1
1On he assessmen o he ene gy pe o mance and en i onmen al beha iou o
2social housing s ock o he adjus men be ween simula ed and measu ed da a: he
3case o mild win e s in he Medi e anean clima e o sou he n Eu ope.
4Rocío Escandón a,*, Ra ael Suá ez a, Juan José Send a a
5a Ins i u o Uni e si a io de A qui ec u a y Ciencias de la Cons ucción, Escuela Técnica Supe io de
6 A qui ec u a, Uni e sidad de Se illa, A . de Reina Me cedes 2, Se ille (41012), Spain.
7Abs ac
8 Cu en Eu opean ene gy policies s ess he po en ial o housing s ock e o i ing o educe global ene gy
9 consump ion. In o de o implemen e icien measu es, i is essen ial o know i s eal in-use ene gy
10 beha iou . In sou he n Eu ope, social housing ep esen s an impo an pe cen age o he esiden ial s ock
11 buil be o e he implemen a ion o he i s ene gy egula ions. Howe e , he e a e ew s udies speci ically
12 analysing hei ene gy beha iou . S anda dized use and occupancy pa e ns a e no usually sui ed o
13 Medi e anean social housing, and his mos ly esul s in es ima ed consump ion exceeding eal
14 consump ion.
15 This esea ch aims o quan i y he he mal com o and ene gy consump ion o he social housing s ock
16 du ing he cha ac e is ically mild win e s in a Medi e anean clima e, based on he moni o ing o
17 ep esen a i e case s udies om sou he n Spain. The esul s show ha he dwellings analysed a e a om
18 con o ming o adap i e com o s anda d EN-15251 and ye , hei limi ed local hea ing sys ems a e a ely
19 u ned on, educing he expec ed ene gy consump ion. In addi ion, eal use and occupancy pa e ns a e
20 de ined o he case s udies, allowing he de elopmen o ene gy simula ion models ha a e be e sui ed
21 o he eal beha iou o his social housing s ock.
22 Keywo ds: social housing s ock, moni o ing, he mal com o , ene gy consump ion, ene gy models
23 calib a ion.
* Co esponding au ho .
E-mail add ess: [email p o ec ed]
2
24 1. In oduc ion
25 O e he las ew decades, Eu opean building egula ions ha e been es ablishing a common amewo k
26 o he encou agemen o ene gy e iciency [1, 2], p omo ing he e o i ing o housing s ock. In he case
27 o sou he n Eu ope, esiden ial s ock buil p io o he i s egula ions which globally limi he ene gy
28 demand o buildings (1976-1979) ep esen s be ween 63 and 76% o he o al housing s ock [3, 4, 5].
29 Mos o hese esiden ial buildings do no inco po a e any speci ic measu es o he mal insula ion in hei
30 en elopes, and a e he e o e obsole e om an ene ge ic poin o iew.
31
32 Social housing ep esen s an impo an pe cen age o he esiden ial s ock in sou he n Eu ope, wi h
33 speci ic socioeconomic cha ac e is ics ha en ail pa icula needs and a use pa e n di e en om he
34 s anda dized one. Field esea ch on social housing in sou he n Eu ope [6, 7, 8, 9] shows ha hese
35 buildings ha e much lowe ene gy consump ion han ha es ima ed by he na ional ene gy assessmen
36 p ocedu es de i ed om Di ec i e 2002/91/EEC and i s upda es [10], based on gene al p emises o
37 in ensi y and habi s o use o buildings and ene gy, no applicable in he case o social housing. An
38 example o his is ha he es ima ed a e age hea ing consump ion o Spanish mul i- amily dwellings
39 (wi h an a e age a ea o 88.7 m2) in he Medi e anean clima e is 3673 kWh pe yea pe dwelling [11],
40 while o low income households (decla ed annual household income below 15,000 €) in sou he n Eu ope
41 i is a ound 1386 kWh pe yea [6].
42
43 The e is ex ensi e esea ch analysing he di e gence be ween eal consump ion and consump ion
44 es ima ed h ough simula ion, ei he by so wa e ecognized by he scien i ic communi y o he o icial
45 so wa e o he di e en coun ies [12,13,14,15]. This ‘pe o mance gap’ has gene ally been a ibu ed o
46 use beha iou and, o a lesse ex en , o a poo iden i ica ion o he cons uc i e cha ac e is ics o
47 buildings [16,17]. In o de o educe he e o a e due o he poo iden i ica ion o he cons uc i e
48 cha ac e is ics, i would su ice o ca y ou he ele an es s in he building, such as ai pe meabili y and
49 in a- ed he mog aphy es s o U-measu emen s [12,18]. Howe e , he e o associa ed o use pa e n
50 de ini ion is much mo e di icul o iden i y and educe.
