A novel residential heating consumption characterisation approach at city level from available public data: Description and case study

A no el esiden ial hea ing consump ion cha ac e iza ion app oach a ci y le el om a ailable public da a: desc ip ion and case s udy Jon Fe nandez1, Luis del Po illo1, I án Flo es1 1Dep . o The mal Enginee ing. Uni e si y o he Basque Coun y, Plaza Ingenie o To es Que edo 1, 48013 Bilbao, Spain, Phone: 34-94 6014028 e-mail: [email p o ec ed] ABSTRACT In he cu en ansi ion owa ds a ne -ze o GHG emissions economy, esiden ial building sec o is one o he sec o s wi h highes po en ial o ene gy e iciency. In o de o assis he implemen a ion o ene gy sa ing policies, in which he ole o local au ho i ies is c ucial, he bo om-up enginee ing app oach is he mos sui able me hod o a eliable cha ac e iza ion o he ene gy consump ion in an exis ing building s ock. In he p esen pape a me hodology is p oposed o cha ac e ize he space hea ing ene gy consump ion o a esiden ial s ock a ci y le el, based on a che ype buildings. The me hodology is ocused on add essing he challenges which o en such app oach in ol es. An e idence-based calib a ion p ocedu e is p oposed o add ess occupan beha iou al pa e ns. A en ion is also gi en o he cons uc ed models esul s alida ion p ocess, by compa ing hem wi h collec ed eal da a. The me hodology is applied o he esiden ial building s ock o he ci y o Bilbao, whe e se en een ypologies o a che ype buildings a e cons uc ed. The esul s demons a e he p ac icabili y o accu a ely ep oducing he exis ing space hea ing ene gy consump ion o a ci y-scale esiden ial building s ock wi h a ailable public da a sou ces. The me hodology has been concei ed o be easily eplicable o any ci y. Keywo ds: Building s ock modelling; A che ype buildings; Residen ial sec o ; Ene gy e iciency; Space hea ing 1. In oduc ion Recen ly he a e age concen a ion o CO2 in he a mosphe e eached 403 ppm, which is a alue abou 40% highe han in he mid-1800s [1]. Based on scien i ic e idence, human-induced global wa ming has al eady eached 1°C abo e p e-indus ial le els, and is inc easing a app oxima ely 0.2ºC pe decade [2]. So he end is clea . Clima e change is a se ious conce n ha dese es immedia e ac ion. In his con ex , he Eu opean Union (EU) has aken a ole o global leade ship in educing g eenhouse gas (GHG) emissions, ecognising o a long ime clima e change as an issue whe e cohe en EU ac ion is needed. The Eu opean Council con i med in Feb ua y 2011 he EU’s objec i e o educing GHG emissions by 80-95% by 2050 compa ed o 1990 [3]. In he Eu opean s a egic long- e m ision o a p ospe ous, mode n, compe i i e and clima e neu al economy [4], he EU a i ied i s in e es in wo king owa ds a ne -ze o GHG emissions economy by mid-cen u y. In his ansi ion owa ds a compe i i e low ca bon economy, building sec o has been iden i ied as one o he sec o s wi h he highes po en ial o ene gy e iciency and GHG emission educ ion. Indeed, in he EU he building sec o causes 36% o he g eenhouse gases (GHG) and is esponsible o 40% o he EU’s inal ene gy consump ion. One eason o hese high a es is ha abou 35% o he EU’s buildings a e mo e han 50 yea s old [5]. Un o una ely, he eplacemen a e o exis ing buildings in Eu ope is low, accoun ing o app oxima ely 1 – 3% pe yea [6]. The e o e he in luence o new buildings in ene gy use and emission educ ion is limi ed, which makes he cu en building s ock a main a ge o ene gy sa ing policies. In o de o implemen such policies, and de e mine p io i y a ge s ha equi e mobilisa ion o in es men , a eliable cha ac e iza ion o he ene gy consump ion dis ibu ion ac oss he building s ock is essen ial. To his e ec , se e al me hods ha e been p oposed, which acco ding o wo in-dep h e iews [7] [8] can be di ided in o wo main app oaches: op-down and bo om-up. The op-down app oach add esses he building s ock ene gy consump ion in an agg ega ed way, by assessing he impac o di e en ac o s, mainly socioeconomic, on i and de e mining hei e ec . On he o he hand, bo om-up me hods a e de eloped om da a o disagg ega ed elemen s, whose impac on ene gy consump ion o a ep esen a i e se o buildings is combined, and hen ex apola ed o he en i e building s ock. Wi hin bo om-up This is he accep ed manusc ip o he a icle ha appea ed in inal o m in Ene gy and Buildings 221 : (2020) // A icle ID 110082, which has been published in inal o m a h ps://doi.o g/10.1016/j.enbuild.2020.110082. © 2020 Else ie unde CC BY-NC-ND license (h p:// c ea i ecommons.o g/licenses/by-nc-nd/4.0/) me hods, s a is ical (SM) and enginee ing me hods (EM) a e dis inguished, depending on he na u e o hei inpu da a. SM a e based on his o ical consump ion da a, while EM ely on building physics de ini ion and simula ion [7]. Each me hod p o ides esul s wi h di e en applicabili y. Bo h op-down and bo om-up SM a e capable o conside in hei esul s he beha iou o occupan s, which is always unp edic able and a ies widely. Ne e heless, hese me hods do no each he le el o de ail which is equi ed o iden i y he speci ic unde lying building elemen s wi h he la ges ene gy e iciency imp o emen po en ial. This in o ma ion, which is c ucial o es ima e how di e en ene gy sa ing measu es could a ec he cu en ene gy consump ion, can only be p o ided by bo om-up EM. Howe e , hese me hods ha e also some limi a ions which mus be o e come o each us wo hy building models. Ka gic e al. [8] concluded ha one o hei main weaknesses is he inhe en unce ain y associa ed wi h many model inpu s, especially hose ela ed wi h he unp edic able na u e o occupan beha iou . Fu he mo e, bo om-up EMs equi e la ge amoun o inpu da a, bo h o cons uc he models and o alida e hem, and o en s udies a e limi ed by he absence o publicly a ailable da a. Thus, model inpu s- ela ed assump ions, which need o be based on au ho ’s expe ience when enough suppo ing da a is no a ailable, could lea e he models wi h oo many deg ees o eedom and lack o c edibili y. The e o e, impo an ocus should be gi en o he eliabili y o he models. Bo om-up EMs ypically classi y an exis ing building s ock in o a se o e e ence o p o o ypical buildings ha cons i u e he modelled domain, whose esul s a e la e ex apola ed o he en i e s ock. The e a e wo main app oaches: i) he selec ion o a sample o eal buildings [9], and ii) he de elopmen o a che ype buildings [10], also called syn he ic o i ual buildings. A che ype buildings a e heo e ical buildings c ea ed by a composi e o se e al cha ac e is ics ound wi hin a ca ego y o buildings wi h simila a ibu es [11]. Po en ially, he use o ac ual building samples could mo e ai h ully e lec he he e ogenei y o an exis ing building s ock han a che ype buildings. Ne e heless, a deep conside a ion o such high deg ee o di e si y could equi e a la ge sample o ep esen a i e buildings. O he wise, an expe analysis is equi ed o a oid ha con ex ual ea u es o he chosen sample jeopa dize he ex apola ion o he modelling esul s. In he end, he choice be ween hese op ions could be made based on expe enqui ies, s a is ical da a a ailabili y, e c. [12] [13]. Se e al au ho s ha e de eloped bo om-up building s ock models o mul iple Eu opean coun ies [14][15] [16][17]. In addi ion, se e al e o s such as TABULA and EPISCOPE p ojec s [18] ha e ied o also con ibu e o a comp ehensi e desc ip ion o exis ing building s ocks a na ional le el. Howe e , wo king wi h such la ge building s ocks necessa ily limi s he ep esen a ion o hei he e ogenei y. Indeed, a single coun y