Energy-consumption-for-space-heating-of-West-German-households-empirical-evidence,-scenario-projections-and-policy-implications.pdf

corresponding author tel 0049 711 780 6151 fax 0049 711 780 3953 e mail address cw ier uni stuttgart de c weber 1formerly at ier energy policy 28 2000 877 894 energy consumption for space heating of west german households empirical evidence scenario projections and policy implications andreas schuler 1 christoph weber ulrich fahl bewag aktiengesellschaft puschkinallee 52 d 12435 berlin germany institute for energy economics and the rational use of energy ier universita t stuttgart he bru hlstr 49a d 70565 stuttgart germany received 22 may 1997 abstract both the technical characteristics of buildings e g insulation standards and the utilisation patterns of households are important determinants of the space heating demand of private households therefore this paper examines to what extent socio economic characteristics of households household size social position income etc are suitable to explain household s behaviour concerning space heating given technical building characteristics on the other hand energy e ciency factors derived from technical simulation models are scrutinised the empirical basis of this analysis is the o cial household survey 1988 for west germany containing about 44 000 data sets therefrom it turns out that the german insulation standards of buildings have been less e ective in the past than assumed on the basis of technical simulations by means of scenario calculations projections of the heating energy demand for space heating in the future are obtained sketching the e ect of di erent possible socio economic developments on the space heating consumption and analysing di erent political options concerning building insulation standards for new buildings and the existing building stock 2000 elsevier science ltd all rights reserved keywords space heating scenarios germany standards energy e ciency 1 introduction in nordic and moderate climatic zones a considerable part of total energy consumption is used for space heat ing purposes for example in west germany more than one third of the total nal energy demand is devoted to space hating and this share amounts even to 75 in the residential sector vdew 1998 hence measures to increase energy e ciency in this eld may contribute substantially to achieve political goals of energy conser vation and climate gas emission reduction in germany therefore during the last two decades regulations on insulation and heating system e ciency have been tight ened several times see table 1 although the stipulated standards increased there is almost no control to see whether standards actually are met properly further the regulations issued so far are only valid for newly con structedbuildings and in the case of major renovations of existing buildings the object of this paperis to analyse empiricallyon the basis of a large scale sample the e ects of standards issued in the past on heating energy consumption there by mainly the question is addressed whether the sample energy consumption is properly predicted by existing technical simulation models ebel et al 1990 or whether other in uencing factors such as utilisation behaviour or improperly ful lled standards have to be accounted for furthermore also the e ects of socio economic factors such as income and age structure on dwelling size and thus indirectly on heating energy consumption are inves tigated from these analyses a series of scenarios for the future evolution of residential heating energy consump tion is deduced thereby scenarios are di erentiated by the socio economic developments assumed by the extent of retro tting in the existing dwelling stock and by the degree of ful llment of the newly issued regulations the paper is organised as follows rst the data base available for the empirical analysis is brie y characterised 0301 4215 00 see front matter 2000 elsevier science ltd all rights reserved pii s0 3 01 4 2 1 5 0 0 0 0 0 74 4 table 1 building regulations concerning space heating ebel et al 1990 regulationsincekmin w m2k wallwindowroofloftcellar ceiling din 4108 19521 561 460 801 01 din 4108 19691 561 100 801 01 din 4108 19741 560 890 680 83 wa k rmeschutzv 19771 063 50 450 450 80 din 4108 bach 1993 gu k lec et al 1994 have developed and used the rst approach to investigate the present heating energy consumption of the west german building stock and its possible future evolution an empirical validation of the models used has however if at all only been performed on the basis of small samples the strength of these models certainly is the detailed evaluation of the e ects of technical measures such as double glazing or roof insulation on the heating energy demand of existing or new buildings however the validity of the behavioural and technical assumptions for the existing building stock has to be scrutinised em pirically 2 especially the technical simulation models do not allow for an analysis of factors in uencing energy relevant aspects of dwelling utilisation and choice e g room temperature dwelling size knowledge of these factors is however crucial if scenarios of future heating energy consumption are to be developed therefore in the following an empirical investigation of households heating energy consumptionis carried out using an existing large scale sample besides avoiding the data collection e ort the use of an existing survey pres ents the advantage of limiting the possible non respon dent and observation biases in particular detailed observation of heating behaviour may lead to a much more conscious choice of room temperature and airing behaviour one suitable existing sample is provided by the house hold survey conducted in 1988 by the german statistical o ce stat bundesamt 1988 this type of large sample 878a schuler et al energy policy 28 2000 877 894 3in particular the sample does not contain households of foreigners nor households with a monthly net income of more than 25 000 dm