Chalmers Publication Library Building-Stock Aggregation through Archetype Buildings: France, Germany, Spain and the UK This document has been downloaded from Chalmers Publication Library (CPL). It is the author´s version of a work that was accepted for publication in: Building and Environment (ISSN: 0360-1323) Citation for the published paper: Mata, É. ; Sasic Kalagasidis, A. ; Johnsson, F. (2014) "Building-Stock Aggregation through Archetype Buildings: France, Germany, Spain and the UK". Building and Environment, vol. 81 pp. 270â282. http://dx.doi.org/10.1016/j.buildenv.2014.06.013 Downloaded from: http://publications.lib.chalmers.se/publication/199226 Notice: Changes introduced as a result of publishing processes such as copy-editing and formatting may not be reflected in this document. For a definitive version of this work, please refer to the published source. Please note that access to the published version might require a subscription. Chalmers Publication Library (CPL) offers the possibility of retrieving research publications produced at Chalmers University of Technology. It covers all types of publications: articles, dissertations, licentiate theses, masters theses, conference papers, reports etc. Since 2006 it is the official tool for Chalmers official publication statistics. To ensure that Chalmers research results are disseminated as widely as possible, an Open Access Policy has been adopted. The CPL service is administrated and maintained by Chalmers Library. (article starts on next page) Building-Stock Aggregation through Archetype Buildings: France, Germany, Spain and the UK Corresponding author É. Mata Division of Energy Technology Department of Energy and Environment Chalmers University of Technology, Gothenburg, Sweden email: [email protected] Desk phone : +46 (0)31 772 52 50 Fax : +46 (0)31 772 35 92 Postal address : Hörsalsvägen 7b, 4th floor SE- 412 96 Gothenburg, Sweden Other authors A. Sasic Kalagasidis, Department of Civil and Environmental Engineering Chalmers University of Technology, Gothenburg, Sweden email: [email protected], F. Johnsson Department of Energy Technology Chalmers University of Technology, Gothenburg, Sweden email: [email protected] ABSTRACT (150 - 200 words) Knowledge regarding the characteristics of national building stocks is fundamental to understanding how the energy performance of the building stock can be improved. To facilitate large diversity and a number of buildings for such analyses, this paper presents a methodology by which national building stocks may be aggregated through archetype buildings. The methodology has been implemented and verified in four EU countries in regions with different climates, namely France, Germany, Spain and the UK. These countries account for about half of the final energy consumption of the EU-28 buildings. The analysis includes the residential and non-residential sectors (residential sector only for Germany). The number of archetypes per country has been defined according to different categories of building type, construction year, climate region and the main fuel source for heating purposes. The accuracy of the description is validated by simulating energy demand using the ECCABS Building Stock Model, and comparing the final energy demand modelled with corresponding statistical data. The total final energy demand calculated for these countries differs from available statistics by between -6% and +2 %, which is considered satisfactory. The proposed description of the building stock is being used as a basis for analyzing the potential for and cost of energy conservation measures. KEYWORDS archetype buildings, EU building stock, energy demand, ECCABS Model, energy simulation 1 1. INTRODUCTION 1.1. Background Knowledge regarding the characteristics of building stocks is fundamental to understanding how the energy performance of the building stock can be improved [1][2]. What are the size and structure of the building stock in the European Union (EU)? Are there sufficiently robust data for the buildings in each member state (MS) upon which to base studies of the energy use of these building stocks? Kohler and Hassler [1] used the German building stock as a case study and concluded that most studies are strongly limited by the absence of reliable statistical data, and international research confirmed the global scale of this knowledge gap [3]. Similar conclusions have been reached by other analysts [4][5][6][7]. Nevertheless, despite the apparent paucity of consistent data, there has been a surge in recent years in the development and use of energy consumption models depicting national building stocks [8]. For the purpose of modelling its energy demand, a building stock can be described through representative buildings in terms of sample buildings or archetypes [2]. Sample buildings are designated as representing actual buildings (with data obtained from measurements) of the existing stock. As the building stock of a country consists of buildings with varying characteristics, an extensive sample of the buildings is required in order to derive an accurate picture of the thermal characteristics