can ha e mul iple clima ic zones, which a ec esiden ial ene gy consump ion. T adi ional cons uc i e solu ions can also di e be ween di e en egions, which in u n could ha e been de eloped in di e en his o ical pe iods and a chi ec u al s yles. E en signi ican ly unequal socioeconomic ac o s can exis be ween e i o ies. All o his makes necessa y o analyse a speci ic building s ock ene gy consump ion om a much close pe spec i e. On he o he hand, in he implemen a ion o ene gy e iciency policies, local au ho i ies ha e a c ucial ole o play. Wi h his in mind, in 2008 he Eu opean Commission launched he Co enan o Mayo s [19], he mains eam Eu opean ini ia i e which in ol e housands o local go e nmen s olun a ily commi ed o implemen EU clima e and ene gy objec i es. In his sense, he de elopmen a ci y le el o bo om-up building s ock models could adequa ely desc ibe he he e ogenei y o he building s ock and se e as a basis o p edic he po en ial ene gy sa ings and ou line measu es o Sus ainable Ene gy and Clima e Ac ion Plans. Recen imp o emen s in geo- e e enced da a a ailabili y and geog aphical in o ma ion sys ems ha e allowed g ea e ocus on u ban con ex s [20], encou aging in ensi e and de ailed spa io- empo al assessmen s. Thus, U ban Building Ene gy Modelling (UBEM) [21] is a g owing ield in building ene gy modelling which accoun s o no only he dynamics o indi idual buildings bu also o he in e -building e ec s and u ban mic oclima e, co e ing a spa ial scale up o an en i e ci y [22]. UBEM is o en based on sophis ica ed me hods which can combine buildings physics modelling, s a is ical in e ence in GIS en i onmen s, CFD p og ams o accoun o u ban clima ic condi ions and ene gy supply simula ion modules [23] [24] [25], which in ol es g ea e da a managemen equi emen s and compu a ional esou ces. In ac , he de elopmen o ci y-scale da abases o he exis ing building s ocks is a c ucial s ep o UBEM o au oma ically gene a e ene gy models o u ban buildings. Building da a needs o UBEM ypically include he GIS oo p in , building heigh , numbe o s o ies abo e g ound, numbe o s o ies below g ound, o al loo a ea, hea ed loo a ea, numbe o dwellings, yea o cons uc ion, yea o e u bishmen , building ype, hea ing sys em ype, annual elec ici y use, annual na u al gas use, e c. [26]. Al hough mo e and mo e ci ies in he wo ld a e making hei building s ock da a publicly accessible a hei open da a po als [27], i is no always a ailable o is di icul o collec . While ex ac ing geome ic building da a om GIS da abases is a ela i ely s aigh o wa d p ocess [23], conside ing explici ly, bu oughly, each u ban building could lead o he gene aliza ion o he building models, losing accu acy on he dynamics o he building ene gy demand [28]. On he o he hand, he conside a ion o he spa ial (geog aphical) dimension on ene gy analyses is c ucial o he op imiza ion o ene gy supply sys ems and he assessmen o ela ed s a egies as peak sha ing, load shi ing, e c. [29]. Ne e heless, demand-side analyses e alua ing o e all annual ene gy sa ings po en ial o e an exis ing building s ock does no necessa ily equi e i . To his e ec , a new me hodological app oach has been de eloped by he au ho s o cha ac e ize an exis ing building s ock a ci y le el, aiming o se a comp ehensi e basis o u he ene gy e iciency in es men mobiliza ions assessmen s. The p oposed me hodology lies on a bo om-up enginee ing app oach, which pe mi s o cons uc and assess a che ype building models ha ep esen an en i e building s ock. In summa y, he p oposed me hodology se es o (a) classi y an exis ing ci y-scale esiden ial building s ock in o a se o ep esen a i e a che ype buildings and, (b) eliably cha ac e ize hei space hea ing ene gy consump ion. The aim is no o ocus on he geog aphical dimension pe spec i e a he han o ai h ully accoun o he dynamics o a che ype buildings models, eaching a success ul ade-o be ween he equi ed de ail o inpu da a and desi ed p ecision. Special ca e has been gi en o i s applicabili y o mul iple ci ies, as models become ac ually use ul only i hey a e sui able o being easily eplicable. The app oach has been concei ed o be applicable o any ci y, p o ided a p ima y unde s anding o he cons uc i e e olu ion in he analyzed geog aphical con ex . Al hough he implemen a ion desc ibed he ein is ailo ed o he Spanish con ex , simila da a sou ces can be ound in many o he coun ies. The kind o da a ha he p ocedu e in ol es is o en publicly a ailable. Real in o ma ion a ailabili y limi a ions ha e also been conside ed, which can some imes make he implemen a ion o he me hodology mo e a duous. Acco dingly, he me hodology also con ibu es o he managemen o da a collec ion ela ed di icul ies by p oposing easy ways o ga he ing u he in o ma ion. In addi ion o add essing he p oblem o da a a ailabili y, he me hodology also ocuses on he o he limi a ions o bo om-up EM. An empi ical me hod is p oposed o add ess he ac ual occupan beha io and calib a e he building models. Due o he capabili y o s a is ical me hods o inco po a e he occupan ’s beha io impac , he me hodology also inco po a es he applica ion o a mul iple linea eg ession o he p ocess o alida ion o he models. In he p esen pape , he a o emen ioned me hodology is applied o he esiden ial building s ock o Bilbao, by cons uc ing a comp ehensi e building s ock agg ega ion model. I demons a es he p ac icabili y o cha ac e izing he ene gy consump ion o a ci y-scale esiden ial building s ock wi h a ailable da a sou ces. This way, a us wo hy basis can be o med o a subsequen assessmen o he impac o di e en ene gy sa ing policies. The models lay he g oundwo k o he assessmen o di e en ene gy conse a ion measu es’ e ec , ep esen ing a use ul decision-making ool o iden i y cos -op imal se s o measu es pe building ype and es ablish he app op ia e ene gy e iciency s a egies. These aspec s a e cu en ly being add essed in u he wo k. The pape is s uc u ed as ollows. Sec ion 2 summa izes he me hodological app oach ollowed o his s udy, di ided in h ee di e en phases. Sec ion 3 desc ibes he implemen a ion o he p ocedu e o he case s udy o Bilbao. Sec ion 4 p o ides and analyzes he esul s. The pape ends wi h a inal Sec ion 5 which ou lines mos impo an conclusions. 2. Me hodological F amewo k A comp ehensi e desc ip ion o he me hodology is p o ided by [30]. The main guidelines a e summa ized he ein o enable he unde s anding o he case s udy p esen ed in his pape , which enables o u he elabo a e he app oach p esen ed in his sec ion. The concep ual scheme o he me hodology is p esen ed in Fig. 1. The me hodology is di ided in o h ee main phases: Building S ock Segmen a ion, Building S ock Modelling and Model Valida ion. Fig. 1. O e all s uc u e o he applied bo om-up me hodological app oach Phase 1 aims o de e mine he a che ype buildings o be modelled and simula ed. The a che ype buildings a e ob ained h ough he combina ion o h ee segmen a ion c i e ia: a) building cons uc ion pe iod ( ela ed o ene gy egula ion) b) building heigh , in e ms o he numbe o s o eys o e g ound c) dwelling use ul a ea The hea ing sys em ype is no in oduced as a segmen a ion pa ame e , as i is p e e able o pe o m ene gy pe o mance simula ions in e ms o ene gy demands, which could be la e con e ed o consump ion e ms by applying he co esponding mean seasonal e iciencies. Thus he de ini ion o excessi e segmen a ion c i e ia is a oided, pe mi ing a deepe classi ica ion o he building s ock based on he p oposed h ee c i e ia. By combining