a primary energy factor (PEF) for a given energy carrier. These conversion factors were quantified in Annex IV of EED [6]. Specifically, for calculating energy savings in kWh electricity in primary energy units using a bottom-up approach based on final energy consumption, the EU default coefficient was set at 2.5. The EU member states may consider this value unless the use of other conversion factors can be justified. This implies that on average in the EU, it takes 2.5 primary energy units to generate and deliver 1 unit of energy of electricity at the end users. This value of the PEF was presented in 2012 and remained the same for several years despite the notable changes in the energy mix for generating electricity, the improvement of power plant efficiencies, and technology advances, which also vary significantly among the EU member states and over time. In particular, the PEFs will change from year to year given the growth of renewables for electricity generation, especially during time periods with rapid growth rates. For example, the EU gross electricity generation from renewables was 22.9% in 2010 and only 5 years later climbed to 30.5% [1], which drove PEFs to significantly lower values.

However, it was not till 2018, during the recast of the EED, that the PEF for electricity was finally updated and lowered to 2.1 [7], using the calculations of the energy balances from the energy system model and grid losses of about 6%. Recognizing the evolving dynamics and the importance of more representative PEF values, the EED mandated that by the end of 2022 and every 4 years thereafter, the Commission shall revise the default coefficient on the basis of observed data. Even then, the challenge remains for national policymakers and other stakeholders to have easy access to information that can facilitate their forward-looking assessment of efficiency measures and other policies during the next 5 years or the following decade.

In the meantime, EU member states have used different PEFs for electricity to define and benchmark buildings energy performance according to EPBD [4]. During the early years of EPBD implementation, it was evident that PEFs were not based entirely on clear calculation steps and scientific arguments [8]. From a study among seven countries, it was revealed that in many cases, there were no clear references or comprehensive information to support the calculation of the specific conversion factors, because the PEFs were to some extent the result of a political process, such as in the case of France, the Netherlands, and Sweden [8].

More recently, the PEF values were calculated for EU Nordic countries for which renewables dominate the energy mix, obtaining the following values: 1.2 in Finland, 1.6 in Sweden, and up to 2.0 in Estonia [9]. Other reported values in the literature range from 1.2 in Finland to 2.5 in Lithuania, Portugal, and Slovenia [10]. A study in Slovenia considered three commonly used methods to calculate a national PEF for the electricity mix in 2020, resulting in a value of 2.4, which accounted for only the nonrenewable share, up to an average total value of 2.8 and to provide linear projections for 2030 and 2040 [11].

Even more complex are the calculations at district level that involve different loads, and various onsite technologies that exploit renewables [12]. Different methodologies are also used for estimating the PEF values for renewable and nonrenewable primary factors in relation to electricity generated from wind turbines [13], photovoltaics [14], and hydropower plants [15].

The importance of consistent updates in order to derive accurate estimates of the PEFs was also underlined by a recent study for assessing the conversion factors over time in different EU member states [16]. The study revealed notable variations in the PEFs among EU member states over the past 20 years, ranging from a decrease of only −7% in France to a notable decrease of −32% in Denmark [16].

In any event, the fact remains that several EU countries either use the default European value or do not regularly update their national conversion values used in energy efficiency, climate change, or decarbonization plans for assessing progress toward short- and long-term targets. This may have a significant impact on the results in countries where there is a strong growth of renewables in the energy mix for electricity generation (such as in the case of Greece), while it may be less important in others where electricity generation in based mostly on fossil fuels (such as in the case of Poland).

For example, the PEF value of 2.9 was introduced in Greece by the national regulation published in 2010 for the EPBD transposition [17] and has not been updated since then. Given that the use of renewables in the Hellenic electricity mix was only 13.7% in 2009 and 10 years later reached 33.2% in 2019 [1], it is evident that the value of the PEF for electricity has significantly changed over the past decade. On the other hand, the value of 3.0 was introduced in Poland by the national ministerial ordinance issued in 2015 [18]. Over this period, the use of renewables in the Polish electricity mix has moderately increased from 12.8% in 2014 to 16.0% in 2019, because its power generation is dominated by solid fossil fuels [1]. For comparison, in the United States, the source (primary) energy conversion factor is calculated as the ratio of the total energy for the various generators (e.g., coal, petroleum, natural gas, nuclear, renewables) to the net generated electricity from domestic production that is delivered to customers (e.g., excluding pumped storage and transmission and distribution losses). For example, for the US national grid in 2019, the numerator is 9.504 × 10¹² kWh, and the denominator is 3.748 × 10¹² kWh, which results in a conversion factor of 2.54 [19].

Looking into the future, the focus is toward zero-emission buildings in order to fully decarbonize the European building stock by 2050 [20], by placing an emphasis on the electrification of buildings, by using clean (green) electricity. Already, the EED has introduced an obligation to annually renovate 3% of the total floor area of public buildings and set up long-term strategies and financing instruments to renovate national building stocks with a priority on low-carbon measures. Progressively more emphasis will be placed on upgrading heating and cooling systems during major building renovations toward the 2050 carbon neutrality target, by minimizing building loads [21] and then using a carbon-free energy source to cover the demand. The evaluation of different scenarios for long-term building renovation plans under the EU Renovation Wave [22] should base the decision-making process on forward-looking data over the next decades, instead of historical data and obsolete conversion factors.