furthermore one person households and households with low income are underrepresented whereas families and households with higher income but below 25 000 dm per month are overrepresented the statistical o ce provides the so called projection factorsa that should enable a correct projection from the sample on the whole population except the categories not represented at all but these have to be used with precaution in particular when applying inferential statistical methods the signi canceof variablescan not be deduced from a sample corrected by the projection factors 4as observed correctly by an anonymous referee in germany never areal large scaleenergy focusedhousehold surveyhas beencarriedout in various data bases limited information are available but these o er no comprehensive basis for analysis table 3 socio demographic variables variabledescription hhshousehold size 1 2 3 4 5 6 person s inchousehold net income per year in dm ageage of reference person 1below 25855 57 225 29958 59 330 341060 62 435 391163 64 540 441265 69 645 491370 and more 750 54 empnumber of gainfully employed persons in the household 0 1 2 3 4 and more employed persons socsocial position of reference person 1farmer5worker 2self employed6 temporarily unemployed 3o cial7non wage earner 4employee munsize of municipality 1below 20 000 inhabitants3100 000 and more inhabitants 220 000 100 000 inhabitants distdistance from place of residence to next city with more than 100 000 inhabitants 00km distance to next city225 to below 60km 1below 25km3more than 60km also di erent income components labour income capital income etc as well as other aggregates gross income have been included in the data set yet they have not been used for the analyses survey original sample size approximately 44 000 house holds is only carried out every ve years with the results of the 1993 survey still not being fully available the survey covers in principle all west german households although it is not fully representative 3 table 2 shows observed relevant variables for the buildings and heating systems the data base contains observations for build ing characteristics as well as six household character istics listed in table 3 a severe limitation of the data base is that the energy consumption for space heating hot tap water and appliances is not observed separately by each application but only the total energy purchases by energy carriers is registered furthermore the age of dwellings is aggregated to vintage classes and there is no information available whether dwellings have been refur bished since they were built dueto thelonglifetimeof buildings thedwellingstock development is relatively slow compared with stocks of most other energy consuming durables e g cars appli ances therefore technical simulations often focus on the in uence of the vintage of buildings furthermore the building type detached house multiple dwelling building tower block etc is important for the energy demand of a building as speci c transmission losses energy loss per dwelling size depend on the ratio between building sur face and dwelling size within the data base used six vintage classes and four building types are distinguished cf table 2 a building class is then de ned as combina tion of vintage class and building type table 4 shows the distributions of dwellings among building classes and the corresponding shares of building classes in the sample many dwellings are located in buildings with three or more dwellings buildings of the type ri ds i 2 where fepuriis the nal energy purchased by energy carrier and ds ithe changes in the energy stock of the household by energy carrier energy storage can be neglected for grid dependent energy carriers thus electricity and gas consumptions of a certain year are identical with the quantity of electricity respectively gas purchased in the same year 3 1 1 energy consumption for appliances it is obviously impossible to calculate an exact amount of energy used for appliances for each household how ever this demand can be eliminated relatively easy by excluding data sets using electricity as heating energy carrier since almost all appliances except some stoves consume solely electricity the e ect of stoves was elimi nated in all samples by excluding all data sets using the same energy carrier for cooking and heating 3 1 2 energy consumption for tap water for the analyses described here an average electricity demand for tap water heating as a function of household size and energy carrier used was applied vdew 1992 by assumptions concerning the annual overall e ciency of electrical tap water heaters p tzner and scha k fer 1994 it is possible to calculate the average useful elec tricity demand for tap water heating related to the house hold size ectw fectw elga el 3 where uectwis the useful energy consumption for tap water fectw elthe nal energy consumption for electri cal tap water heating and ga elthe average annual e ciency of electric tap water heaters assuming independence of the useful energy demand from the energy carrier used makes it possible to calculate the nal energy demand as a function of house hold size and energy carriers fectw ec ectw ga ec 4 where fectw ecis the energy carrier dependent nal energy consumption for tap water and ga ecthe energy carrier dependentaverage e ciency of tap water heaters the annual e ciency used for gas and coal boilers was derived from prognos 1992 3 1 3 energy consumption for space heating assuming that the energy purchased is approximately identical to the energy consumed and applying the as sumptions described above as well as the calculated values for tap water consumption to the data base it is possible to determine the energy demand for space heat ing for each household according to fecsh hh fep rhh fectw hh 5 where fecsh hhis the nal energy consumption for spaceheating by a speci c household fepurhhthe nal energy purchased by a speci c household and fectw hh the nal energy consumption for tap water of a speci c household 3 1 