he segmen a ion c i e ia, a clus e map is gene a ed, which pe mi s o iden i y he mos ep esen a i e dwelling g oups. To ca y ou he analysis, he main da a sou ce is he Popula ion and Housing Census om he Na ional S a is ics Ins i u e [31]. I should be no ed ha , as usually happens wi h his ype o sou ces, mos o i s da a e e s o indi idual dwellings, no en i e buildings, while he a che ypes o be cons uc ed and simula ed a e en i e buildings, and no indi idual dwellings. The e o e, di e en homogeneous g oup ep esen a i eness mus be i s assessed in e ms o indi idual dwellings, and a e wa ds a ela ion mus be se be ween hose ep esen a i e dwellings and he ela ed a che ype buildings, which a e cons i u ed by he o me s. In Phase 2, each a che ype building is de ined by i s physical and echnical pa ame e s, as well as o he da a inpu s equi ed o he model cons uc ion. The cons uc ion o a che ype building models o en equi es la ge amoun o da a as inpu , bu he e a e ce ain building ea u es which ha e highe in luence han o he s on i s he mal beha iou . The ollowing h ee main ac o s ha e been iden i ied as he mos signi ican a ec ing he he mal pe o mance o a building, which dese e special a en ion: i) building acade composi ion, ii) u ban o m whe e a building is loca ed, which a ec s he building o m and con igui y and, iii) occupan beha iou p o iles. In addi ion, a use ul empi ical calib a ion me hod is p oposed. Finally, Phase 3 explo es inno a i e me hods o ob ain eal dwelling consump ion da a, in o de o compa e hem wi h he a che ype buildings simula ions esul s and hus alida e he models. An o iginal su ey is sugges ed o easily collec ac ual uel consump ion in o ma ion om indi idual inhabi ed dwellings. In addi ion, a bo om-up s a is ical eg ession app oach is p oposed o ob ain he hea ing demand o di e en dwelling ypologies. Se e al s udies exis which u ilize pos al code le el ene gy consump ion da a in a eg ession [32][33]. Ins ead, his me hodology adap s o he a ailable da a sou ces and al e na i ely applies he eg ession app oach a egional scale, by using agg ega ed municipal-le el domes ic consump ion alues. 3. Me hodology Applica ion: Case S udy In o de o show he applica ion o he p oposed app oach, he de eloped me hodology is applied o he esiden ial building s ock o Bilbao. Loca ed in he p o ince o Biscay, Basque Coun y egion, Bilbao is he la ges ci y in no he n Spain, wi h a popula ion o app oxima ely 350,000 inhabi an s. The u ban a ea, loca ed in a i e alley, is su ounded by wo moun ain anges on he no he n and sou he n sides. The municipali y is di ided in o eigh dis ic s, which a e indica ed in Fig. 2. Fig. 2. Spa ial dis ibu ion o he dis ic s o Bilbao Hea ing ene gy used in he esiden ial s ock is mainly p o ided om na u al gas (app ox. 50%). Apa om na u al gas, he mos used ene gy ca ie s o space hea ing a e elec ici y (30%) and oil de i a i es (20%, including gasoil and LPG). Gasoil is mainly used in communal hea ing sys ems, while na u al gas and elec ici y a e mos ly used in dwelling’s indi idual sys ems. The es o uels, such as wood o coal, a e e i ied negligible. 3.1. Building S ock Segmen a ion The p esen analysis ocuses on main dwellings, i.e. dwellings which a e used all yea ound, ac ing as usual home, and excludes seconda y (occasionally used) and emp y dwellings. Main dwellings ep esen mo e han he 90% o Bilbao’s esiden ial s ock [31], and accoun o a s ock o 147,655 houses, which is he a ge o he s udy. Fig. 3 shows he dis ibu ion o Bilbao’s main dwellings acco ding o hei cons uc ion pe iod. Almos he hal o he exis ing building s ock was cons uc ed wi hin a ime ame o only 20 yea s, be ween 1960 and 1980, du ing he so-called De elopmen alism pe iod. In addi ion, app oxima ely he 80% o he houses we e cons uc ed be o e he applica ion o he Spanish NBE-CT-79 egula ion [34], which came in o o ce in 1979 and wi h which he u iliza ion o building he mal insula ion commenced. On he o he hand, he numbe o houses ha comply wi h he subsequen (mo e s ingen om he ene gy poin o iew) CTE egula ion [35], in o ce since 2006, is e y low. Fig. 3. Dis ibu ion o Bilbao’s main dwellings acco ding o hei cons uc ion pe iod. The e is li e a u e which, when classi ying an exis ing building s ock, di e en ia es be ween single- amily houses and mul i- amily houses [36][37]. Ne e heless, Bilbao has e y ew single- amily houses (p obably due o he complica ed opog aphic na u e and his o ical land sca ci y), which a e less han 1% o he en i e building s ock [38], and he e o e his segmen a ion c i e ion is excluded o building classi ica ion. By combining he h ee segmen a ion c i e ia p oposed in he me hodology, Table 1 is gene a ed, which pe mi s o iden i y he mos ep esen a i e g oups o dwellings. Table 1. Bilbao’s main dwelling clus e map Only dwelling ypes wi h a sha e g ea e han 2% a e selec ed, esul ing in he 17 ep esen a i e dwellings shown in Table 1, which a e named om “RD 1” o “RD 17”. All oge he , hey ep esen he 87.90% o he o al esiden ial s ock o Bilbao, which is conside ed ep esen a i e enough. Dwelling’s loo a ea Cons uc io n Pe iod S o eys o e g ound (incl. g ound loo ) < 60 m2 60 - 90 m2 90 - 120 m2 > 120 m2 < 1960 < 3 0.47% 0.09% 0.05% 0.06% 4 – 6 5.85% RD1 6.37% RD2 1.32% 0.23% 7 - 9 3.72% RD3 8.38% RD4 3.42% RD5 2.40% RD6 > 10 0.22% 0.68% 0.28% 0.06% 1960 – 1980 < 3 0.10% 0.00% 0.00% 0.00% 4 - 6 4.79% RD7 5.56% RD8 0.92% 0.00% 7 - 9 5.10% RD9 12.93% RD10 2.82% RD11 0.83% > 10 2.90% RD12 9.13% RD13 3.59% RD14 0.81% 1981 – 2005 < 3 0.00% 0.06% 0.00% 0.00% 4 - 6 0.65% 3.99% RD15 0.47% 0.00% 7 - 9 1.10% 4.93% RD16 1.52% 0.11% > 10 0.00% 2.01% RD17 0.53% 0.00% > 2006 < 3 0.00% 0.00% 0.00% 0.00% 4 - 6 0.13% 0.80% 0.00% 0.00% 7 - 9 0.13% 0.36% 0.08% 0.00% > 10 0.00% 0.05% 0.00% 0.00% The a che ypes o be modelled and simula ed a e en i e buildings, no indi idual dwellings, and he e o e a ela ion is es ablished o ans o m each ep esen a i e dwelling in o an a che ype building. To do so, he ollowing app oach is applied: -E e y dwelling con ained in each a che ype building is equal and co esponds o one o he de ined ep esen a i e dwellings. -E e y s o ey o he a che ype building is equal: i has he same numbe o dwellings, and he e o e has he same plo a ea. So he las aspec o be de ined is he o al amoun o dwellings pe building ype, and i s dis ibu ion along he building s o eys. The dwellings quan i y o each a che ype building is de e mined aiming o be as nea es as possible om he a e age alues ob ained om [38]. A he same ime, a che ype’s numbe o s o eys is also de ined, adjus ing i wi hin he ange o he co esponding building ype. I is assumed ha he e is no dwelling in he g ound loo . Table 2 shows he ob ained dwellings dis ibu ion pe a che ype building. Table 2. Dwellings dis ibu ion pe a che ype building 3.2. Building S ock Modelling The ollowing main inpu da a ca ego ies a e p esen ed he ein: i) geome ical cha ac e is ics, ii) cons uc ion cha ac e is ics and, iii) occupan beha iou p o iles. 3.2.1. Geome ical cha ac e is ics The main aim o his sec ion is o de e mine he geome ical p ope ies o he a che ype buildings i.e. hei shape and con igui y. To his e ec , Bilbao’s u ban mo phology is analysed. His o ical u ban de elopmen Ini ially, Bilbao’s his o ical u ban de elopmen is assessed o be e unde s and and classi y he di e en u ban ab ic zones o he ci y. The e olu ion o he ci y has esul ed in clea ly di e en ia ed u ban mo phologies, which can be iden i ied and associa ed wi