Given the continuous and significant technological progress and improvements in power generation, the aggressive efforts throughout the EU to increase the use of renewables in the fuel mix for electricity generation, along with lower transmission and distribution losses, the PEF values that were calculated from years-back data are outdated. Beyond the inaccuracies introduced in the current analysis of energy balances, there is an even greater need to have simple correlations for future projections that can be easily adapted in relevant long-term studies in order to make more-realistic plans. Accordingly, the purpose of this work is to quantify the PEFs for all EU-27 member states, document their historical evolution, and derive simple correlations that can facilitate future projections.

The work elaborated here considers different calculation approaches for estimating the primary energy factor for electricity based on readily available data and a standardized approach. The results are compared with the officially reported EU average values and representative national examples. Furthermore, the work derives easy-to-implement correlations of national conversion factors for estimating primary energy use in future energy efficiency and building renovation plans. The results document significant annual variations over the years that can significantly influence the projections of primary energy savings and emphasize the need to consistently calculate, at least annually update, and officially publish the relevant conversion factors at both national and EU levels. This is most relevant in the era of the electrification and decarbonization of buildings and other sectors.

2. Calculation Methods

The PEFs are not among the data that are consistently published or readily available by Eurostat or other agencies, institutions, or departments of the European Commission. Accordingly, EU member states can introduce a different PEF in their energy and climate assessment plans and other relevant studies, such as the national cost-optimal exercises and long-term renovation strategies, provided that there is sufficient justification. As a result, there may be inconsistencies that may in some cases lead to diverging and confusing outcomes. These discrepancies were also documented by a recent literature review, which revealed no consensus on how to calculate or use the conversion factors [23].

A landmark study that was performed on behalf of the European Commission provided valuable insights into the different calculation approaches given that the estimation of the PEF is characterized by a high level of complexity, while even the official EU calculation approach has not been very transparent [24]. This study eventually inspired the development of a European standard on a harmonized methodology for the calculation of the PEF for electricity [25].

Recognizing the calculation complexities, the question that motivated this work was to consider approaches that are based on simple-to-access information, from officially published data on an annual basis and available in electronic format by Eurostat [1]. The data can be easily tracked through the energy balances for specific end uses with flow (Sankey) diagrams [26] and the interactive tool that can be used to easily handle the complete national energy balances [27].

The primary energy consumption refers to the transformation input for electricity-only producers and electricity production from combined heat and power (CHP) units. This covers all inputs into the transformation plants that are modified to produce electricity and/or heat [28]. The final electricity consumption refers to the delivered electricity to all sectors, excluding the use by the energy-producing industries and distribution losses. This way, the relevant calculations can be easily performed and verified with annual updates in order to properly quantify and even track the evolution of PEFs.

Accordingly, the work reported herein considers three simple definitions (PEF1-3), using the following equations. The first (PEF1) is a simple ratio of primary energy consumption (PEC_p) used as transformation input for electricity generation to the final electricity consumption (FEC_el) delivered for all end-use sectors. The abbreviations given in the descriptive definitions for the various parameters that appear in the following equations are based on the common framework of the EU regulation for energy statistics in the European Community [29]. The primary energy consumption is the sum of the transformation input for electricity-only units and the allocated total transformation input for CHP producers used for electricity generation. The simplified calculation for the allocation of the CHP input is equivalent to the ratio of the electricity from CHP producers to the total CHP output (i.e., electricity and heat production) that are consistently published and readily available in [1]. The final electricity consumption is the amount of energy delivered to all sectors, which already excludes the distribution and transmission losses, self-consumption, or other statistical differences. The other two equations account for the net imports, either as an additional resource to the primary energy (PEF2) or a reduction from the delivered electricity (PEF3).

$$\mathrm{PEF}1=\frac{PE{C}_p}{FE{C}_{el}}$$

(1)

$$\mathrm{PEF}2=\frac{PE{C}_p+Net Imports}{FE{C}_{el}}$$

(2)

$$\mathrm{PEF}3=\frac{PE{C}_p}{FE{C}_{el}-Net Imports}$$

(3)

The main advantage of the three Equations (1)–(3) is that they are straightforward and simple to use and utilize data that are annually published in national energy balances [1,26]. This way, it is easy to replicate and consistently update on an annual basis without elaborate calculations. One would expect similar results from the three calculation approaches in the event that there were very small net electricity imports, such as in the case of an average value for the EU-27. On the other hand, in the event that there are large net imports, either for countries that are major electricity exporters (e.g., the net balance in France represents about 10% of the gross electricity generation and about 7% in Germany) or others that are major electricity importers (e.g., Luxemburg imports three times more electricity than it produces), the different approaches will result in diverging values and should be used with caution. In addition, the net imports do not fully capture the differences between importing “green” electricity versus exporting electricity that is produced mainly from a fuel mix that is dominated by fossil fuels.