of the building stock. Therefore, establishment of such a sample would require significant efforts towards measuring and quantifying the parameters of a building sample. However, archetypal buildings are statistical composites of the features found within a category of buildings in the stock [3] derived from available data of the national building stock. Thus, despite the paucity of real building data, archetype buildings can be derived. Although these descriptions do not include all relevant parameters for determining the energy demand, the literature provides descriptions of the building stock for several EU countries according to certain categories of samples and/or archetypes (cf. for France: [9][10]; for Germany: [11][12]; for Spain: [13][14]; and for the UK: [15][16][17][18][19]). For instance, only heated floor areas or the number of buildings are provided. On an EU level, there are examples of a more comprehensive description of a set of existing buildings [20], including the noteworthy effort to gather relevant data in the European TABULA Project [21], which has recently mapped data of existing residential buildings of 13 MSs. However, TABULA [21] has only very recently presented a quantification of the frequency in the building stock of the existing buildings described (no quantification is given in [20]). Therefore, despite such buildings constituting so-called typologies, they are not representative in a way that would allow an extrapolation of the description of the buildings to represent all residential and non- residential buildings in each of the EU MSs investigated. Besides different use and building traditions, a description of the building stock should reflect the climate zone where the building stock is located. The studies by Ciscar [22] and Tsikaloudaki [23] propose that the climates within EU could be represented by three or five zones, respectively, dependent on whether only heating [22] or both heating and cooling [23] degree days are calculated. To the authors’ knowledge, no studies representative of the EU have described the building stocks of these zones climatically. Thus, there is much additional work to be done in order to arrive at a description of the characteristics of EU buildings on a level which may serve as a basis for modelling the effect of improvements in energy performance. Buildings in the EU are a particularly interesting object of study, not only because the EU has a concrete regulatory framework and a set of energy and environmental 2 targets, but also because the building sector accounts for 35%–40%1 of the final total energy consumption (25%–27% residential, 10%–13% non-residential) in EU-28 and 25%–40% of the associated carbon dioxide (CO ) emissions2 (15%–27% residential, 11%–21% non- 2 residential), as shown in Figure 1. 45% 40% 35% 30% 25% Non-residential Sector 20% Residential Sector 15% 10% 5% 0% 7 r r r r 2-UE otces gnidliuecnarF ni otces gnidliuynamreG ni otces gnidliuniapS ni otces gnidliudetinU nimodgniK B B B B Figure 1. Percentage of residential and non-residential buildings in the final total energy consumption (i.e. all sectors) of the EU-28 and selected MSs. 2011 data from the Eurostat database [24]. This study assesses the possibility of describing the EU building stock for the purpose of forming a basis for analyzing the effect and costs of applying different energy conservation measures (ECMs) to the entire EU building stock3 starting with key member states. More specifically, the aims are as follows: A. To review the existing building stock data for EU, identifying key issues and main data gaps for the purpose of defining archetype buildings B. To describe a methodology for building stock aggregation applicable to selected EU countries and to the residential and non-residential sectors C. To apply and evaluate the methodology by describing and aggregating building stocks of four selected EU countries, then comparing the modelled energy use in the stocks with the statistical data available. The four MSs selected for the study include France, Germany, Spain and the UK which together account for 58% and 52%, respectively, of the final energy demand and CO 2 emissions of buildings in EU-28, as shown in Figure 2. Since they belong to each of the EU climate zones given in [22][23], they are also representative of all EU climates. In accomplishing all three aims, a step forward is made by means of structuring and compiling 1 All the ranges given in this paragraph cover the percentages given in the databases [24][34] for EU-27 (as a total) and for the six most densely populated countries: France, Germany, Italy, Poland, Spain and the UK. Thus, there may be MSs for which the percentages differ from the ranges given here. 2011 data from the Eurostat database [24]. 2 Year 2011 data from the Odyssee database [34]. 3 The authors are currently working on an analysis of the potentials and costs for energy saving and corresponding CO emission reductions for Spain [113] and on similar studies for the three other countries 2 investigated. Since these works are not yet available, for an application of the modeling methodology to analyze energy use of residential buildings in an EU country not included in this paper (Sweden), see [114]. 