h he di e en neighbou hoods. On he one hand, in Bilbao a ela ion exis s be ween he p edominan u ban mo phology o an a ea and i s opog aphy. Thus, he la si es o he ci y cen e, in he middle o he alley and a ound he i e , show a solid block u ban ab ic, ypical o 19 h cen u y Ensanche (ci y expansion in Spanish, which g ew up in o hogonal layou beyond he exis ing Old Town). In u n, in he hillsides o he pe iphe y he linea block is p edominan , which be e accommoda es o he di icul opog aphy. On he o he hand, a ela ion be ween he cons uc ion pe iod and he u ban block ypes is also no ed. Due o he his o ical de elopmen o he ci y, whose demog aphic and cons uc ion boom coincided wi h he p e ailing a ionalis s yle, he linea block is p edominan in he pe iphe al neighbou hoods c ea ed om Cons uc ion Pe iod S o eys o e g ound (including g ound loo ) S o eys o e g ound (excluding g ound loo ) Rep esen a i e dwellings pe s o ey To al dwellings ob ained pe a che ype building A e age dwellings pe building - [38] Be o e 1960 4 - 6 loo s 4 2 8 8.40 7 - 9 loo s 7 2 14 13.92 1960 - 1980 4 - 6 loo s 4 3 12 13.07 7 - 9 loo s 6 3 18 18.17 > 10 loo s 12 3 36 37.92 1981 – 2005 4 - 6 loo s 4 3 12 12.46 7 - 9 loo s 6 3 18 18.24 > 10 loo s 13 3 39 39.06 he 60s, du ing he De elopmen alism pe iod. In such a eas, he building wid h, i.e. he s uc u al span be ween acades, is o en be ween 8 – 10 me e s long. Such na owness, in luenced by he hygienic ideal o he a ionalism, allowed o c ossed en ila ion in he whole dwelling. On he con a y, he acade leng h o buildings loca ed wi hin solid block u ban ab ic can change signi ican ly. In he Old Town, whe e he oldes buildings a e loca ed ( om 19 h cen u y and be o e), he acades a e in gene al sho , wi h a ypical leng h o app oxima ely 8 me e s. In con as , la e 19 h cen u y buildings in he Ensanche, which cu en ly cons i u e he ci y cen e, show usual acade leng hs be ween 16 – 24 me e s. Conside ed U ban Blocks Based on he p e ious assessmen , Table 3 shows he h ee u ban o m ypologies conside ed in he p esen analysis: solid block, linea block and de ached buildings. A building con igui y index, i.e. he en elope’s po ions in con ac wi h o he buildings o he en i onmen , is alloca ed o each u ban o m. Table 3. U ban o m ypes conside ed o u ban o m alloca ion o a che ype buildings In Bilbao, he in luence o sola i adia ion is less se e e han in o he Spanish ci ies. Based on [39] measu ed sola i adia ion da a, ob ained du ing he 1983 – 2005 pe iod, Bilbao is he Spanish capi al ci y wi h he lowes sola i adiance, wi h a daily a e age o 3.54 kWh/m2. In addi ion, i s pa icula opog aphy, being loca ed wi hin a alley and su ounded by hills, makes i less exposed o he sola ays. The e o e, he modelling o ex e nal shading c ea ed by he en i onmen a ound a building is excluded om he analysis, which enables o include di e en u ban layou s such as closed o open blocks wi h in e nal cou ya ds and linea blocks in o a unique u ban o m ypology, as i can be seen in Table 3. Fo he p esen analysis, all hese u ban layou s a e conside ed equi alen in e ms o hei buildings shape and con igui y. On he o he hand, he small in e nal cou ya ds (“chimney cou ya ds”) ha migh exis wi hin a solid block layou a e excluded om he assessmen . Shape and Con igui y Alloca ion o A che ype Buildings Fi s ly, he geog aphic dis ibu ion o each ep esen a i e dwelling is analysed, wi h he aim o iden i ying he ci y dis ic s whe e a speci ic ep esen a i e dwelling is concen a ed he mos ( o his e ec , “a che ype building” e m and i s co esponding “ ep esen a i e dwelling” can be in e changeably used). I shall be no ed ha , a his poin , ha ing de ined building heigh and dwelling use ul a ea as segmen a ion c i e ia con ibu es o a clea e concen a ion o a che ypes in speci ic dis ic s. As an example, Table 4 indica es he dis ic dis ibu ion ha co esponds o each RD acco ding o [31], in which he colou ed cells show he dis ic s whe e each RD is mainly concen a ed. U ban Fo m Cha ac e is ics Examples Alloca ed Con igui y Index Solid Block Typical o he 19 h cen u y Ensanche. The building is in eg a ed wi hin a building compac layou , wi h a unique acade acing he s ee . 3 ou o 4 en elope sides in con ac wi h o he buildings Linea Block Typical a ionalis block amed wi hin he hygienic ideal o ligh , en ila ion and sunligh , while add essing he lack o housing h ough simple buildings se ializa ion. 2 ou o 4 en elope sides in con ac wi h o he buildings De ached Buildings Buildings wi h no en elope’s po ion in con ac wi h o he buildings No en elope’s po ion in con ac wi h o he buildings Table 4. RD 7 ep esen a i e dwelling dis ibu ion by dis ic Aging maps a e buil o each neighbou hood o ci y zone, o which he public da a p o ided by he cadas e is used ega ding yea o cons uc ion o he buildings. Two examples o he cons uc ed building’s aging maps a e shown in Fig. 4, which co espond o O xa koaga - Txu dinaga and U iba i dis ic s. Fig. 4. Examples o maps wi h cons uc ion pe iod o buildings Ci y Dis ic s Rep esen a i e Dwelling ype Deus o U iba i O xa koa ga Txu dina Begoñ a Ibaion do Aband o Rekald e Basu o Zo oz To al analysed dwelling RD 1 11.90 % 16.29% 14.80% 23.65% 20.91% 12.45% 75.65% RD 2 33.93 % 9.74% 6.89% 29.96% 13.33% 6.14% 77.23% RD 3 13.96 % 6.76% 1.80% 24.55% 26.35% 17.45% 9.12% 82.32% RD 4 5.76% 4.55% 2.07% 21.63% 40.27% 13.85% 11.87% 87.62% RD 5 9.45% 90.55% 90.55% RD 6 7.37% 92.63% 92.63% RD 7 9.98% 24.94% 25.02% 14.80% 4.42% 17.99% 2.86% 82.75% RD 8 21.66 % 30.23% 1.96% 12.21% 6.32% 20.13% 7.49% 84.23% RD 9 10.21 % 10.38% 14.91% 19.69% 17.55% 4.70% 15.24% 7.33% 67.38% RD 10 16.46 % 8.71% 1.36% 27.27% 14.66% 5.18% 14.33% 12.02% 72.73% RD 11 25.35 % 10.38% 17.96% 2.99% 25.95% 2.79% 14.57% 83.83% RD 12 13.29 % 9.93% 14.90% 47.11% 5.37% 3.49% 3.22% 2.68% 75.30% RD 13 20.42 % 14.66% 10.49% 23.47% 12.54% 2.88% 9.53% 6.01% 71.09% RD 14 17.12 % 20.02% 16.58% 7.10% 6.57% 10.87% 16.36% 5.38% 70.08% RD 15 7.79% 47.21% 4.88% 17.33% 3.84% 18.95% 83.49% RD 16 16.08 % 2.59% 4.77% 6.95% 32.78% 4.05% 16.29% 16.49% 81.64% RD 17 4.91% 31.01% 4.65% 24.81% 28.94% 5.68% 84.75% Table 8. Fuel consump ion dis ibu ion in he Basque Coun y [49] 3.3.2. Fuel Consump ion Su ey An excellen me hod o he alida ion o modelled building’s hea ing demand is o ga he ac ual ene gy consump ion da a om inhabi ed houses o Bilbao, which is con ained in ene gy bills. To his e ec , a su ey is conduc ed, which aims o collec ac ual uel consump ion in o ma ion om indi idual inhabi ed dwellings, by asking he su ey esponden s abou hei annual uel use. Ne e heless, he complexi y ha his ype o su ey en ails is no i ial. Dwelling billing da a is o en conside ed p i a e in o ma ion, wi hou an ob ious disposi ion o sha e i . In addi ion, answe ing he su ey implies he necessi y o looking o a ce ain amoun o bills which may co e a leas one-yea pe iod, and summing he di e en mon hly-in oiced consump ions, which is ob iously edious. An inno a i e app oach is p oposed, aiming o o e come hose p oblems and acili a e da a collec ion: -I is only eques ed a pic u e o he g aph ha is included in any bill, which con ains his o ical in o ma ion abou he uel consump ion o he las wo yea s. The g aph shows he di e en mon hs in x-axis and he co esponding uel consump ion in kWh as y-axis. Thus, equi ed esponse imes a e signi ican ly educed, wi hou subs an ially a ec ing he quali y o he collec ed da a. -In addi ion o he men ioned g aph, h ee supplemen a y da a a e