The fourth calculation approach (PEF4) is based on the methods proposed by [24] and among the ones adapted in [25]. This is a more complex method, in line with the Eurostat primary energy calculation, while the allocation for the electricity generated from CHP follows the accounting method used by the International Energy Agency that splits the primary energy input to the CHP outputs according to their output energies.

$$\mathrm{PEF}4=\frac{PE{C}_p-PE{S}_{p,CHP} }{GEG \left(100-SC\%-GL\%\right)}$$

(4)

where PES_p,CHP is the primary energy share for the heat output in CHP that works as a heat bonus during power generation, GEG is the gross electricity generation that corresponds to the output from all types of power plants and is adjusted for the net imports, SC% is self-consumption, and GL% is the grid distribution and transmission losses that are expressed as a percentage of the gross electricity generation. The amount of primary energy that is allocated to the electricity generated from the CHP power plant corresponds to the conversion efficiency of the electricity process divided by the conversion efficiency of the entire CHP process. Given that the thermal efficiency of the heat generation process is significantly higher than the efficiency of the electricity process, the result of this approach is that the heat generation is attributed a larger share of primary energy. The denominator in Equation (4) refers to the net electricity consumed.

3. Results

This section presents representative information for the available data and results from the calculations performed using the four approaches and the follow-up analysis. The representative results focus on the EU-27, Greece, and Poland. Similar elaboration has been performed for all EU-27 member states. The detailed results are presented in Figure A1.

3.1. PEFs Data

The results from the calculation of the primary energy factors (PEF1-4) using the four calculation methods (Equations (1)–(4)) cover the period starting from 1990, which marks the first year for which relevant data are available. Representative results are illustrated in Figure 2. The data are used to fit simple linear regressions that are elaborated in the following section. The results from the standardized approach using PEF4 are presented for the entire time series of available historical data (PEF4-h) and for the most recent data during the most recent decade (PEF-10y), covering the period 2009–2019.

The overarching results for EU-27 clearly illustrate that all calculation approaches are consistently very close to each other. The first three (PEF1-3) simplified calculation approaches are practically the same in that there are very small net electricity imports for the whole European Union. For example, EU-27 reported only 0.12% net electricity imports in 2019 [1].

Diverging results are observed on a national basis in the event that there are large net electricity imports. For example, Greece has an energy balance that is heavily dependent on electricity imports, reaching 19.8% of the total final electricity used in 2019, from an average of only 2.7% in the 1990s [1].

On the other hand, Poland was historically a net electricity exporter (averaging −3.2% in the 1990s) until recently. Over the past 4 years, there has been a reversing trend of Poland importing more electricity, reaching 7.6% of the total final electricity use in 2019. The upward shift of PEF4 is attributed to the dominant role of CHP in the national power generation and its attributes in the calculation approach. The gross electricity generation covered by CHP was historically around 97% in the 1990s and reached 86% in 2019 [1].

Given all the results from the EU-27 member states presented in Figure A1, similar apparent discrepancies are observed between the simplified calculation approaches (PEF1-3) and the standardized approach (PEF4) for some countries. The following discussion provides a brief insight utilizing the available information for the energy balances by country [1,26]. For a group of nine countries (i.e., Bulgaria, Czechia, Denmark, Finland, Hungary, Latvia, Poland, Slovakia, and Slovenia) there is a clear and consistent upward shift in the calculated values with PEF4, compared with the others. A common attribute that can explain some of these variations is the dominant role of CHP. Under these circumstances, one should exercise caution when using the standardized calculation approach or consider alternative paths of handling the CHP allocation [24,25]. For example, in Denmark, CHP generated over 90% of the electricity in the 1990s and currently around 42%, with the rest coming from renewable energies (basically wind) or imports. Bulgaria averaged 22% from CHP in the 1990s down to 12% in 2019, while reporting very high transformation losses of 37%. Finland and Hungary are predominantly net importers of electricity, from 12.4% and 11.9% in the 1990s to 24.5% and 31.2% in 2019, respectively. For Romania, the notable dropping trend of the results using PEF4 is attributed to the significant contribution of electricity from renewables that ranged from 27% in the 1990s to 42% in 2019, along with the large transformation losses of 33% in the 1990s to 24% in 2019, while the electricity from CHP averaged 55% in the 1990s and dropped down to 13% in 2019. Combined with the practically constant variations of PEF1-3 as a result of the very small net imports and small energy-use variations (e.g., a standard deviation of 2210 GWh during the past decade that averaged 61,745 GWh gross electricity generation), the results from the four approaches have practically converged in recent years. For Croatia, the role of CHP has increased from 14% in the 1990s to 28% in 2019, and the imports have grown from 56% in 1990 to 70% of final electricity use in 2019, while exports have increased from 7% of generated electricity in 1990 to 41% in 2019.

Very close agreement between all four calculation approaches is observed for nine countries, specifically Austria, Belgium, Cyprus (no electricity imports/exports), Germany, Greece, Ireland, Italy, Portugal, and Spain. For France, only the results from PEF1 exhibit a notable differentiation among the estimated values, since it does not account for the net imports. The results from Malta are consistently the same among the four calculations till 2015, which marks the year of the first electricity imports, by connecting the island to the European main grid through the Malta–Sicily interconnector. Malta imported 50% of its electricity needs in 2015, reaching 72% in 2016, and it remained stable, averaging 31% during the past 3 years. For Sweden, the divergence of PEF4 results has been more evident since 2004, while the upward trend over the past decade can be attributed to the increased exports from 6.6% in 2009 to 20.9% in 2019. The Netherlands exhibits very small variations over the years for all four methods.