3 the information, as well as defining a methodology to be used in building stock modeling with the above geographical and sectorial scope. 15% Building sector in France Building sector in Germany 42% Building sector in Spain 21% Building sector in United Kingdom Building sector in the Rest of EU 9% 13% Figure 2. Share of the buildings of the four selected MSs in the final energy consumption of the EU-28 building sector (12831 TWh for residential and non-residential). 2011 data from the Eurostat database [24]. The following section of the paper addresses Aim A by reviewing building stock data and availability, providing a framework of definitions to be used throughout the paper. Section 2 presents the methodology for building stock aggregation, which is implemented in four selected EU countries; thus, this section addresses Aims B and C. Section 3 focuses on Aim C by giving a cross-country comparison of the building stock with respect to the building characteristics determining the energy demand. Section 4 discusses the results and Section 5 summarizes the findings with respect to the three aims cited above. 1.2. Building stock data and availability Building stocks are generally divided into Residential (R), also called domestic or household sector, and Non-Residential (NR) buildings, also known as the tertiary or services sector. In the R sector, due to the constant transferring between these categories [25], allocations of main, secondary residences and vacant units are difficult to analyze. However, since there has been greater political attention paid to the R sector, especially in social housing [1], there are generally better statistics and knowledge on the R sector than on the NR sector. The NR sector has until recently mostly been documented in terms of isolated buildings for technical or cultural reasons (public buildings, industrial monuments, etc.), with the primary focus on buildings that are perceived as individually outstanding works of art [1]. Information available on the building stock has typically been gathered through two basic approaches:  Census data, compiled for the establishment of a register of new building construction statistics. This type of register includes all new buildings but generally provides only basic information on the stock based on the designed construction project, such as the use of the building, the number of buildings or area. Such information is typically reported in national and subsequently international statistics. 4  Surveys, which are additional studies of the existing buildings through a number of selected buildings carried out with a specific purpose. Surveys provide a wide range of post-occupancy information about buildings, including their technical characteristics, fuel usage and occupant behaviors. Such information is required for the categorization of the stock and is a prerequisite for modelling energy use in the building stock. As a consequence, the countries that have assessed energy-saving and CO -mitigation 2 potentials have conducted major surveys of their building stocks, e.g., the UK [26], Scotland [27], Belgium [28], and Sweden [29]. Individual billing data and sub-metering may also be available and can be used as complementary information for characterizing building stocks. In recent years, energy certificates [30][31][32] and other information linked to geographical information systems [33] have appeared as additional means of data gathering, although mostly on a regional scale. Data from the above approaches is compiled on the EU and international levels and three readily accessible databases provide data on the building sector, namely Eurostat [24], ODYSSEE [34], and the Greenhouse Gas and Air Pollution Interactions and Synergies (GAINS) Database [35]. These databases are updated on a regular basis and their main contents are summarized in Table 1. It is clear from the table that Odyssee provides the most complete dataset, including energy consumption for R and NR buildings per fuel and end use. However, not all data categories are included in the database. Eurostat provides only final energy consumption for R and NR buildings and is quite similar to corresponding data by Odyssee. The GAINS Database also includes data disaggregated by single- and multi-family dwellings, as well as data projections according to some scenario-based assumptions. In all cases, data on CO emissions are only given on an aggregated level; e.g. Eurostat only gives 2 total greenhouse gas emissions (GHG) (in CO eq) with no details for the building sector. 