51
52 Many o he s udies abou occupan beha iou ocus on e o i ed dwellings (low-ene gy) in cen al and
53 no he n Eu ope (clima es wi h se e e win e s), whe e eal consump ion is equen ly much highe han
3
54 es ima ed consump ion due o he Rebound e ec [19, 20]. Howe e , in many o he cases, he opposi e
55 occu s: eal consump ion is much lowe han ha es ima ed, because use s spend a less ime in he
56 dwellings han he ime es ablished by he s anda dized use pa e ns o he di e en coun ies, known as
57 he P ebound e ec [13, 20].
58
59 In he case o social housing, he di e gence be ween eal and es ima ed consump ion is e en g ea e due
60 o an addi ional a iable: when use s occupy hei dwellings, do hey do so in condi ions o he mal
61 com o ? The s anda dized use pa e ns assume ha use s li e in he mal com o condi ions, bu in social
62 housing in sou he n Eu ope, al hough i belongs o a mild win e clima e zone, i is usual o o go com o
63 and li e in uel po e y condi ions [6, 7]. In addi ion, in he speci ic case o sou he n Spain, only 5% o
64 mul i- amily dwellings buil be o e 1979 ha e a cen alized hea ing sys em, and 13% ha e a local hea ing
65 sys em [3]. These da a demons a e ha his housing s ock does no ha e he mal condi ioning means o
66 ensu e he main enance o he mal com o condi ions.
67
68 Due o he uel po e y ound in his social housing s ock, in es men s in ene gy e o i ing (usually
69 ocused on imp o ing he insula ion o i s en elope) [21, 22, 23] a e o en ine ec i e om an ene gy
70 sa ing poin o iew wi h excessi e payback pe iods. Al hough i is e y di icul o educe ene gy
71 consump ion when i is so low, his does no mean ha e o i ing is no bene icial, bu simply means ha
72 ins ead he inal aim ough o be he imp o emen o he mal com o condi ions in he dwellings, whe e
73 he p o i ma gin is e y high [24].
74
75 The main objec i e o his esea ch is o quan i y he eal condi ions o he mal com o in he speci ic
76 con ex o social housing in sou he n Eu ope (mild win e clima e), based on he adap i e model
77 es ablished by s anda d EN-15251 [25] and ene gy consump ion. To his end, a speci ic me hodology
78 based on in-use moni o ing has been de eloped and applied o h ee case s udies ep esen a i e o he
79 ypology, cons uc i e sys em and clima e o sou he n Spain. Th ough his moni o ing, eal use and
80 occupancy pa e ns a e de ined in o de o de elop ene gy simula ion models adjus ed o he eal
81 beha iou o his housing s ock, signi ican ly educing he ‘pe o mance gap’ be ween eal and es ima ed
82 consump ion. The conclusions ob ained a e expec ed o be he po en ial s a ing poin o a u u e change
4
83 in he s anda dized use and occupa ion pa e ns, and he ene gy e o i ing policies o social housing
84 s ock in he Medi e anean clima e.
85
86 2. Me hodology
87 In his wo k he me hodology used has been de eloped based on p e ious esea ch in o de o u he
88 examine he cha ac e iza ion o he cu en ene gy condi ions o social housing s ock in sou he n Eu ope
89 [26]. To do so, a de ailed ene gy and en i onmen al assessmen o he dwellings was ca ied ou h ough
90 moni o ing and subsequen ene gy simula ion.
91 2.1. Moni o ing
92 Di e en moni o ing me hodologies ha e been alida ed by he scien i ic communi y. Two
93 me hodologies a e dis inguished ocusing on he measu emen o en i onmen al pa ame e s [27]: spo
94 measu emen s aken only once o long- e m measu emen s aken a speci ic in e als, om minu es o
95 hou s. The ype o measu emen depends on he ype o da a equi ed o he in es iga ion.
96
97 In gene al, spo measu emen s a e used when he aim is he e alua ion o he mal com o h ough
98 su eys and i is necessa y o link hese wi h he in e nal and ex e nal en i onmen al condi ions o ha
99 poin in ime, o when i is no possible o ca y ou long- e m measu emen s due o economic o in usion
100 cons ain s. Long- e m measu emen s a e undamen al when he pu pose o he moni o ing is o
101 de e mine he beha iou o he building o e a gi en pe iod o ime such as a season o a ull yea . In
102 addi ion, hese ypes o measu es a e necessa y o e alua e he mal com o ollowing an adap i e model,
103 one o he main aims o his esea ch. Thus, his moni o ing me hodology includes long- e m
104 measu emen s ( empe a u e, ela i e humidi y, CO2 le el, and ene gy consump ion) which a e
105 complemen ed wi h spo es s (ai pe meabili y and in a- ed he mog aphy).
106 2.1.1. Ai pe meabili y es and In a- ed he mog aphy
107 Dep essu isa ion es s we e ca ied ou in each case s udy in o de o e i y he ai igh ness o he
108 en elope o he dwellings, acco ding o no m UNE EN-13829 [28]. The Blowe Doo equipmen used in
109 his es was ins alled in he en ance doo o he dwellings and con olled om inside he esiden ial uni .