eques ed, which a e easy o answe : dwelling add ess, numbe o inhabi an s and ki chen ype (elec ical o uel- i ed). Dwelling add ess se es o ob ain dwelling’s a ea om he cadas e. In o al, 65 dwellings domes ic uel consump ion da a a e collec ed, mainly o Bilbao bu also om municipali ies o i s me opoli an a ea. The e o o using his sample is es ima ed using he Eq. 1 [50]. P o ided Biscay’s building s ock olume, his sample gi es a ma gin o e o o 5 kWh/m2, wi h an associa ed 90% con idence in e al. Whe e n is he su ey’s size, N is he size o he popula ion (building s ock olume), Z is he accep ed de ia ion om he mean alue o achie e he assumed con idence le el (gi en by he gaussian dis ibu ion o m), e is he e o ma gin and σ2 is he expec ed a iance o he popula ion. Two di e en app oaches a e applied o analyse collec ed da a. The i s me hod, shown in Fig. 8, is based on DHW consump ion es ima ion using summe pe iod uel consump ion, and elies on he assump ions ha du ing summe he e is no space hea ing- ela ed uel use, and ha DHW consump ion is no likely o signi ican ly a y wi h seasons. The second app oach consis s on he applica ion o he a e age uel use dis ibu ion (%) by domes ic end-use, p e iously shown in Table 8. Compa ing he ob ained no malized annual hea ing consump ion alues (kWh/m2), a ai consis ency is e i ied be ween bo h app oaches. Fig. 9 shows how he ob ained alues om each app oach a e e y cohe en , conside ing he unde lying a iabili y ha ac ually exis s behind a e age uel dis ibu ion sha es o Table 8. Ne e heless, i seems ha in he cases o highe space hea ing demands, he DHW es ima ion app oach unde es ima es he consump ion e e ed o space hea ing in compa ison wi h he end-use sha es o Table 8. Basque Coun y, Coas al Clima ic Zone Fuels consump ion dis ibu ion (%) Space Hea ing 59% DHW 30% Cooking 11% n=N·z2·σ2 (N−1)·e2+z2·σ2 Eq . 1 Fig. 8. Da a analysis app oach based on es ima ed DHW summe consump ion Fig. 9. Compa ison o he annual space hea ing ene gy consump ions ob ained wi h he wo app oaches 3.3.3. Reg ession model This sec ion aims o ob ain he space hea ing demand o di e en dwelling ypologies h ough a bo om- up s a is ical eg ession app oach. This me hod is based on eal agg ega ed le el consump ion da a, so i s ou pu cons i u es an addi ional basis o he cons uc ed models’ alida ion. The mul iple linea eg ession is a s a is ic modelling echnique used o adjus linea models be ween a dependen a iable (y) and mo e han one explana o y o independen a iables (xi). The aim is o ob ain bi coe icien s om an equa ion sys em consis ing o j equa ions as Eq. 2. Due o he lack o consump ion da a a pos al code le el, he eg ession app oach is applied in a egional scale, by using municipal-le el domes ic consump ion alues om municipali ies loca ed in he Basque Coun y, a ound Bilbao. The implemen a ion o Local Ac ion Plans (LAPs), wi hin he amewo k o Agenda 21 p ocesses, has led o he measu emen and publica ion o se e al sus ainabili y indica o s, which o en include na u al gas municipal consump ion. DHW and cooking a e age uel use sha es in Table 8 a e discoun ed om he esiden ial na u al gas consump ion, hus ob aining he domes ic space hea ing na u al gas consump ion. To ans o m i in o demand e ms, mean seasonal e iciencies based on [51] a e used, which a e a e aged o each municipali y acco ding o he hea ing sys ems and uel dis ibu ions. yj=b0+b1·x1j+b2·x2j+b3·x3j+ . . . + bn·xnjEq . 2 In o de o ex apola e, municipali y by municipali y, he hea ing demand sa is ied by na u al gas in o a o al domes ic hea ing demand, na u al gas hea ed dwellings loo a ea’s sha e is conside ed, o e o al municipal dwellings loo a ea [31]. This way, in his eg ession app oach he dependen a iables (y) a e he agg ega ed space hea ing demands o he di e en municipali ies, no malized o e he o al dwelling s ock loo a ea (kWh/m2). The xi a iables a e he sha es (%) o each dwelling ypology’s loo a ea, o e he o al s ock loo a ea in each municipali y. The pu sued bi coe icien s would he e o e be he hea ing demands (kWh/m2) o each dwelling ypology. Municipali ies wi h a popula ion lowe han 4,000 inhabi an s ha e been excluded om he analysis. Due o hei lowe amoun o dwellings, i is conside ed ha he lowe accu acy o he s a is ical da a could o he wise dis o he applica ion o he eg ession me hod. To ally, municipal esiden ial na u al gas consump ion is ob ained om 17 municipali ies o he Basque Coun y. All o hem a e loca ed in he same coas al clima ic zone, so he consis ency o he app oach is ensu ed. A e age municipal consump ions a e ob ained om he consump ion alues o di e en yea s wi hin he 2005 – 2011 his o ical pe iod, depending on he a ailable in o ma ion in each municipali y. Conce ning dwelling ypologies, i would be desi able he p esen eg ession analysis o include he same 17 ep esen a i e dwelling ypes used o cons uc modelled a che ype buildings. Howe e , he dwelling b eakdown de ail le el o be in oduced in he analysis depends on he amoun o a ailable da a, i.e. agg ega ed municipal consump ion da a. The e o e, a simpli ied classi ica ion mus be pe o med. Due o he low sha e ha he dwellings cons uc ed a e 2006 imply, hey a e excluded om he analysis. The es o dwellings a e classi ied in o he h ee dwelling ypologies shown in Table 9, based on hei cons uc ion pe iod. I is analysed he possibili y o u he dis inguishing he dwellings acco ding o hei numbe o s o eys, bu he sha es o 1 – 2 s o eys buildings, which include single- amily houses, a e gene ally obse ed negligible. Ul ima ely, he numbe o he dwelling ypologies in oduced in he eg ession app oach complies wi h he ule o humb o ensu ing a minimum a io o i e samples pe each independen a iable [52]. In his case, he a io is 17:3. Table 9. In oduced dwelling ypologies in he eg ession analysis 3.3.4. Assessmen o Ene gy Pe o mance Ce i ica es The e is no doub ha he in oduc ion o Ene gy Pe o mance Ce i ica es (EPCs) in he EU cons i u es an in e es ing and ex ensi e da a sou ce o ene gy policy planning pu poses. Recen s udies poin o a wide spec um o po en ial applica ions o EPCs da a han o iginally in ended, including he alida ion o building s ock models [53]. The e o e, i is assessed he applicabili y o he EPCs as an addi ional sou ce o building ene gy modelling esul s alida ion. An EPCs egis y is publicly a ailable in he webpage o he Basque Go e nmen [54]. Ins ead o he comple e EPCs, he da abase p o ides he inal ene gy e iciency label o each dwelling, which includes CO2 emissions and p ima y ene gy consump ion alues. Acco ding o dwellings ene gy ce i ica ion p ocedu e in Spain, hose alues co espond o he space hea ing, space cooling and DHW end-uses. Conside ing ha cooling end-use could be conside ed negligible in Bilbao, a calcula ion p ocedu e is se based on he applica ion o he co esponding con e sion ac o s, o es ima e mean annual space hea ing consump ion. To ally 228 EPCs a e analysed. The ob ained mean annual space hea ing consump ion is inconsis en wi h he es o collec ed eal da a: i is signi ican ly highe han he alues assessed in he li e a u e e iew and uel consump ion su ey. The e o e, i seems ha he cu en p ac ice on EPCs d i es o an o e es ima ion o hea ing ene gy demand. Such pe o mance gap, i.e. he di e ence be ween es ima ed and ac ual ene gy pe o mance, has been widely assessed in he li e a u e [55][56][57][58] [59]. In ac , he e is a b oad ag eemen ha buildings wi h poo he mal pe o mance end o consume less han p edic ed in he EPCs [60][61][62], mainly due Typology 1 Typology 2 Typology 3 Building Cons uc ion Pe iod Be o e 1960 1960 - 1980 1981 – 2005 o he o e es ima ion o ac o s usually se by s anda ds and o he assump ions used in EPC a ings [63] [64][65][66]. The e o e, he possibili y o building ene gy model’s esul s alida ion h ough EPC da abases is excluded om he p oposed me hodology. 