2 Another database, the MURE II Policy Database [36], provides information on energy efficiency policies and measures carried out in the MSs of the EU, thus compiling all building regulation codes in force in each MS. In addition to these databases, some European projects (e.g. EPA-NR [37], EL-TERTIARY [38], and TABULA [21]), have compiled the information available on the building stock of a given country or set of countries. Ó Broin [39] has mapped the available data, indicators, and models related to the energy demands of European buildings. Pérez-Lombard et al. [6] have reviewed the data on energy use in buildings worldwide over the past 30 years (not continuously but in snapshots when the information became available). The recent BPIE Data Hub [40] is an open data portal for statistical data on the European building stock and includes all countries. Data are to be presented according to categories, such as building stock inventory, building stock performance and climatic zones, as well as different building types and owner profiles. However, the current BPIE Data Hub Portal still does not provide data for all categories. Although these studies and databases listed in Table 1 provide a valuable overview of the EU building stock, until recently the data could not be used directly for an archetype-based modelling of the energy performance of the stock because of the lack of physical description of the buildings required for such an analysis and because there has been no quantification of the number of buildings in the building stock. However, TABULA and Data Hub are now updated on a regular basis and their contents have significantly increased by the end of 2013. Yet, these data were not available at the time when the present work was carried out (2010-2013) and, thus, could not be applied here. 5 Table 1. Summary of international data sources for the energy consumption levels and characteristics of the European building sector. Odyssee[34] Eurostat[24] GAINS[35] Years included From 1980 From 1990 2005-30 Building characteristics Stock of dwellings Yes(1, 2, 3, 4) Yes(1, 3) Total floor area Yes(1, 3, 5, 6) Yes(1, 3) Floor area of dwellings (average) Yes(1, 3) Energy consumption levels Final energy consumption, residential sector Yes(7, 8, 9) Yes(9) Yes(3, 10, 11) Final energy consumption, tertiary sector Yes(5, 7, 9) Yes(9) Yes(8, 10, 12 ) CO emissions 2 CO emissions Yes(6) Yes 2 Total CO emissions (with electricity) Yes(6) Yes 2 (1)Data provided disaggregated into single-family dwellings and multifamily dwellings. (2)Data provided disaggregated into individual/collective central heating and room heating, as well as oil/coal/gas/district heating/electric/wood space heating. (3)Data provided disaggregated into existing and new. (4)Only permanently occupied dwellings. (5)Data provided disaggregated into hotels/ restaurants, health and social actions, education/research, administration, private services, offices and trade (wholesale and retail). (6)Data provided disaggregated into households and services. (7)Data provided disaggregated into space heating, hot water and cooking. (8)Data provided disaggregated into coal, oil, gas, heat, wood and electricity. (9)Data provided also with climatic corrections. (10)Data provided disaggregated into space heating and hot water, lighting and appliances. (11)Data provided disaggregated into cooling and heating. (12)Data provided disaggregated into cooling, heating and ventilation. 2. METHODOLOGY FOR BUILDING STOCK AGGREGATION The description of a building stock through archetype buildings proposed in this study follows four steps: segmentation, characterization, quantification and validation of the final energy demand in the building stock for a reference year. For the latter, a detailed and dynamic Building Stock Model is used which was previously validated as described in [41]. Data were compiled through several surveys conducted on a country basis, for which corresponding reports are available for downloading [42][43][44][45]. 2.1. Segmentation In the segmentation process, the number of archetype buildings required to represent the building stock of the country is determined. The number of archetype buildings is obtained from the combination of the different segmentation criteria:  building type, defined from the use of the building, its layout (one or several floors) and the way it is attached to neighbouring buildings (e.g. detached, semi-detached or terrace houses);  construction year, determined from the updates of the building regulation codes but also according to historical events and changes in construction technologies;  main heating system; and  climate zone (within a country), defined in accordance to the climate zoning suggested for winter periods in the building regulation codes. Meteorological data from the most densely populated city in the climate zone is considered representative of the climate in the zone and is generated by Meteonorm [46]. 