5
110
111 This in o ma ion was complemen ed wi h he cap u e o images o dwelling en elopes using a
112 he mog aphic came a, ollowing no m UNE EN-13187 [29]. The aim o his es is o sea ch o he mal
113 beha iou pa e ns o he en elopes in he buildings unde s udy.
114 2.1.2. In si u ambien measu emen s
115 The in si u da a collec ion o en i onmen al a iables is essen ial o he e alua ion o he en i onmen al
116 beha iou o he case s udies. Fo his eason, his esea ch has moni o ed ai empe a u e, ela i e
117 humidi y and CO2 le els inside he dwellings. Two WOHLER CDL 210 indoo da a-logge s we e placed
118 in each dwelling (one in he li ing oom and he o he in he main bed oom) o measu e he a iables
119 e e y 30 minu es o a ull yea .
120
121 Da a p o ided by h ee me eo ological s a ions belonging o he Spanish S a e Me eo ological Agency
122 [30] we e used o analyse ex e nal en i onmen al a iables. The a iables o empe a u e, ela i e
123 humidi y, sola adia ion, wind speed and di ec ion, and p ecipi a ion we e measu ed e e y 30 minu es
124 o e a yea .
125
126 In addi ion o he analysis o en i onmen al a iables in win e his esea ch also s udies he le el o
127 he mal com o in he dwellings acco ding o he adap i e model es ablished by EN-15251 [25]. This
128 model o adap i e he mal com o analysis ocuses on he in e ac ion be ween people and buildings
129 ( h ough indoo empe a u e) and on hei expec a ions (based on ou doo empe a u e) [31]. In o de o
130 ob ain he ou doo e e ence empe a u e (TeR), he daily weigh ed a e age was calcula ed acco ding o
131 equa ion 1. Fo he calcula ion o he op imum ope a ing empe a u e (Top), equa ion 2, es ablished in EN-
132 15251, was used. Case s udies belong o ca ego y III, associa ed wi h exis ing buildings, which means an
133 accep abili y band o +/- 4 ºC in ela ion o he calcula ed Top.
134 (1) TeR = (1 - P- * Ted-1 + P * TeR-1
135 whe e:
136 TeR: unning mean empe a u e o oday
137 Ted-1: daily mean ex e nal empe a u e o p e ious day;
138 TeR-1: unning mean empe a u e o he p e ious day;
139 P; is a cons an be ween 0 and 1. Recommended o use 0.8.

6
140 (2) Top = 0.33 * TeR + 18.8
141
142 2.1.3. Ene gy consump ion measu emen s
143 Ano he key aspec o he e alua ion o he beha iou o he case s udies is he in si u da a collec ion o
144 ene gy a iables. As is usually he case wi h social housing in sou he n Spain, none o hese case s udies
145 ha e gas- uelled hea ing sys ems, and he e o e he esea ch ocuses on elec ici y consump ion. The
146 de ailed ene gy consump ion (sub-me e ing) was moni o ed by a gene al consump ion me e in he
147 elec ical panel and se e al indi idual me e s in he socke s o some appliances (mainly local hea ing and
148 cooling sys ems). Consump ion was measu ed in kWh e e y 15 minu es o e a yea . This helps o
149 es ablish ela ionships be ween com o le el, indoo and ou doo condi ions, and use beha iou . In
150 addi ion, in o de o comple e and con as he moni o ed consump ion, his o ical da a o gene al
151 elec ical consump ion o he case s udies we e compiled using pas elec ici y bills.
152 2.1.4. Use pa e ns
153 In o de o a oid he usual di e gence be ween eal and es ima ed consump ion (P ebound e ec ),
154 moni o ing da a was used o de ine use pa e ns. A mixed me hodology was used in his esea ch o
155 cap u e he echnical and social aspec s om use p ac ices quali a i ely and quan i a i ely [32]. In o de
156 o explain and alida e hese aspec s he quali a i e da a ob ained om he use su eys and he use
157 pa e ns ob ained om he quan i a i e da a measu ed in he dwellings ( empe a u e, CO2 concen a ion
158 and elec ic consump ion) we e c oss- e e enced.
159
160 In o de o analyse he quan i a i e da a, he hou ly a e age alues o he en i e win e pe iod we e used
161 [33]: empe a u es in deg ees cen ig ade (ºC), CO2 concen a ion in pa s pe million (ppm), hea ing local
162 sys ems ope a ion on a scale o 0 o 1 (0 means o and 1 on) and gene al elec ical consump ion in
163 kilowa -hou (kWh). This esea ch ocused on he analysis o weekdays, which is when pa e ns a e
164 usually epea ed he mos , whe eas a weekends beha iou is usually mo e i egula .
165
166
7
167 2.2. Ene gy simula ion
168 Fo he simula ion o he en i onmen al and ene gy beha iou o he case s udies, he me hodology
169 equi es wo ini ial asks: cons uc ion o he ene gy model o ep oduce he ooms s udied and an ene gy
170 model adjus men p ocess.