4 expected heating energy consumption as detailed information on technical standards of dwellings lack in the data base used the expected heating energy consumption has to be determined using useful energyfactors these factors are relatedto vintageclasses and building types vintage class and building type stand here as proxies for the insulation standard and for the ratio between dwelling surface and dwelling size insula tion standard and the ratio mentioned are important determinants for transmission losses and consequently also for the energy demand the useful energy factor indicates the useful energy demand of a building mea sured per square metre dwelling size dependent on the building class factors have been published for buildings under original conditions and considering average reno vation of the buildings for instance certain replacement rates for low standard windows by well insulated win dows cf table 6 multiplied by the average dwelling space of the appro priate building class and divided by energy carrier dependent average annual overall e ciency of the heat ing system geiger and wegner 1989 5 the useful energy factor uef leads to the average nal energy demand for space heating fecsh of the respective building class fecsh efbc dsizbc ga ec 6 where fecshis the nal energy consumption for space heating uefbcthe building class dependent useful a schuler et al energy policy 28 2000 877 894881 table 6 useful energy factors according to ebel et al 1990 vintage class building type kwh a before 1919 before 1919 1919 19481949 19571958 19681969 19781979 19831984 1987total dh245188202258146143119120174 rh 20016916716115012194156 ah1931661581911741189876155 lh 150168154153120 136 tb 105120 136 ebel et al describe the german building stock in detail the energy conservation caused by renovations is calculated with a simulation model on the bases of empirical data and assumptions based on the german insulation standards half timbered house solidly built house average value weighted by dwelling size calculated commonly for lh and tb dh detached house 1 or 2 dwellings rh row house ah apartment house up to 41 2 oors lh large apartment house 5 8 oors tb tower block above 8 oors 6another mathematical approach similar to the one described in eq 8 is laid down in the guideline vdi 2067 vdi 1991 the utilisation intensity there is expressed as product of six correction factors specifying deviations from the average utilisation and heating system the factors are meant to describe additional or reduced heat demand for airing reduced heating time restricted heating space deviations of the heating utilisation concerning the room temperature regulation of radiators and heating system as well as additional heat sources instead of useful energy factors vdi 2067 uses a detailed analysis of the technical standard and characteristics of the building din 4701 din 1983 to calculate the solely technical heat demand however this requires very detailed information on the building as thisinformation mostly lacks for a large number of buildings vdi 2067 is not used to analyse building stocks but mainly for single buildings the objective of analyses based on the vdi methodology is often to makethe costs of the heating systemof a buildingmore transparentand thus to support the owners when deciding on heating system and building related investments 7in 1988 the numberof degree days was 2692 in west germany the average number of degree days is 3030 schi er 1992 as a result the climate correction term is 1 126 unfortunately variations of the cli mate within germany cannot be observed as for data secrecy require ments there is no information available on the location of the sample buildings within germany energy factor dsizbcthe building class dependent aver age dwelling size in m2 and ga ecthe energy carrier dependent average e ciency of heating systems instead of using the average dwelling size of a building class dsizbc the dwelling size of each household dsizhh might be used to determine the nal energy demand of a particular household as both useful energy factor and annual e ciency of the heating system re ect only average conditions the calculated nal energy de mand of a speci c household does not correspond to theactual nalenergydemandofthesamehouseholdinthe data base determined according to eq 6 therefore the utilisation intensity ui has to be introduced into eq 7 to get the nal energy demand for speci c house holds fecsh hh this leads to the following equation 6 fecsh hh c efbc dsizhh ihh ga 7 where c is the climate correctionterm fecsh hhthe nal energy consumption for space heating by a speci c household dsizhhthe dwelling size of a household in m2 and uihhthe utilisation intensity of a household the utilisation intensity accounts for deviations of the average utilisation of a dwelling of the respective class caused e g by the chosen temperature of rooms the heating period the heated space and airing however it includesalso technical deviations from the average build ing concerning e g insulation orientation of the build ing size of windows and heating system characteristics finally it considers thedivergencebetweentheclimate of the building s location and the average climate in west germany by the climate correctionterm c thevariation of the climate of the year observed from long term aver age conditions can be eliminated 7 inserting eq 5 into eq 7 and solving for uihhleads to ihh ga fep rhh fectw hh c efbcdsizhh 8 hence an estimate of the utilisation intensity uihh may be obtained for each household if the useful energy factor of the building is known there is a detailed de scription of german useful energy factor available split into ve building types and seven vintage classes ebel et al 1990 disadvantageously the building types and vintage classesdo not tally exactly withthe building classes of the 882a schuler et al energy policy 28 2000 877 894 table 7 useful energy factors adapted to vintage classes registered in the sample vintage class building t