6 It is estimated that these criteria provide a good representation of the energy demand of the buildings. Besides, they facilitate data compilation (i.e. matching the form of data sources) as identified in the literature [9][13][14][20][47][48][49]. The segmentation was applied to both R and NR buildings in all countries studied, except for Germany where the R stock alone was included. Table 2 summarizes the archetype buildings obtained for each county. Table 2. Segmentation categories and resulting number of archetype buildings obtained for the four countries investigated. C, commercial; L, leisure; MFD, multifamily dwelling; NR, non-residential; O, office; R, residential; SCL, sports, culture and leisure; SFD, single-family dwelling; T, terraced house; X, other. Categories France Germany Spain UK 99 archetypes 122 R archetypes 120 archetypes 252 archetypes Building 3 R[10]: SFD, 5 [54]: SFD3, prefabricated 2 R: SFD, MFD; type private and public houses, T, MFD4, MFD <10 4 NR: C, SLC, 6 R [56]: detached, semi- MFDs; floors, MFD >10 floors5 O, X. detached, T, bungalow; 3 5 NR [50][51]: C, NR [57]: O, warehouses, E, H, O, SLC1 retail10 Construction 3 R [10]: before 10: before 19186, 4: before 1975, 7 [58][59][60][61]: before year 1975, before framework buildings before 1975-79, 1980- 1985, 1986-1991, 1992-95, 1975 refurbished, 1918, 1919-48, 1949-57, 2005, 2006-089 1996-2002, 2003-06, 2007- after 1975; 1958-68, 1969-79, 1980-83, 10, and after 2010 3 NR [52][53]: 1984-94, 1995-2002, 2003- before 1977, 09 1977-2000, after 2000 Heating 2 R [10]: electric Non-applicable7 Non-applicable7 2 [56]: central and non- system heating, other central11 source for heating2 Climate zone 3 (H1to H3) [52]: 3: Essen, Stuttgart, Munich 8 5 (A to E) [55]: 4 [62]: London, Paris, Toulouse, Málaga, Seville, Birmingham, Newcastle, Marseille Barcelona, Glasgow Madrid, Burgos (1) Which account for 80% of the final energy use of the NR sector in this country. (2) For NR buildings there are no data available that would allow a division by category with respect to the sources of energy for heating purposes. (3) Including two-family dwellings. (4) With maximal four floors and eight to ten apartments. (5) Mostly from the 1960s and 1970s. (6) Except framework buildings. (7) Data availability did not allow a segmentation of the building stock according to their type of heating system. (8) Found by adjusting the four recommended zones for calculating the heating load after the DIN EN 12831 to the political division of the territory in sixteen states [33]. (9) This last period of construction is only applicable to the NR sector. (10) These NR building types are known as the Valuation Office’s bulk classes and cover about 70% of all ratable NR buildings, but exclude for instance hospitals, schools, churches, etc. (11) Assumed that premises which currently have non-central heating system are all built before 1985. The average internal temperature for centrally heated dwellings is assumed to be 17.5 °C while it is 14°C for non-centrally heated premises. 2.2. Characterization In the characterization step, each archetype is described by its technical characteristics (i.e. 23 inputs as listed in [41]) based on the parameters from the segmentation provided in Table 3. The four countries include a total of 593 archetype buildings. Although all inputs and assumptions are of relevance, such a matrix of data is rather extensive and would be too 7 comprehensive to fit into a journal paper. Instead, a summary of a common procedure for gathering all the data as well as the sources used, on a country basis, is provided. In addition, selected country-specific values of the data are also provided for the purpose of understanding the levels of detailing as well as the differences between the countries. Further details on the characterization of the archetypes, including values for all inputs, are given in the corresponding country reports [42][43][44][45]. Table 3. Segmentation categories and relevant building data depending on the category. Categories Relevant building data dependent on the category Building type Effective heat capacity of the building Floor area External surface area Internal gains Minimum desired indoor temperature Maximum desired indoor temperature Sanitary ventilation rate Window area Construction year Average U-value of the building Ventilation rate Heating system Indoor temperatures Fuels used Climate zone Average U-value of the building Outdoor climate data Reports from official authorities responsible for dwellings (e.g. Ministry of Dwellings/Energy/Environment in the four countries) provide information on the physical characteristics of the buildings. National building energy codes have been used to determine the indoor conditions and thermal properties of the building envelope, i.e., the period of construction is translated into U-values according to the building codes on a country-basis (seen mainly from the construction year in Table 2). The reliance on the building codes implies that the buildings have been constructed in each period of time according to the regulations and that they are operated according to their technical requirements; nevertheless, the literature does not provide any systematic check on regulatory compliance. The nature of the task is data intensive; thus, the references in this section are numerous but provided for transparency; furthermore, mostly national sources and other branch-specific journals have been used to obtain building data. Unknown surfaces of envelope (such as facades, cellars, roofs and windows) and building geometry are obtained according to the so-called 3CL-DPE Method [63], based on the floor area heated, the number of levels/floors, assumptions of the building form and