171 2.2.1. Ene gy model cons uc ion
172 The ene gy model was de eloped wi h so wa e DesignBuilde ( .4.7.0.027), which uses Ene gyPlus
173 [34], a simula ion p og am ecognized by he US DOE [35].
174
175 Models ec ea ing he whole dwelling and i s bounda y condi ions we e gene a ed o each o he case
176 s udies moni o ed. The clima e da a used o he ene gy simula ion we e p o ided by h ee me eo ological
177 s a ions belonging o he Spanish S a e Me eo ological Agency [30], whose da a we e p e iously
178 alida ed compa ing hem wi h spo measu emen s aken ou side he case s udies. The cons uc i e
179 de ini ion o he en elope (sec ion 3.2) and he use and occupa ion pa e ns o he case s udies (sec ion
180 3.3) we e also aken in o accoun o he de elopmen o he model. In addi ion o occupa ion habi s, he
181 models also include ligh ing pa e ns, use o local hea ing sys ems, na u al en ila ion (window opening),
182 and use o sun p o ec ion (shu e s, awnings…).
183
184 Ano he aspec ha helps o educe he unce ain y o he model is he in oduc ion o he a e o
185 in il a ion measu ed in he ai pe meabili y es .
186 2.2.2. Ene gy model adjus men
187 As Cip iano e al. [36] co ec ly poin ou , he li e a u e e iew shows he absence o a gene ally accep ed
188 me hodology o he calib a ion o ene gy building models. Howe e , some au ho s de ine calib a ion
189 me hodologies based on he manual adjus men o he models [37, 18], while o he s base hem on a mul i-
190 s age guided p ocedu e [36, 38].
191 This is a complex ask, in ol ing many pa ame e s wi h a signi ican deg ee o unce ain y. Many o he
192 s udies on calib a ion ocus on he impo ance o in si u measu emen s as a ool o he adjus men o
193 ene gy simula ions [36, 38]. The e o e, as desc ibed in he p e ious sec ion, all he in o ma ion collec ed
194 in si u in his esea ch was inco po a ed in o he ene gy models. Howe e , he e a e s anda d c i e ia o
8
195 de e mine when a model can be conside ed o be calib a ed, ocusing on how he model adjus s o
196 measu ed da a and analysing he deg ee o e o [39, 40, 41].
197
198 Usually, he li e a u e abou he calib a ion o ene gy model ocuses on he adjus men o he hea ing
199 ene gy consump ion [36]. Howe e , his would make no sense in hese case s udies, since he hea ing
200 ene gy consump ion is e y insigni ican in he social housing p o ile in sou he n Spain ( he dwellings a e
201 usually in condi ions o ee e olu ion). The e o e, his esea ch ocuses on he manual adjus men o
202 en i onmen al a iables du ing he win e pe iod, speci ically in he in e io empe a u e (as his is he
203 a iable in ol ed in he mal com o analysis). The esul s o ene gy simula ions and moni o ing we e
204 compa ed bo h g aphically ( o a ep esen a i e week) and s a is ically ( o he en i e win e pe iod). The
205 g aphs p o ide in o ma ion abou he ime pe iods in which he wo da ase s di e ge, while he s a is ical
206 indices p o ide a nume ical index o assess how closely he simula ion ma ches he measu ed da a.
207
208 ASHRAE Guideline 14-2002 [39] was ollowed in he s a is ical alida ion o he model, es ablishing wo
209 e o indica o s, he Mean Bias E o (MBE) and he Coe icien o Va ia ion o he Roo Mean Squa e
210 E o (CVRMSE) alues, acco ding o equa ions (3) and (4):
211 (3) MBE =
Ni
i=1(Mi Si)
Ni
i=1Mi
212 whe e:
213 Mi: measu ed da a a ins ance n;
214 Si: simula ed da a a ins ance n;
215 Ni: coun o he numbe o da es used in he calib a ion.
216 (4) CVRMSE =
Ni
i=1
(  )2
Ni
1
Ni Ni
i=1Mi
217 whe e:
218 Mi: measu ed da a a ins ance n;
219 Si: simula ed da a a ins ance n;
220 Ni: coun o he numbe o da es used in he calib a ion
221
222 ASHRAE Guideline 14 conside s a building model calib a ed when hou ly MBE alues all wi hin ±10%
223 and hou ly CVRMSE alues all below 30%. The MBE p o ide an indica ion o e o s a e aged o he
9
224 mean o measu ed alues bu may be in luenced by o se ing e o s, while he CVRMSE index is a
225 measu e o accumula ed e o no malized o he mean o he measu ed alues.