Finally, there is a small group of countries (i.e., Estonia, Lithuania, and Luxemburg) for which there are unique features resulting in erratic variations that may mandate a closer look at the available national data. Large differences between electricity imports and exports are not realistically captured by net imports and differences of importing “green” electricity versus exporting electricity that is generated from mainly fossil fuels. In particular, Estonia generated all of its electricity from CHP and was a major exporter (−36.3%) in the 1990s, but currently, CHP contributes only 19% and has transformed into a major importer of electricity (29.5%). However, looking at the evolution of future projections, there is a clear converging trend among the different methods. Lithuania was a net exporter, but as of 2010, electricity imports have dominated the national energy flow by about three times. In this case, the results are presented for only the gross electricity generation. Finally, Luxemburg has some unique features, with limited electricity generation capacity from renewables and natural gas with CHP; it has provided some limited exporting capabilities since 2000, which climbed to 75% in 2006; and it imported 88% of its electricity in 2019.

3.2. PEFs Correlations

Another challenge that is encountered while developing national energy and climate plans or performing an impact assessment on building renovation measures is that they have a long lifetime that runs over several years and even extends to several decades. This type of work will require elaborate background work in order to make detailed forward-looking models and data analysis for the evolution of the energy mix for power generation, efficiency improvements, etc. On the other hand, national studies and assessments of different energy-saving measures and scenarios are typically performed by making a simplified assumption for a constant conversion factor to avoid the complexities of estimating and documenting the evolution and future of PEFs.

However, the performance assessment of different scenarios using a default (outdated) PEF or even recently updated values may result in different decisions that unfavorably treat the use of electricity several years into the future. For example, energy performance assessments of different scenarios and technologies for buildings renovation have been performed in Greece using the default value of 2.9 for the PEF of electricity that was derived from data from 15 years ago. This means that all electrically driven HVAC equipment were unfavorably treated during the recommendation assessments for building energy certification, e.g., replacing a fuel fired boiler with a heat pump. Currently, the PEF in Greece is estimated at 2.1, which is about 30% lower. Furthermore, given that this kind of a renovation will have a lifetime of 15 or 20 years, it is not sufficient to use a more recent value for the PEF (i.e., 2.1) looking forward and making decisions that will carry into the next decade. Instead, one needs to use a more realistic estimate for the future PEFs that would be representative during the equipment’s operational period.

To bridge the knowledge gap and overcome the complexities for the elaborate modeling of national energy systems, this work derives simple regressions based on historical trends and the past decade that reflect the evolution of the national energy mix and policies. This provides an easy-to-handle approach for forward-looking and more-realistic near- and long-term projected estimates. The results are summarized in

Table 1. Linear regressions for the PEFs and goodness-of-fit measure (R²) for the available national data (Figure A1) using the four calculation methods.