4. Resul s The cons uc ed 17 a che ype buildings models a e simula ed by Design Builde .4.7.0.027 [47] so wa e. Fig. 10 summa izes he no maliza ion pe hea ed loo a ea o he annual space hea ing demands o a che ype buildings, p esen ed pe cons uc ion pe iod and u ban o m. The space hea ing demands show a co ela ion wi h he cons uc ion pe iod and u ban o m. On he one hand, he e olu ion o cons uc i e ea u es ends o educe he hea ing demand. Fo example, wi h he same u ban o m and simila building dimensions, bu di e en açade composi ion, A ch. 7 (U = 1.26 W/ m2K) and A ch. 15 (U = 0.63 W/m2K) show an annual hea ing demand 15% and 52% lowe , espec i ely, han A ch. 2 (U = 1.77 W/m2K). On he o he hand, due o he highe con igui y, buildings loca ed in solid blocks p esen lowe demands han he ones placed as linea blocks o de ached buildings. Fo example, wi h he same acade composi ion and simila building dimensions, bu di e en u ban o m, A ch. 1 (linea block) shows a hea ing demand 70% highe han A ch. 4 (solid block). In addi ion, he mos ecen buildings seem o be dis inguished by a lowe hea ing demand, hanks o he in oduc ion o i s ene gy egula ions and ela ed he mal insula ion. Fig. 10. Annual space hea ing demands o a che ype buildings, by cons uc ion pe iod and u ban o m Sec ion 3.2 abo e summa ized he ou pu s o he dep h assessmen conduc ed o de e mine he model pa ame e s which a e conside ed mos in luen ial. The esul s ob ained in he p esen analysis suppo such assump ion. Ne e heless, he e a e o he inpu pa ame e s which has no been such exhaus i ely assessed, ei he due o lack o a ailable da abases o because s anda ds o egula ion alues ha e been applied. In o de o analyse he e ec o hei unce ain y, he local sensi i i y analysis [67] is conduc ed o ob ain he no malised sensi i i y coe icien s, which ep esen he pe cen age change in he annual space hea ing demand o he models gi en a 1% change in he inpu pa ame e . The analysed inpu pa ame e s a e he ollowing: acade glazed a ea (%), in il a ion a e, hea ing se back empe a u e, in e nal gains, occupancy and s o eys’ wall o loo he mal b idge’s linea he mal ansmi ance. Table 10 shows he a e age o he no malised sensi i i y coe icien s o all a che ype buildings, in ela ion o he space hea ing demand. Table 10. A e age o he no malised sensi i i y coe icien s o all a che ype buildings I is e i ied ha he hea ing se back empe a u e ep esen s, by a , he pa ame e which he esul s a e mo e sensible o, ollowed by in il a ion a e. In a much lowe s ep, in e nal gains can be ound, wi h a simila in luence o equipmen and ligh ing, and occupancy. Finally, he esul s sensi i i y o acade glazed a ea and wall o loo he mal b idge Ψ alue is e y simila and negligible. Fig. 11 shows he desc ibed o de o magni ude o he esul s sensi i i y o each pa ame e , as well as he oscilla ion o no malised sensi i i y coe icien s depending on he a che ype building. I is e i ied ha he a iabili y o he esul s due o hese pa ame e s, in hei usual ange, is no ably less han ha caused by he u ban o m, açade composi ion o he occupan beha iou . Fig. 11. A che ype buildings esul s’ sensi i i y o each inpu pa ame e Using a weigh ing ac o o each modelled building a che ype, based on i s sha e o e he o al esiden ial s ock, an annual space hea ing demand o 43.08 kWh/m2 is ob ained on a e age, which is conside ed o co espond o a space hea ing ene gy consump ion o 49.58 kWh/m2 pe yea . A mean seasonal e iciency o 86.89% is es ima ed o Bilbao’s space hea ing, based on [51] e e ence da a, and conside ing he hea ing sys em dis ibu ion in Bilbao [31]. These esul s a e compa ed o alida ion da a ob ained in Sec ion 3.3. Table 11 shows ha he esul ing a e age space hea ing ene gy consump ion is close o he alues ob ained om uel consump ion su ey and li e a u e: i is 1% abo e han he ou come o he su ey conside ing he DHW-based app oach and 10% abo e he su ey ou pu based on end-use a io applica ion app oach. A e age - All A che ype Buildings No malised sensi i i y coe icien , Sj Facade glazed a ea (%) -0.078 In il a ion a e (ACH) 0.704 Hea ing se back empe a u e (ºC) 1.250 Equipmen and ligh ing (maximum W/m2) -0.196 Occupancy (occupan s pe m2) -0.204 Ψ wall o loo he mal b idge 0.078 Table 11. Compa ison o he annual hea ing ene gy consump ion On he o he hand, ocusing on a disagg ega ed le el, Fig. 12 shows he compa ison be ween he modelled space hea ing demands, a e aged pe cons uc ion pe iod, and he demands ob ained om he eg ession app oach. The coe icien o de e mina ion was R2 = 0.51, meaning a mode a e adjus men . The model was globally signi ican wi h a p- alue o 0.022, and all a iables ejec ed he null hypo hesis o an alpha alue o 0.01. The assump ion o independence o esiduals was analysed h ough he Du bin-Wa son es , ob aining a alue o 2.16 ha demons a es he lack o au oco ela ion. The assump ions o linea i y and homoscedas ici y we e also e i ied h ough he sca e plo o esiduals agains p edic ed alues. Compa ing he eg ession esul s wi h he ou pu s om he a che ype buildings modelling, he demand alues co esponding o he cons uc ion pe iod be o e 1960 and he De elopmen alism pe iod a e e y simila , being he modelled demand sligh ly highe in he case o pos -1981 buildings. Fig. 12. Compa ison be ween he modelled space hea ing demands, and he ones ob ained om he eg ession Ne e heless, he ob ained mean annual hea ing demand o modelled 1981 – 2005 a che ype buildings, which is 29.38 kWh/m2, co esponds o an annual consump ion o 33.81 kWh/m2, which is only 5% highe han he mean ene gy consump ion (32.13 kWh/m2) o he comple e se o 107 dwellings o med by he ou eal buildings used o occupan beha iou calib a ion. The e o e, he di e ence wi h he demand ob ained om he eg ession app oach may be caused by he limi ed numbe o samples in oduced in such app oach, which combined wi h he usual lowe sha e o pos -1981 buildings o e he o al s ock, could d i e o g ea e e o in hei esul . Finally, he esul s a e analysed a an agg ega ed le el. The En i onmen al Depa men o he Ci y Council is con ac ed wi h he aim o collec ing u he in o ma ion abou he his o ical e olu ion o he esiden ial na u al gas consump ion in Bilbao. Thanks o i s collabo a ion, he ull 2004 – 2017 his o ical consump ion end is comple ed, which is con e ed o esiden ial space hea ing demand e ms by applying he same assump ions desc ibed o eg ession. The ob ained disagg ega ed ene gy demand es ima es a e ex apola ed o he whole esiden ial s ock by mul iplying he esul s by he numbe o houses which i he desc ip ion o each building a che ype. Table 12 shows ha he ob ained o al space hea ing demand o Bilbao is e y simila o he mean demand alue de i ed om eal consump ion da a o 2004 – 2017 pe iod. The di e ence is only abou 4%. Building Modelling Resul Fuel Consump ion Su ey Li e a u e App oach based on es ima ed DHW App oach based on end-use a io SECH-SPAHOUSEC A lan ic Zone [48] 49.58 kWh/m2 48.98 kWh/m2 44.67 kWh/m2 44 kWh/m2 Table 12. Compa ison o Bilbao’s esiden ial hea ing ene gy demand 5. Conclusions An app op ia e ene gy cha ac e iza ion o an exis ing building s ock can help o implemen