226
227 3. Case s udies: social housing
228 Whene e a de ailed analysis is ca ied ou , in ol ing long campaigns o moni o ing and e alua ion o he
229 he mal en elope o buildings, i is complica ed o ha e a la ge sample o case s udies, due o inancial
230 and ime cons ain s [27]. Fo his eason, based on a p e ious analysis o he social housing s ock o
231 sou he n Spain [42], he s a is ically mo e ep esen a i e mo phological and cons uc i e building
232 ypologies we e de e mined and some neighbou hoods we e selec ed by he Andalusian Go e nmen as
233 cases o in e es o i s ene gy e o i ing plan [43]. Th ee dwellings belonging o hese cases o in e es
234 we e selec ed as case s udies in his esea ch. The main cha ac e is ics o he clima e in which hey a e
235 loca ed, mo phological ypology, cons uc i e cha ac e is ics o i s en elope and he use pa e ns will be
236 desc ibed below.
237 3.1. Loca ion and clima e
238 The h ee case s udies a e loca ed in places wi h a Medi e anean clima e ( igu e 1a), speci ically h ee
239 Spanish ci ies wi h he lowes alues in he scale o win e clima e se e i y: Se ille (win e clima e zone
240 B), Huel a and Malaga (win e clima e zone A) ( igu e 1b). Table 1 summa izes he main clima e
241 cha ac e is ics o hese loca ions.
242
(a)
(b)
Figu e 1. Medi e anean clima e map: (a) in Eu ope, (b) in sou he n Spain [44].
243
16
342 hou ly empe a u e, ela i e humidi y and CO2 concen a ion e olu ion du ing he coldes mon h o he
343 win e pe iod. In he empe a u e g aphs an adap i e com o band has been included ollowing s anda d
344 EN-15251. The CO2 g aphs also included a e e ence limi below which indoo ai quali y is conside ed
345 good (1200 ppm), ollowing he FSIAQ (Finnish Classi ica ion o indoo clima e 2000 [47]) s anda d o
346 ca ego y S3 (a less demanding ca ego y which may be associa ed wi h exis ing buildings).
347
348 In case 1 (Se ille), du ing he win e pe iod moni o ed (Decembe 2014 - Feb ua y 2015) he coldes
349 mon h was Janua y. Du ing his pe iod ( igu e 5), wi h mild ou doo empe a u es anging om 5 ºC o
350 18 °C, inside he dwelling he empe a u es we e always a ound 14-15 °C, and so always below he lowe
351 limi se by adap i e he mal com o s anda d EN-15251 (a ound 18 °C). In pa icula , he a e age
352 Janua y in e io empe a u e was 15 °C in he li ing oom and 14.2 °C in he main bed oom. The
353 beha iou o he wo moni o ed ooms was e y simila , wi h he excep ion o some imes o nigh when
354 a local hea ing sys em was used in he bed oom. As i is a e y shaded dwelling, he e is no much sola
355 gain du ing he day epea edly making indoo empe a u es all below ou doo ones.
356
357 Analysing o he en i onmen al a iables ( igu e 6), an in e nal ela i e humidi y be ween 60 % and 80 %
358 was ound, compa ed o an ex e nal one a ying om 40 % o 90 %. The in e io CO2 concen a ion
359 shows wo e y di e en ways o use o he ooms moni o ed. While he li ing oom is ou inely kep
360 below 1250 ppm (wi h he excep ion o high occupancy pe iods) due o spo adic use, he bed oom
361 eaches 5000 ppm o a high pe cen age o nigh s because o mo e con inued use.
362
363 In case 2 (Malaga), du ing he win e pe iod moni o ed (Decembe 2015 - Feb ua y 2016) he coldes
364 mon h was Feb ua y. In his mon h ( igu e 7), wi h wa m ou doo empe a u es anging om 10 ºC o 20
365 ºC, he empe a u es inside he dwelling we e always a ound 18-19 ºC, and almos always below he
366 lowe limi se by he adap i e he mal com o s anda d EN-15251 (a ound 20 °C in his case). The
367 a e age in e io empe a u e o Feb ua y was 18.1 ºC in he main li ing oom and 18.8 ºC in he main
368 bed oom. The beha iou o he wo moni o ed ooms was e y simila , wi h he excep ion o some
369 pe iods whe e a local hea ing sys em was used in he main li ing oom. This dwelling main ains a e y
370 s able empe a u e h oughou he whole day.
371

17
372
373
374 Figu e 5. Case 1 - Se ille: hou ly empe a u e e olu ion du ing Janua y.
375
376 Figu e 6. Case 1 - Se ille: hou ly ela i e humidi y and CO2 concen a ion e olu ion du ing Janua y.
377
378 Figu e 7. Case 2 - Malaga: hou ly empe a u e e olu ion du ing Feb ua y.
18
379
380 Figu e 8. Case 2 - Malaga: hou ly ela i e humidi y and CO2 concen a ion e olu ion du ing Feb ua y.
381
382 Figu e 9. Case 3 - Huel a: hou ly empe a u e e olu ion du ing Janua y.
383
384 Figu e 10. Case 3 - Huel a: hou ly ela i e humidi y and CO2 concen a ion e olu ion du ing Janua y.
385
19
386 An in e nal ela i e humidi y be ween 50 % and 80 % was ound ( igu e 8), compa ed o an ex e nal one
387 anging om 30 % o 95 %. The in e io CO2 concen a ion shows a e y simila way o use o he wo
388 ooms moni o ed, p obably because he doo s o all he ooms a e always open linking hem. Bo h a e
389 ou inely kep below 1250 ppm due o low occupancy o hese ooms. Ano he key aspec o he good
390 quali y o he in e io ai in his case s udy is he con inuous ope a ion o he na u al en ila ion sys em.