EU Member State	Simplified Approach			Standardized Approach
	PEF1	PEF2	PEF3	Historical Data (PEF4-h)	Past 10 Year Data (PEF4-10y)
Austria	y = −0.0113x + 24.204 R2 = 0.788	y = −0.0058x + 13.187 R2 = 0.607	y = −0.0028x + 7.2634 R2 = 0.182	y = −0.0106x + 22.945 R2 = 0.698	y = −0.0163x + 34.355 R2 = 0.403
Belgium	y = −0.0339x + 70.676 R2 = 0.727	y = −0.0296x + 62.229 R2 = 0.818	y = −0.023x + 49.013 R2 = 0.919	y = −0.0341x + 71.072 R2 = 0.750	y = −0.0569x + 117.11 R2 = 0.349
Bulgaria	y = 0.0104x − 16.656 R2 = 0.444	y = −0.0018x + 7.6578 R2 = 0.002	y = −0.0282x + 60.194 R2 = 0.671	y = −0.0342x + 73.448 R2 = 0.581	y = −0.0766x + 158.59 R2 = 0.728
Croatia	y = −0.0166x + 34.728 R2 = 0.336	y = −0.014x + 29.67 R2 = 0.431	y = −0.0171x + 36.275 R2 = 0.432	y = −0.0244x + 50.655 R2 = 0.504	y = −0.014x + 29.679 R2 = 0.251
Cyprus	y = −0.0321x + 67.533 R2 = 0.850	y = −0.0321x + 67.533 R2 = 0.850	y = −0.0321x + 67.533 R2 = 0.850	y = −0.0304x + 64.103 R2 = 0.846	y = −0.0355x + 74.306 R2 = 0.716
Czechia	y = 0.0133x − 22.897 R2 = 0.213	y = 0.0033x − 2.954 R2 = 0.031	y = −0.0166x + 36.52 R2 = 0.788	y = −0.0077x + 19.911 R2 = 0.137	y = −0.0591x + 123.4 R2 = 0.767
Denmark	y = −0.0394x + 80.738 R2 = 0.608	y = −0.0315x + 64.746 R2 = 0.785	y = −0.0279x + 57.582 R2 = 0.951	y = −0.0436x + 90.124 R2 = 0.530	y = −0.0949x + 193.5 R2 = 0.816
Estonia	y = −0.0032x + 10.84 R2 = 0.002	y = 0.0053x − 6.6245 R2 = 0.008	y = 0.0115x − 19.558 R2 = 0.037	y = −0.0887x + 183.08 R2 = 0.679	y = −0.0866x + 179.17 R2 = 0.171
Finland	y = −0.0153x + 32.51 R2 = 0.592	y = −0.011x + 24.006 R2 = 0.619	y = −0.0077x + 17.557 R2 = 0.626	y = −0.0099x + 22.212 R2 = 0.308	y = −0.0341x + 71.057 R2 = 0.785
France	y = −0.0165x + 36.73 R2 = 0.767	y = −0.0137x + 30.962 R2 = 0.771	y = −0.0067x + 16.67 R2 = 0.463	y = −0.0154x + 34.359 R2 = 0.681	y = −0.0219x + 47.559 R2 = 0.499
Germany	y = −0.0317x + 66.383 R2 = 0.966	y = −0.0351x + 73.161 R2 = 0.976	y = −0.0396x + 82.066 R2 = 0.978	y = −0.0263x + 55.773 R2 = 0.942	y = −0.0297x + 62.745 R2 = 0.670
Greece	y = −0.0493x + 101.63 R2 = 0.915	y = −0.0443x + 91.729 R2 = 0.924	y = −0.0372x + 77.674 R2 = 0.902	y = −0.0342x + 71.593 R2 = 0.891	y = −0.0376x + 78.055 R2 = 0.538
Hungary	y = −0.0397x + 82.245 R2 = 0.751	y = −0.0318x + 66.542 R2 = 0.847	y = −0.0183x + 39.884 R2 = 0.583	y = −0.0592x + 122.04 R2 = 0.821	y = −0.0969x + 197.85 R2 = 0.890
Ireland	y = −0.0482x + 99.094 R2 = 0.926	y = −0.0472x + 97.143 R2 = 0.943	y = −0.0461x + 94.919 R2 = 0.950	y = −0.0445x + 91.748 R2 = 0.923	y = −0.028x + 58.399 R2 = 0.583
Italy	y = −0.0189x + 40.059 R2 = 0.899	y = −0.0199x + 42.075 R2 = 0.901	y = −0.025x + 52.598 R2 = 0.895	y = −0.0021x + 6.4902 R2 = 0.074	y = −0.0113x + 24.869 R2 = 0.607
Latvia	y = 0.0084x − 15.764 R2 = 0.205	y = −0.0062x + 13.687 R2 = 0.273	y = −0.0274x + 56.517 R2 = 0.268	y = −0.009x + 20.053 R2 = 0.085	y = 0.0131x − 24.4 R2 = 0.118
Lithuania	y = −0.0484x + 101.67 R2 = 0.075	y = −0.1368x + 277.64 R2 = 0.653	y = 0.025x − 46.704 R2 = 0.190	y = −0.0488x + 102.07 R2 = 0.388	y = −0.1358x + 277.14 R2 = 0.619
Luxemburg	N/A	y = −0.0058x + 12.983 R2 = 0.153	y = 0.0311x − 60.151 R2 = 0.114	y = −0.019x + 39.974 R2 = 0.324	y = −0.0098x + 21.632 R2 = 0.108
Malta	y = −0.121x + 246.22 R2 = 0.736	y = −0.1097x + 223.65 R2 = 0.762	y = −0.089x + 182.38 R2 = 0.732	y = −0.1202x + 244.68 R2 = 0.738	y = −0.3076x + 622.01 R2 = 0.821
Poland	y = −0.0209x + 44.576 R2 = 0.968	y = −0.021x + 44.727 R2 = 0.973	y = −0.0211x + 45.059 R2 = 0.978	y = −0.0242x + 51.216 R2 = 0.976	y = −0.0551x + 113.94 R2 = 0.984
Portugal	y = −0.0236x + 49.429 R2 = 0.550	y = −0.0234x + 49.071 R2 = 0.681	y = −0.0236x + 49.59 R2 = 0.736	y = −0.018x + 38.355 R2 = 0.478	y = 0.0199x − 37.926 R2 = 0.222
Romania	y = 0.0051x − 7.3927 R2 = 0.032	y = −0.0011x + 5.0718 R2 = 0.002	y = −0.0121x + 27.077 R2 = 0.202	y = −0.0814x + 167.33 R2 = 0.936	y = −0.0949x + 194.62 R2 = 0.902
Slovakia	y = −0.0252x + 53.337 R2 = 0.487	y = −0.0266x + 56.182 R2 = 0.736	y = −0.0332x + 69.472 R2 = 0.791	y = −0.0182x + 39.836 R2 = 0.316	y = −0.0269x + 57.196 R2 = 0.597
Slovenia	y = −0.0305x + 63.88 R2 = 0.810	y = −0.0261x + 54.923 R2 = 0.886	y = −0.0173x + 37.259 R2 = 0.732	y = −0.0315x + 66.352 R2 = 0.763	y = −0.081x + 166.08 R2 = 0.899
Spain	y = −0.0259x + 54.574 R2 = 0.875	y = −0.0261x + 55.046 R2 = 0.847	y = −0.0267x + 56.088 R2 = 0.780	y = −0.0292x + 61.216 R2 = 0.925	y = −0.0151x + 32.682 R2 = 0.365
Sweden	y = −0.0044x + 11.029 R2 = 0.091	y = −0.0096x + 21.419 R2 = 0.511	y = −0.0146x + 31.502 R2 = 0.719	y = 0.0089x − 15.273 R2 = 0.288	y = 0.0293x − 56.426 R2 = 0.580
The Netherlands	y = −0.0054x + 12.632 R2 = 0.189	y = −0.0082x + 18.507 R2 = 0.554	y = −0.0128x + 27.753 R2 = 0.740	y = −0.0052x + 12.705 R2 = 0.244	y = −0.0237x + 50.072 R2 = 0.443
EU-27	y = −0.0209x + 44.576 R2 = 0.968	y = −0.021x + 44.727 R2 = 0.973	y = −0.0211x + 45.059 R2 = 0.978	y = −0.0242x + 51.216 R2 = 0.976	y = −0.0251x + 52.997 R2 = 0.971