he necessa y ene gy e iciency policies in he mos e ec i e way. Recen imp o emen s in geog aphical in o ma ion sys ems ha e led o a ise in models de eloped o u ban building s ocks. Ne e heless, da a a ailabili y on which hey ely is no always a ailable, and he sophis ica ion o ce ain models could become a disad an age o being eplicable. A he same ime, local au ho i ies can cons i u e a signi ican bo om- up suppo in implemen ing he impe a i e ene gy sa ing policies. Aiming o con ibu e o such ield, his pape desc ibes an easy- o-implemen me hodology o he cha ac e iza ion o esiden ial hea ing ene gy consump ion a ci y le el. Pa icula a en ion has been gi en o i s applicabili y, so he me hodology lies on a ailable public da a sou ces and assumes eal limi a ions o in o ma ion ools a ailabili y. The me hod has been concei ed o be applicable o any ci y, p o ided a p ima y unde s anding o he cons uc i e e olu ion in he analysed geog aphical con ex . The p esen ed me hod lies on a bo om-up enginee ing app oach, which o en in ol es se e al challenges o deal wi h. This app oach equi es la ge amoun o inpu in o ma ion o model cons uc ion and alida ion p ocesses. To con ibu e o da a collec ion ela ed di icul ies, he in oduced me hodology p oposes o iginal ways o ga he ing u he in o ma ion, as bene i ing om he con en o uel consump ion in oices. Ano he weakness o his kind o app oaches lies on he unp edic able na u e o human beha iou , which can lead o impo an gaps be ween ac ual and modelled ene gy consump ion. In he de eloped me hodology, an empi ical calib a ion p ocedu e is p oposed o cons uc occupan beha iou p o iles, based on eal consump ion da a om buildings wi h communal hea ing sys ems. The lack o anspa ency ega ding o he model inpu s- ela ed assump ions based on au ho ’s expe ience, as well as he common u iliza ion o s anda d alues o ce ain inpu pa ame e s, o en lead he model ou pu s o lack c edibili y i igo ous alida ion is no pe o med. In he desc ibed me hodology impo an ocus is gi en o he p ocess o e i ying models esul s, by collec ing eal da a o an e idence-based es ing. Valida ion me hods, a bo h disagg ega ed and agg ega ed le el, allow he eliabili y o he esul s o be ensu ed. A mul iple linea eg ession app oach is also p oposed, applied a a eginal scale ins ead o he usual pos al code le el. In addi ion, sensi i i y analysis helps o delimi he in luence o subjec i e assump ions. The desc ibed me hod was implemen ed o he esiden ial building s ock o Bilbao, whe e a o al o 17 ep esen a i e buildings we e iden i ied. The esul s cons i u e a da a se which is signi ican ly aligned wi h ga he ed eal consump ion in o ma ion. Applying he me hodology o Bilbao demons a es ha he p oposed me hod can accu a ely ep oduce he exis ing hea ing ene gy consump ion, p o iding a good es ima e o he he mal pe o mance o he esiden ial s ock. The implemen a ion o he p esen ed p ocedu e p o ides an agg ega ed hea ing ene gy demand ha only di e s a ound 4% om eal his o ical da a which, being in he o de o magni ude o he di e ences achie ed in o he analyses [10][14][25][32] [58], is conside ed success ul. Accu a e modelling o buildings ene gy demand cons i u es a key ool o assis ene gy e o i ing policies. The ou come o his s udy p o ides Bilbao’s baseline ene gy consump ion es ima ion, disagg ega ed pe building ype, o ming a basis o u he in es iga ions aimed a explo ing he economic e ec s o di e en ene gy e iciency s a egies, iden i ying cos -op imal se s o measu es pe building ype and es ima ing he equi ed in es men o he achie emen o di e en ene gy a ge s. This esea ch is cu en ly being conduc ed and he au ho s expec o publish i s esul s soon. In building s ocks modelling h ough a che ype buildings, one o he challenges is o ind a limi ed numbe o essen ial building pa ame e s as segmen a ion c i e ia, which pe mi o main ain a easonable compu a ional ime. The esul s show ha ene gy consump ion a ia ion be ween buildings in di e en Es ima ed Residen ial Space Hea ing Demand (MWh) Bilbao A e age 2004 - 2017 461,584 Building Modelling Resul (MWh) 481,915 u ban mo phologies can be much la ge han he di e ences caused by he a ying acade solu ions due o di e en cons uc ion pe iods. The u ban o ms ha e a signi ican impac on he ene gy pe o mance o a building, and ha is why i is c ucial o include, wi hin segmen a ion c i e ia o Phase 1, pa ame e s ha could indi ec ly be ela ed wi h di e en u ban mo phologies, such as building heigh and dwelling loo a ea. The ou lines o such ela ion mus be sea ch wi hin he his o ical e olu ion o a ci y and he de elopmen o he di e en a chi ec u al s yles. Al hough he applica ion o he p esen ed me hodology is e i ied comple ely easible wi h exis ing da a sou ces, he e a e some limi a ions ela ed o he lack o su icien da abases ha make i mo e a duous i s implemen a ion. As i is b oadly men ioned in he li e a u e [22][26] he e is oom o imp o emen in he de elopmen o he exis ing building s ock ea u es da abases, as well as in he knowledge o domes ic ene gy consump ion cha ac e is ics. Fu he mo e, a g ea po en ial exis s o IEEs o become an impo an da a sou ce o building ene gy modelling. The applicabili y o Ene gy Pe o mance Ce i ica es (EPCs) is also e alua ed as an addi ional way o alida ion o building ene gy modelling esul s. Ne e heless, i seems ha he cu en p ac ice on EPCs d i es o an o e es ima ion o hea ing ene gy demand, which is aligned wi h ou comes o o he s udies [55][57][61]. The e o e, he possibili y o building ene gy model’s esul s alida ion h ough EPC da abases is excluded om he p oposed me hodology Ne e heless, he esul s demons a e he p ac icabili y o cha ac e izing he hea ing ene gy consump ion o a ci y-scale esiden ial building s ock wi h a ailable da a sou ces, aiming o p o ide a basis o he in es iga ion o ene gy sa ings po en ial and ela ed CO2 emission educ ions. Acknowledgemen s The au ho s would like o acknowledge he na u al gas consump ion da a p o ided by he En i onmen al Depa men o he Ci y Council o Bilbao. This esea ch did no ecei e any speci ic g an om unding agencies in he public, comme cial, o no - o -p o i sec o s. Re e ences [1] In e na ional Ene gy Agency. CO2 emissions om uel combus ion. Highligh s. 2017. [2] In e go e nmen al Panel on Clima e Change (IPCC). Global Wa ming o 1.5 ºC. Summa y o Policymake s. 2018. [3] Eu opean Commission (EC). Ene gy Roadmap 2050 [COM/2011/885]. 2011. [4] Eu opean Commission (EC). A Clean Plane o all. A Eu opean s a egic long- e m ision o a p ospe ous, mode n, compe i i e and clima e neu al economy. 