391 This sys em keeps a low con inuous en ila ion a e h ough he ki chen window and ex ac ed by he
392 en s in he ba h ooms.
393
394 In case s udy 3 (Huel a), du ing he moni o ed win e pe iod (Decembe 2016 - Feb ua y 2017) he
395 coldes mon h was Janua y. Du ing his mon h ( igu e 9) he ex e nal empe a u es we e wa m, be ween
396 10 ºC and 20 °C, while inside he dwelling he empe a u es we e measu ed a a ound 13-14 °C, a below
397 he lowe limi se by adap i e he mal com o s anda d EN-15251 (a ound 20 °C in his case). In
398 pa icula , he Janua y a e age in e io empe a u e was 14.7 °C in he li ing oom and 13 °C in he main
399 bed oom. The wo moni o ed ooms exhibi almos pa allel beha iou , wi h he excep ion o nigh - ime,
400 when he empe a u e dec ease in he bed oom is g ea e .
401
402 Analysing o he en i onmen al a iables ( igu e 10), an in e nal ela i e humidi y be ween 50 % and 85
403 % was ound, compa ed o an ex e nal one a ying om 30 % o 70 %. The in e io CO2 concen a ion
404 shows wo sligh ly di e en ways o using he moni o ed ooms. Du ing he day he doo s o all ooms
405 a e kep open which homogenizes he quali y o he ai . Howe e , a nigh he bed oom eaches highe
406 CO2 concen a ions, up o 2000 ppm, while he li ing oom s ays below 1250 ppm. In gene al, he in e io
407 ai main ains good quali y due o he ac ha he use s en ila e he dwelling na u ally e e y mo ning
408 ( om 8.00 o 11.00 h, as seen om he su ey).
409 4.2. Com o analysis
410 The he mal com o le el in he h ee case s udies du ing he win e pe iod has been analysed acco ding
411 o he adap i e model es ablished by s anda d EN-15251 [26]. In addi ion, he pe cen age o hou s o
412 discom o and he a e age de ia ion o he in e io empe a u es wi h espec o he com o band ha e
413 been con as ed wi h he pe cen age o hou s o use o he local hea ing sys ems in each dwelling.
414
20
Pe iod
Decembe
2014
Janua y
2015
Feb ua y
2015
Win e
2014/15
% hou s o discom o
Li ing oom
99.7
100
100
99.9
Bed oom 1
100
100
99.5
99.9
A e age de ia ion Indoo empe a u e – Com o band (ºC)
Li ing oom
-3.4
-3.7
-2.7
-3.3
Bed oom 1
-3.9
-4.2
-3.9
-3.8
% hou s o use o local hea ing sys ems
Po able
ai hea e
5.4
4.4
3.3
4.4
Bed oom 2
hea pump
9.6
14.7
4.6
9.9
415 (a)
Pe iod
Decembe
2015
Janua y
2016
Feb ua y
2016
Win e
2015/16
% hou s o discom o
Li ing oom 1
92.5
87.5
99.4
92.9
Bed oom 1
*
100
95.4
96.1
A e age de ia ion Indoo empe a u e – Com o band (ºC)
Li ing oom 1
-1.3
-1.6
-2.0
-1.6
Bed oom 1
*
-1.3
-1.3
-1.3
% hou s o use o local hea ing sys ems
Li ing oom 1
hea pump
12.9
15.2
18.2
15.3
Li ing oom 2
hea pump
13.6
17.7
20.8
17.3
* Loss o moni o ing da a due o equipmen ailu e.
416 (b)
Pe iod
Decembe
2016
Janua y
2017
Feb ua y
2017
Win e
2016/17
% hou s o discom o
Li ing oom
100
100
100
100
Bed oom 1
98.7
100
100
99.5
A e age de ia ion Indoo empe a u e – Com o band (ºC)
Li ing oom
-3.6
-4.6
-3.0
-3.8
Bed oom 1
-4.8
-5.8
-4.2
-5.0
% hou s o use o local hea ing sys ems
Li ing oom
hea pump
7.1
22
2.8
10.9
Bed oom 1
hea pump
0.1
0.3
0.1
0.2
417 (c)
418 Figu e 11. E alua ion o he mal com o and use o hea ing sys ems du ing he win e pe iod: (a) case 1 -
419 Se ille, (b) case 2 - Malaga, (c) case 3 - Huel a.
420
421 In cases 1 and 3 (Se ille and Huel a), a high pe cen age o hou s o discom o was de ec ed h oughou
422 he win e pe iod in he wo ooms moni o ed, a ound 100 % ( igu e 11a, 11c). In case 3, despi e ha ing
423 highe ex e nal empe a u es han in case 1, in e io empe a u es e en u he emo ed om he com o
424 band a e obse ed. This is mainly due o hea losses o case 3 h ough he oo , as i is on he op loo . In
425 bo h cases, in spi e o being in a clea si ua ion o discom o , he pe cen age o hou s o use o local
426 hea ing sys ems is e y low. Du ing he coldes pe iod (Janua y) his pe cen age is less han 15 % in
21
427 Se ille and 22 % in Huel a. In case 3 he use o he local hea ing sys em in he main bed oom is almos
428 ze o al hough his is he oom u hes away om he com o band.