for all EU-27 member states, using the different calculation methods. In the case of the standardized approach (PEF4) the analysis is performed for the entire time series of available data, resulting in a linear regression with the historical data (PEF4-h), and one using the most recent data during the past ten years (PEF-10y). This past decade corresponds to a more ambitious trend given that in most cases it reflects the significant progress made in EU-27 to support the targets set by the introduction of the Renewable Energy Directive (RED) in 2009 and its recast in 2018 [30]. The relevant national data are illustrated in Figure A1 and can be used to interpret the correlation coefficients and future trends.

4. Discussion

A key element in the evolution of the PEFs has traditionally been the technological advances and improved efficiency of power generation plans. However, the most notable changes come as a result of the growing annual deployment of renewables in EU-27, especially since the introduction of RED in 2009 [30], reaching more than 22% of the gross final energy consumption in 2020 [1]. Even more ambitious is the proposed target for scaling up the uptake of renewables to 45% by 2030, in accordance with the REPowerEU plan [31]. The uptake of renewable energy will extend beyond central power generation by facilitating renewable energy communities and prosumers (the right to produce and consume renewable energy), stimulating its use for heating and cooling, the transport sector, etc.

4.1. Role of Renewables

The historical evolution of the energy mix for power generation is illustrated for the EU-27 and two example countries (Figure 3). In 2019, renewables covered 34.6% of gross EU-27 electricity generation (Figure 3a), increasing by 67.7% during the past decade, which is also reflected by a drop of about 10.9% among the four PEFs (Table 1). Greece reached a share of 33.2% in 2019 (Figure 3b) and exhibited a notable increase in the contribution of renewables by 91.5% during the period (2009–2019), with a decrease of about 19.9% in PEFs (Table 1). On the other hand, the share of fossil fuels in electricity generation remains high in some countries. For example, in Poland, the growth of renewables has been very slow in the past, and despite the notable growth by 182% during the past decade, it reached only 16% of gross electricity generation in 2019 (Figure 3c), which is still dominated by solid fossil fuels. This is also reflected by the slow change of PEF over the years, with a most notable decrease of about 18.1% (Table 1) during the past decade (2009–2019). In anticipation of an even-faster growth of renewables toward the target of a decarbonized Europe by 2050, the ambitious PEF4-10y correlations presented in Table 1 will be more representative to use for assessing future plans.

4.2. Comparison against Reported PEFs

The calculated PEFs are compared against officially reported factors in the EU, Greece, and Poland, with very good agreement that gives confidence to the results. While the year of publication of the relevant conversion factors is known, it is not evident or clearly stated from which chronological period the data used in the calculations were derived. One should keep in mind that usually the calculations are based on at least a couple of years’ old data, since it takes that long to release the official statistics and energy balances. Accordingly, the officially reported data should be compared against the estimates in the range of 2 to 4 years before the reported publication date.

The EU-27 default PEF value of 2.5 that was published in 2012 [6] compares very well with the estimated values using the correlations in Table 1 that average about 2.61–2.56 from PEF1-3 and 2.57–2.50 from PEF4, during the period 2009–2011. The updated value of 2.1 that was published in 2018 [7] appears to be a realistic and suitable near-term value for this decade and compares very well with the findings of this work (Figure 2a).

In Greece, the national primary energy conversion factors for all energy carriers were published with the 2010 legislation and regulation for EPBD transposition, setting a value of 2.9 for electricity [17]. This is in very good agreement with the estimated values resulting from the correlations presented in Table 1, which average 2.92–2.76 from PEF1-3 and 2.82–3.70 from PEF4, during the period 2007–2009 (also see Figure 2b). In Poland, the national conversion factor for electricity was calculated at 3.0 and published in 2015 [18], which is consistent with the estimated averages of 2.76–2.66 from PEF1-3 and 3.18–3.09 from PEF4, during the period 2012–2014 using the correlations from Table 1 and illustrated in Figure 2c.

These examples give confidence that the estimated values are realistic. There is no need to statistically quantify the exact differences, because it is not clearly reported how the reference values have been derived. This includes the national values and the EU-27 default value [24]. More than anything else, the approach for the calculation of the EU default value should be clearly documented so that it can be replicated as a reference. In any event, the intent of this comparison was to give confidence that the future projections are based on prudent historical estimates. It is expected that there will be future year-to-year uncertainties and deviations, but at least the historical estimates and the long-term projections appear to be reasonable.

4.3. Simplified Methods and Standardized Approach

This work is not trying to challenge the accuracy or relevance of the standardized method [25] but rather comparing the results from PEF4 with other, easier-to-use-and-implement methods. The differences among the four calculation approaches are negligible on the EU-27 level and for some countries, such as Austria, Cyprus, Greece, Ireland, and Spain (Figure A1). However, there are also notable inconsistencies between the four methods at the national level, such as in Bulgaria and in Denmark. Similar findings have also been reported even between elaborate calculation methods, especially in terms of how to treat renewables in the accounting process, concluding that no PEF calculation method can claim absoluteness [24].