attachment to neighbouring buildings’ of each building as defined by the 3CL-DPE Method. The fuel use is allocated to the R and NR archetypes separately. In all four countries, it has been possible to separately allocate different fuels and technologies for space heating and hot water. Both for the fuel types and the efficiencies of the different heating technologies, data were found only at level of aggregation that would not allow for differentiation of each archetype. For French buildings, U-values are determined for the different construction periods according to building regulations; for the R sector from [64][65][66] and for the NR sector from [67]. There are no data for the U-values of the oldest R and NR buildings per climate zone, and the U-values of the H3 climate zone are assumed to be 15% higher than those of the other two zones [68]. Ventilation rates for R buildings are between 0.23 and 0.51 l/s/m2 as 8 extracted from [67], and for NR buildings these are between 0.37 and 1.47 l/s/m2 from [43] [68][69]. Since no data have been found for France, natural ventilation rates for R buildings are 0.40 l/s/m2 as calculated according to the 3CL Method [10] and for the NR buildings from [43]. Heat gains from occupants are calculated considering metabolic gains at low activity (i.e. sitting) [70], except for the sports, culture and leisure (SCL) buildings where heat gains from occupants correspond to higher activity levels. Heat gains from lighting, equal to their electricity consumption, for R buildings are calculated by adding 3W/m2 to each 100 lux of installed illuminance [67], with a minimum illuminance of 300 lux required in households (since the heated floor area of the household is different for each archetype, the heat gain also differs between archetypes). For NR buildings, heat gains from lighting range from 3.1 to 11.5 W/m2 and are calculated from the lighting consumption [71] assuming the operating schedules and occupancy patterns shown in [43]. Heat gains from appliances in R buildings are 2.1 W/m2 for an average SFD of 103.8 m2 and 3.4 for an average MFD of 66.0 m2 [21]. For NR buildings, gains from appliances range from 1.1 to 4.7 W/m2 [44][72]. Due to the lack of data, building materials in France are assumed to be similar to those of Catalan buildings because of their proximity and common building technology [73]. The distribution of energy sources within R buildings is taken from [74], where the percentages of each heating system per climate zone are specified. Thereafter, these values have been separated into SFD and MFD according to the division proposed by [75]. Regarding NR buildings, the percentages for climate zone H1 are taken for offices (O) and commercial (C) buildings from [76], and for educational (E) buildings from [77]. The remaining archetype buildings have the same average shares of 47.5% electricity, 28.1% gas, 19.8% oil, and 4.6% other fuels as provided by [78]. Data regarding the efficiency of the boilers installed in the buildings were lacking at a national level; therefore, average values were considered based on [79]. For German R buildings, all U-values are taken from [54]. In Germany, natural ventilation is prevalent for supplying fresh air to the buildings. The share of buildings with a mechanical ventilation system is less than 1.5% and only around half of them have heat recovery (the latter is disregarded). Additionally, 10% of the newly constructed buildings are equipped with ventilation systems [80]. Sanitary (natural) ventilation rates4 are between 0.60 ACH and 0.85 ACH, depending on the construction year. Values for heat gains from appliances and occupancy range, respectively, were between 1.78 W/m2 and 2.66 W/m2 and between 1.11 W/m2 and 2.00 W/m2 for the different building types according to [81]. Heat gain from lighting is 0.27 W/m2 for all buildings, calculated from the energy use for lighting per person and per year [82], as well as the average number of person per household given in Eurostat [24] and the average floor area of a German dwelling [80]. Hot water demand is between 1.16 W/m2 and 3.38 W/m2 for the different building types. The distribution of different types of heat and hot water generators/producers was acquired from [80]. The efficiencies of the heating and hot water systems were extracted from [83]. For Spanish buildings, the period of construction was translated into U-values according to the building codes [55][84][85]; for buildings constructed before the implementation of the first thermal regulation in 1975, the average U-value is based on [20][86]. Ventilation rates for R buildings are between 0.42 and 0.51 l/s/m2 as taken from [55] and for buildings constructed before 1975 these are taken as the infiltration rate of 4 ACH reported in [55]; for NR buildings, between 0.55 and 0.83 l/s/m2 from [86]. Heat gains from occupants, lighting 4 Here, ventilation rates are given the units provided in the original reference (ACH or l/s/m2). Transformation between these two units requires some assumptions with respect to the volume of indoor air (thus heated floor area) that would distort the original data. 9