429
430 In case 2 (Malaga), as he clima e is milde han in case 1 (bu simila o case 3), he pe cen age o hou s
431 o discom o is sligh ly lowe han in he o he cases bu i is s ill high. This pe cen age was abo e 90 %
432 du ing he whole win e pe iod in he wo ooms moni o ed ( igu e 11b). Al hough he pe cen age o
433 hou s o discom o does no di e much om he o he cases, he g aph shows ha he a e age de ia ion
434 o he in e io empe a u e wi h espec o he com o band is much lowe in case 2 (-1.6 ºC in he li ing
435 oom and -1.3 °C in he bed oom) han in cases 1 (-3.3 °C in he li ing oom and -3.8 °C in he bed oom)
436 and 3 (-3.8 °C in he li ing oom and -5.0 °C in he bed oom). Howe e , in his case he pe cen age o
437 hou s o use o local hea ing sys ems is highe han in case 1 and simila o he case 3 li ing oom du ing
438 he colde pe iods, eaching 20 % in Feb ua y. In spi e o his, he pe cen age o hou s o use o he
439 hea ing sys ems does no nea ly co e he pe cen age o hou s o discom o .
440 4.3. Ene gy consump ions
441 The elec ici y consump ion o he h ee case s udies has been analysed comp ehensi ely and in de ail
442 du ing he win e pe iod, ocusing on he consump ion o he local hea ing sys ems. In o de o e alua e
443 he hea ing consump ion, he addi ion o he consump ion o all local hea ing sys ems (kWh) has been
444 di ided by he hea ed a ea, which in his case is he o al a ea o he dwelling excep o he ki chen and
445 ba h ooms ( he doo s o he ooms a e always kep open). In he h ee case s udies, e y low le els o
446 o al consump ion and no signi ican pe cen ages o he use o local hea ing sys ems we e de ec ed.
447
448 In cases 1 and 2, i is wo h no ing ha he Ho Wa e p oduc ion sys em is elec ic. In case 1 (Se ille)
449 speci ically he wa e hea e shows e y signi ican consump ion ( igu e 12a), a ound 24 %, well abo e
450 hea ing consump ion, a ound 13 %. Also in his case, he high consump ion o wo compu e s, a ound 19
451 %, is signi ican since he use s usually wo k a home. In case 2 (Malaga) ( igu e 12b), a balanced
452 dis ibu ion o consump ion can be obse ed, in which hea ing accoun s o a ound 20 % o o al
453 consump ion. In case 3 (Huel a) ( igu e 12c), he li ing oom hea ing sys em consump ion is no able as i
454 accoun s o 29 % o he o al consump ion o he dwelling. In his case, he global consump ion o all
455 de ices wi hou indi idual me e s (‘o he s’ in igu e 12) ep esen s a mo e signi ican pe cen age han in

22
456 he o he cases (52 %), due o he ac ha i has less elec ic equipmen (Ho Wa e is hea ed by gas) and
457 he consump ion om ligh ing and cooking acqui es g ea e ele ance.
458
459 Looking a he o al consump ion da a, i is appa en ha case 1 (occupied by a young couple who wo k a
460 home) consumes he same as case 2 (occupied by a amily o 4 membe s) and almos double ha o case
461 3 (occupied by a e i ed couple). The economic le el o he use s o case 1 is sligh ly highe han he
462 o he s. When analysing he hea ing consump ion, case 2 consumes he highes amoun in ela ion o he
463 space hea ed, as well as aking up he highes pe cen age o use o he hea ing sys ems ( igu e 11).
(a)
Pe iod
To al consump ion
(kWh)
Hea ing consump ion
(kWh/m2 hea ed)
Decembe 2014
355.89
1.12
Janua y 2015
394.80
1.15
Feb ua y 2015
305.84
0.47
Win e 2014/15
1056.53
2.75
(b)
Pe iod
To al consump ion
(kWh)
Hea ing consump ion
(kWh/m2 hea ed)
Decembe 2015
268.01
1.46
Janua y 2016
327.23
1.57
Feb ua y 2016
410.44
1.42
Win e 2015/16
1005.68
4.45
(c)
Pe iod
To al consump ion
(kWh)
Hea ing consump ion
(kWh/m2 hea ed)