For most countries, the straightforward calculations using the consistently reported annual data (PEF1-3) provide comparable results with the more detailed standardized approach (PEF4). However, in some cases, the results reveal notable deviations and should be used with caution. For example, overlooking the role of net imports (PEF1) will provide erroneous results in countries that are mainly electricity importers (e.g., Lithuania, Luxemburg). Similar problems will occur for countries that are mainly electricity exporters (e.g., Croatia, Estonia). When the more accurate but also more complex, standardized calculation (PEF4) is taken as a reference, the differences among the more simplified calculation approaches in 2019 range for PEF1 between −45.7% in Demark and 4.2% in Spain, for PEF2 between −55.5% in Lithuania and 14.8% in Malta, and for PEF3 between −33.3% in Denmark and 41.6% in Luxemburg, averaging 0.8% according to the EU-27 data.

This work developed simple correlations for all EU-27 member states that can be used as a first estimate of PEF evolution, in order to facilitate future projections and forward-looking plans. The annual variations of the PEF values have been most significant during the past decade, as a result of the growth of renewables in power generation. For example, given the results for PEF4-10y, the faster evolutions are projected in Bulgaria, Denmark, Lithuania, and Slovenia. This may be considered as a more ambitious future projection, compared with the correlations using all the historical data (PEF4-h).

The use of input data from different energy balances should be carefully addressed. The work reported herein is based on Eurostat data. However, energy balances from different reporting organizations (e.g., Eurostat, International Energy Agency, United Nations Statistics Division) may have deviations. This may be attributed to different calculation principles (e.g., use the physical energy content or the substitution method), conversion factors (e.g., net or gross calorific values for different energy carriers), or other definitions (e.g., treatment of electricity and heat in the transformation processes), among others [32].

4.4. Case Study

The practical implications of the present work and findings are demonstrated through the following case studies. They consider the assessment of energy-efficiency measures for improving the energy performance of the buildings sector, which constitutes a pillar of relevant efforts in most EU-27 national energy and climate plans, and they drive the policies and long-term strategies toward 2050.

In Greece, the buildings sector used 6.2 Mtoe, or 40.6%, of the total final energy consumption [1] and about 35 TWh, or 68%, of the total electricity available for final consumption in 2019 [33]. The average annual electrical energy–use intensity is 23–47 kWh/m² in residential buildings [34] and 17–145 kWh/m² in nonresidential buildings [35], depending on their typology that includes the building use (e.g., single-family houses, offices, hotels), construction period, and location-climate zone.

The implementation of various energy conservation measures (ECMs) can provide significant annual savings for electricity demand by up to 1.32 TWh for residential buildings [34] and 1.41 TWh for nonresidential buildings [35]. Using the default PEF value of 2.9 for Greece, this could translate to potential primary energy savings from all types of buildings as (2.73 × 2.9) TWh that are equivalent to a total of 0.68 Mtoe per year. This is consistent with the estimated new annual energy savings in the buildings sector of 0.79 Mtoe for the period 2021–2030 in the national energy and climate plan [36]. For a policymaker, these annual savings can be interpreted as a contribution of a 3.3% annual drop toward the 20.55 Mtoe target of total primary energy use in 2030.

However, this may give a false impression of the actual impact and the progress achieved toward the national and EU targets. For example, using the PEF4 linear projection, the value for 2025 is estimated at 2.34 using the fit to the historical data (PEF4-h) and at 1.92 using the more ambitious trend line from the recent decade (PEF4-10y). This could translate to potential primary energy savings from all buildings as (2.73 × 2.34) TWh that are equivalent to a total of 0.55 Mtoe per year for the historical PEF trend. Similarly, using the ambitious PEF trend, the primary energy annual savings become 5.23 TWh or 0.45 Mtoe. The results are illustrated in Figure 4 by using the different PEF values that were derived from the correlations presented in Table 1 to project the values for the corresponding periods. At the beginning, the annual energy savings in 2021 are overestimated using the default value by 31.2% compared to PEF1, 24.2% to PEF2, 14.0% to PEF3, 14.7% to PEF4-h, and 28.8% to PEF4-10y. Higher differences will be observed by 2030, reaching 46.5% compared to PEF1, 37.9% to PEF2, 25.6% to PEF3, 25.3% to PEF4-h, and 40.4% to PEF4-10y. It is evident that the cumulative primary energy savings will fall short of the national target and distort the effectiveness of ECMs in the buildings sector.

In Poland, the buildings sector used 28.8 Mtoe, or 40.1%, of the total final energy consumption [1] and 78 TWh, or 52%, of the total electricity used in 2019 [33]. Coal will continue to play a major role in the Polish energy market for many years to come [37], which is also reflected by the trends of the solid fossil fuels in the energy mix (Figure 3c) and the low slope of the PEF correlations (Figure 2c). The goal is for renewables to cover 32% of Poland’s electricity consumption by 2030.