2018. [5] Eu opean Commission (EC). Buildings [Online]. A ailable: h ps://ec.eu opa.eu/ene gy/en/ opics/ ene gy-e iciency/buildings [accessed 15.01.2019] [6] POMBO, O.; RIVELA, B.; NEILA, J. The challenge o sus ainable building eno a ion: assessmen o cu en c i e ia and u u e ou look, Jou nal o Cleane P oduc ion, 2016, ol. 123, pp. 88-100. [7] SWAN, L. G; UGURSAL, I. Modeling o end-use ene gy consump ion in he esiden ial sec o : A e iew o modeling echniques, Renewable and Sus ainable Ene gy Re iews, 2009, ol. 13, pp.1819-1835 [8] KAVGIC, M.; MAVROGIANNI, A.; MUMOVIC, D.; SUMMERFIELD, A.; STEVANOVIC, Z.; DJUROVIC-PETROVIC, M. A e iew o bo om-up building s ock models o ene gy consump ion in he esiden ial sec o , Building and En i onmen , 2010, ol. 45, pp. 1683-1697. [9] GHIASSI, N.; MAHDAVI, A. Reduc i e bo om-up u ban ene gy compu ing suppo ed by mul i a ia e clus e analysis, Ene gy and Buildings, 2017, ol. 144, pp. 372-386. [10] NAGELI, C.; CAMARASA, C; JAKOB, M.; CATENAZZI, G.; OSTERMEYER, Y. Syn he ic building s ocks as a way o assess he ene gy demand and g eenhouse gas emissions o na ional building s ocks, Ene gy & Buildings, 2018, ol. 173, pp. 443-460. [11] MONTEIRO, C. S.; PINA, A.; CEREZO, C.; REINHART, C.; FERRAO, P. The use o mul i-de ail building a che ypes in u ban ene gy modelling, Ene gy P ocedia, 2017, ol. 111, pp. 817-825. [12] BALLARINI, I.; CORGNATI, S.P.; CORRADO, V. Use o e e ence buildings o assess he ene gy sa ing po en ials o he esiden ial building s ock: The expe ience o TABULA p ojec , Ene gy Policy, 2014, ol. 68, pp. 273-284. [13] O icial Jou nal o he Eu opean Union. Guidelines accompanying Commission Delega ed Regula ion (EU) No 244/2012 o 16 Janua y 2012 supplemen ing Di ec i e 2010/31/EU o he Eu opean Pa liamen and o he Council on he ene gy pe o mance o buildings by es ablishing a compa a i e me hodology amewo k o calcula ing cos -op imal le els o minimum ene gy pe o mance equi emen s o buildings and buildings elemen s (2012/C 115/01). 19 h Ap il 2012. [14] MATA, E.; SASIC KALAGASIDIS, A.; JOHNSSON, F. Building-s ock agg ega ion h ough a che ype buildings: F ance, Ge many, Spain and he UK, Building and En i onmen , 2014, ol. 81, pp. 270-282 [15] SARTORI, I.; SANBERG, N. H.; BRATEBO, H. Dynamic building s ock modeling: gene al algo i hm and exempli ica ion o No way, Ene gy and Buildings, 2016, ol. 132, pp. 13-25. [16] GANGOLELLS, M.; CASALS, M.; FORCADA, N.; MACARULLA, M.; CUEVA, E. Ene gy mapping o exis ing building s ock in Spain, Jou nal o Cleane P oduc ion, 2016, ol. 112, pp. 3895-3904. [17] STREICHER, K. N.; PADEY, P.; PARRA, D.; BÜRER, M. C.; SCHNEIDER, S.; PATEL, M. K. Analysis o space hea ing demand in he Swiss esiden ial building s ock: Elemen -based bo om-up model o a che ype buildings, Ene gy and Buildings, 2019, ol. 184, pp. 300-322. [18] In elligen Ene gy Eu ope P ojec "EPISCOPE" [online]. A ailable: h p://episcope.eu [accessed 21.09.2019 [19] Co enan o Mayo s o Clima e & Ene gy. Co enan ini ia i e [Online]. A ailable: h ps:// www.co enan o mayo s.eu/abou /co enan -ini ia i e/o igins-and-de elopmen [accessed 15.01.2019] [20] ÖSTERBRING, M.; MATA, E.; THUVANDER, L.; MANGOLD, M.; JOHNSSON, F.; WALLBAUM, H. A di e en ia ed desc ip ion o building s ocks o a geo e e enced u ban bo om-up building s ock model, Ene gy and Buildings, 2016, ol. 120, pp. 78-84. [21] REINHART, C.F.; CEREZO DAVILA, C. U ban building ene gy modelling - A e iew o a nascen ield, Building and En i onmen , 2016, ol. 97, pp. 196-202. [22] HONG, T.; CHEN, Y.; LUO, X.; LUO, N. LEE, S.H. Ten ques ions on u ban building ene gy modelling, Building and En i onmen , 2020, ol. 168, 106508. [23] CEREZO DAVILA, C.; REINHART, C.F.; BEMIS, J.L. Modelling Bos on: A wo k low o he e icien gene a ion and main enance o u ban building ene gy models om exis ing geospa ial da ase s, Ene gy, 2016, ol. 117, pp. 237-250. [24] BRAULIO-GONZALO, M.; JUAN, P.; BOVEA; M.D.; RUA, M.J. Modelling ene gy e iciency pe o mance o esiden ial building s ocks based on Bayesian s a is ical in e ence, En i onmen al Modelling & So wa e, 2016, ol. 83, pp. 198-211. [25] D’ALONZO, V.; NOVELLI, A.; VACCARO, R.; VETTORATO, D.; ALBATICI, R.; DIAMANTINI, C.; ZAMBELLI, P. A bo om-up spa ially explici me hodology o es ima e he space hea ing demand o he building s ock a egional scale, Ene gy & Buildings, 2020, ol. 206, 109581. [26] CHEN, Y.; HONG, T.; LUO, X.; HOOPER, B. De elopmen o ci y buildings da ase o u ban building ene gy modelling, Ene gy & Buildings, 2019, ol. 183, pp. 252-265. [27] BUFFAT, R.; FROEMELT, A.; HEEREN, N.; RAUBAL, M.; HELLWEG, S.; Big da a GIS analysis o no el app oaches in building s ock modelling, Applied Ene gy, 2017, ol. 208, pp. 277-290. [28] FRAYSSINET, L.; MERLIER, L.; KUZNIK, F.; HUBERT, J.L.; MILLIEZ, M.; ROUX, J.J.: Modelling he hea ing and cooling ene gy demand o u ban buildings a ci y scale, Renewable and Sus ainable Ene gy Re iews, 2018, ol. 81, pp. 2318-2327. [29] HAPPLE, G.; FONSECA, J.A.; SCHLUETER, A.; A e iew on occupan beha iou in u ban building ene gy models, Ene gy & Buildings, 2018, ol. 174, 276-292. [30] FERNANDEZ, J.; DEL PORTILLO, L.; FLORES, I.; HERNANDEZ, P. Assessing esiden ial hea ing ene gy consump ion in a ci y-scale: a bo om-up me hodological app oach. 11CNIT. XI Na ional and II In e na ional Enginee ing The modynamics Cong ess. P oceedings book, 2019, pp. 122-133. [31] INE (Spanish Na ional Ins i u e o S a is ics). Popula ion and Housing Census [Online]. A ailable: h ps://www.ine.es/censos2011/ ablas/Inicio.do [accessed 18.01.2019] [32] MASTRUCCI, A.; BAUME, O.; STAZI, F.; LEOPOLD, U. Es ima ing ene gy sa ings o he esiden ial building s ock o an en i e ci y: A GIS-based s a is ical downscaling app oach applied o Ro e dam, Ene gy and Buildings, 2014, ol. 75, pp. 358-367. [33] HOWARD, B.; PARSHALL, L.; THOMPSON, J.; HAMMER, S.; DICKINSON, J.; MODI, V. Spa ial dis ibu ion o u ban building ene gy consump ion by end use, Ene gy and Buildings, 2012, ol. 45, pp. 141-151. [34] Bole ín O icial del Es ado. Real Dec e o 2429/79, de 6 de julio, po el que se ap ueba la No ma Básica de la Edi icación NBE-CT-79, sob e Condiciones Té micas en los edi icios. 6 h July 1979. [35] MITYC. Código Técnico de la Edi icación, 2006. [36] GOUVEIA, J.P.; FORTES, P.; SEIXAS, J.; P ojec ions o ene gy se ices demand o esiden ial buildings: Insigh s om a bo om-up me hodology, Ene gy, 2012, ol. 47, pp. 430-442. [37] MCKENNA, R.; MERKEL, E.; FEHRENBACH, D.; MEHNE, S.; FICHTNER, W., Ene gy e iciency in he Ge man esiden ial sec o : A bo om-up building-s ock-model-based analysis in he con ex o ene gy-poli ical a ge s, Building and En i onmen , 2013, ol. 62, pp. 77-88. [38] INE (Spanish Na ional Ins i u e o S a is ics). Popula ion and Housing Census [Online]. A ailable: h p://www.ine.es/censo/es/inicio.jsp [accessed 18.01.2019] [39] AEMET. A las de Radiación Sola en España u ilizando da os del SAF de Clima de EUMETSAT. 2012. [40] Nie es Basu o Fe o. Los maes os de ob as en la cons ucción de la ciudad: Bilbao, 1876-1910. 1999 [41] E aikal P og am. GUÍA AZALA: Rehabili ación Ene gé ica In eg al de la En ol en e de los Edi icios en Euskadi. 2015. A ailable: h p://www.ga aioak.ejg .euskadi.eus/ 41-18971/es/con enidos/ in o macion/guia_azala/es_de /index.sh ml [accessed 09.06.2019] [42] Depa amen o de Empleo y Polí icas Sociales del Gobie no Vasco. Di ección de Vi ienda. Ca álogo de Rehabili ación Ene gé ica. 2014. [43] Iñigo Lizundia. La Cons ucción de la A qui ec u a Residencial en Gipuzkoa du an e la Época del Desa ollismo. Uni e sidad del País Vasco – Euskal He iko Unibe si a ea. 2012. [44] Bole ín O icial del Es ado. Ley 8/2013, de 26 de junio, de ehabili ación, egene ación y eno ación u banas. 28 h June 2013. [45] Bole ín O icial del Es ado. Real Dec e o Legisla i o 7/2015, de 30 de oc ub e, po el que se ap ueba el ex o e undido de la Ley de Suelo y Rehabili ación U bana. 31s Oc obe 2015. [46] O icial Jou nal o he Eu opean Union. Di ec i e 2012/27/UE. 25 h Oc obe 2012. [47] DesignBuilde simula ion ool. DesignBuilde So wa e L d, UK. A ailable online: h ps:// designbuilde .co.uk [48] IDAE. P oyec o SECH-SPAHOUSEC. Análisis del consumo ene gé ico del sec o esidencial en España. 2011. [49] EVE. Cla es ene gé icas del sec o domés ico en Euskadi. 2013. [50] TRIOLA, M.F.; Es adís ica. 2009. [51] IDAE. Escala de cali icación ene gé ica. Edi icios exis en es. 2011.