The national plan for improving energy efficiency has set a target for 2030 to reduce the primary energy consumption by 23%, or 27.3 Mtoe, reaching 91.3 Mtoe [37]. However, this is characterized by modest ambition in its contribution to the EU target [38]. Even under this conservative scenario, the renovation of residential and nonresidential buildings is expected to have a significant contribution toward meeting the target of about 118.6 Mtoe annual primary energy savings. For example, the aim is to increase the share of residential buildings that are thermally insulated from 58.8% in 2015 to 70% in 2030. The recently published detailed long-term renovation strategy [39] will significantly contribute to this process. Accordingly, it is estimated that approximately 7.5 million thermomodernization investments will be carried out by 2050, including 4.7 million deep renovation projects and step-by-step renovation measures spread over time. The strategy assumes an average annual rate of thermomodernization of approximately 3.8%, assuming that 65% of the buildings will achieve an annual primary energy-use intensity of less than 50 kWh/m² by 2050.

During this decade (2021–2030), the final energy consumption in Poland is expected to decrease in residential buildings by 4.1 Mtoe, while the savings in nonresidential buildings are expected to reach 1.2 Mtoe. In this direction, the plan is to reduce the use of coal in dwellings by replacing inefficient and manually fed boilers for space heating, reducing their number from 18.5% in 2015 to less than 7.5% in 2030.

The use of electricity is expected to reach 21.7% of the energy carriers used by 2030 [37]. Given an equivalent breakdown of the total final energy savings in the buildings sector (5.3 Mtoe), the allocated annual electricity savings can reach 1.15 Mtoe per year. During the 2021–2030 period, using the default PEF value of 3.0 for Poland, this translates to potential primary energy savings of 3.45 Mtoe per year and an annual contribution of a 3.8% drop toward the 2030 target of 91.3 Mtoe primary energy use [37]. On the other hand, this will be an overestimated value using the projected PEFs from the correlations presented in Table 1 and illustrated in Figure 5. For example, using the PEF4-h correlation, the projected 2025 conversion factor is 2.21, which translates to primary energy savings of (13.38 × 2.21) TWh, or 2.54 Mtoe, in 2025 (Figure 5). The more ambitious projection uses PEF4-10y results in a value of 2.36, which translates to primary energy savings of 31.6 TWh, or 2.72 Mtoe, in 2025 (Figure 5).

The different annual primary energy savings are illustrated in Figure 5 as a result of using the different PEF values calculated from the correlations presented in Table 1. Annual savings are overestimated when using the default conversion factor by 22.1% compared to PEF1, 23.8% to PEF2, 19.5% to PEF3, 23.1% to PEF4-h, and 13.9% to PEF4-10y. Differences will be higher by 2030, reaching 28.4% compared to PEF1, 30.1% to PEF2, 25.8% to PEF3, 30.3% to PEF4-h, and 30.4% to PEF4-10y. As a result, the cumulative primary energy savings will not achieve the anticipated savings from the buildings sector or contribute toward meeting the national target.

4.5. Future Directions

Accounting for the origin of the imported electricity is an issue in terms of how it was generated. This is even more relevant given that the EU is aiming to improve cross-border electricity interconnections. Security over electricity supply is one facet, but higher interconnections are also advantageous in terms of facilitating the integration of more renewables into the power grid. In this direction, the EU has set an interconnection target of at least 15% by 2030 [40]. This will further encourage imports and exports among neighboring countries. PEFs tend to increase when imports are taken into account, as notably illustrated in the cases of Luxemburg (from 1.3 to 1.8) and Latvia (from 1.4 to 1.7) [41]. In other cases, they decrease, such as in Hungary, where the PEF drops from 2.5 to 2.2.

Marginal conversion factors to estimate emissions will soon be a very important issue as renewables with high variability play a major role in the energy mix for electricity generation [42], along with grid stability, demand response, and load manipulation, such as EV charging during specific hours of the day [41]. However, with the exception of the transmission system operator in Denmark and France, which release hourly data, there is very limited, if any, availability of hourly conversion factors for other EU member states [23]. Therefore, a simplified marginal approach may be more manageable [43]. For some applications, such as the energy performance assessment of buildings, it may be more practical to pursue and derive seasonal PEFs. They may better match the assessment of different measures and scenarios by using technologies and energy carriers that are targeted for heating (winter) or cooling (summer) periods. In addition, they correlate better with seasonally dependent power production from renewables. For example, higher solar radiation availability in summer will lead to higher electricity output from photovoltaics. In Italy, the seasonal variations of PEF range from 1.5 to 2.36, the higher share of renewables during the summer resulting in a median value as low as 1.8, while during winter, the PEF values rise up to 2.1 [44].

5. Conclusions

The paper reviewed different calculation approaches for estimating the primary energy factor for electricity, using official annual statistics for EU-27. The analysis included the historical evolution of the calculated conversion factors, which revealed a clear and significant drop of PEF values, in most countries. This reflected the positive impact from the increasing use of renewables for electricity generation throughout Europe. The main objective of the work was to derive simple linear correlations for projecting the PEF values, which can then be used to facilitate more-realistic forward-looking calculations. The estimates of future conversion factors become even more relevant for assessing different technologies for the electrification of buildings in national energy efficiency, climate change, or decarbonization plans in EU-27 member states. Two case studies, for Greece and Poland, were used to illustrate the potential impacts on the assessment of electricity savings from energy-efficiency measures in the buildings sector. For example, the quantified cumulative primary energy savings may significantly deviate from the results using the national default PEF values by up to 40% in Greece and 30% in Poland. By using the proposed correlations, it is possible to easily make more-realistic annual calculations of the PEFs for adapting the estimates of future primary energy savings. This will give more confidence to national authorities and policymakers on the actual progress toward meeting national targets.