Decembe 2016
165.87
0.77
Janua y 2017
206.46
1.76
Feb ua y 2017
133.68
0.38
Win e 2016/17
506.01
2.91
464 Figu e 12. E alua ion o elec ic consump ion du ing he win e pe iod: (a) case 1 - Se ille, (b) case 2 -
465 Malaga, (c) case 3 - Huel a.
23
466 Table 5. S a is ical alida ion o he ene gy models.
MBE Li ing oom
MBE Bed oom
CV Li ing oom
CV Bed oom
Case 1 (Se ille)
-5.33 %
-5.06 %
10.14 %
11.27 %
Case 2 (Malaga)
-0.23 %
-3.17 %
4.08 %
4.33 %
Case 3 (Huel a)
-4.78 %
-1.73 %
8.22 %
5.61 %
ASHRAE S anda ds
MBE < 10 %
CV(RMSE) < 30 %
467
468
469 (a)
470
471 (b)
472
473 (c)
474 Figu e 13. G aphic alida ion o he ene gy models: (a) case 1 - Se ille, (b) case 2 - Malaga,
475 (c) case 3 - Huel a.
24
476 4.4. Ene gy model adjus men
477 Following da a collec ion o all case s udies, he ene gy models we e cons uc ed and subsequen ly
478 simula ed o e i y he adjus men o he measu ed da a. Fi s , he esul s o ene gy simula ions and
479 moni o ing o a ypical week we e compa ed in g aphs. Acco ding o he esul s ob ained, he ene gy
480 model achie es he adjus men o a iables on which he e is some deg ee o unce ain y (due o he
481 impossibili y o measu emen ), such as he na u al en ila ion a e h ough opening windows and he
482 occupan s’ me abolic a e. The inal esul is an adequa e g aphic adjus men o he ene gy models ( igu e
483 13), which ook in o accoun he deg ee o e o o he measu emen equipmen (+/- 0.5 ºC), ep esen ed
484 in he g aph by e o ba s.
485
486 In addi ion, he s a is ical alida ion o he ene gy models has been e i ied du ing he whole win e
487 pe iod, ollowing he indica o s es ablished by he ASHRAE [35] ( able 5). The adjus men be ween he
488 ene gy models and he measu ed da a shows ha he moni o ing o he buildings is essen ial o ob aining
489 a alid adjus men , and so g ea ly educing he numbe o unce ain a iables. The e o e, o he ene gy
490 e alua ion o e o i ing measu es i is essen ial o ha e a calib a ed simula ion model, e en mo e so in
491 case s udies wi h a use pa e n so a emo ed om he s anda ds.
492
493 5. Conclusions
494 In sou he n Eu ope, he e is ex ensi e social housing s ock wi h a no able need o e o i ing due o
495 ene gy obsolescence. Ene gy simula ion inco po a ing s anda dized use pa e ns canno be used o
496 asce ain he mos app op ia e measu es o hese pa icula dwellings o a oid he P ebound e ec a e
497 e o i ing. Gi en ha he case s udies e alua ed demons a e a a ie y o use p o iles, e o i ing
498 decisions should no be simpli ied by using a single use p o ile. Na ional egula ions should de ine
499 di e en use p o iles, including he numbe and age o use s, he in ensi y o use o he dwellings, and
500 he clima ic zone in which i is loca ed ( o conside he di e en c i e ia o he mal com o ).
501
502 A e moni o ing he en i onmen al and ene gy beha iou o some case s udies (du ing he win e
503 pe iod), he ollowing esul s ha e been ob ained:
25
504 Despi e he mild clima e condi ions du ing he ha shes win e mon hs, he a e age indoo
505 empe a u es measu ed a e low, oscilla ing be ween 13 ºC (Huel a) and 18.8 ºC (Malaga).
506 The low alues o he in e io empe a u e mean ha o mos o he ime he dwellings a e in
507 discom o condi ions, and he pe cen age o hou s ou side he com o band anges om 92.9 %
508 (Malaga) o 100 % (Huel a).
509 In spi e o hese inadequa e indoo he mal condi ions, use s gene ally ejec he use o hei
510 local hea ing sys ems, mainly o inancial easons. The pe cen age o hou s o use o he local
511 hea ing sys ems anges om 9.9 % (Se ille) o 17.3 % (Malaga).
512 Despi e he wa m Medi e anean win e s, he low ene gy pe o mance o he p e-1980 social housing
513 en elope (wi hou any kind o he mal insula ion) causes indoo he mal condi ions a emo ed om
514 com o . This, added o he lack o global and e icien hea ing sys ems and o inancial cons ain s o
515 use s, means ha his housing s ock is comple ely exposed o he isk o alling in o a si ua ion o uel
516 po e y. The cu en ene gy policies mus ake in o accoun he en i onmen al beha iou o social
517 housing in sou he n Spain desc ibed in his esea ch, and ensu e ha ene gy e o i ing g an s ocus on
518 passi e measu es ha en ail he imp o emen o he mal com o ins ead o ene gy sa ings.
519
520 This esea ch also demons a es ha he moni o ing o in-use dwellings pe mi s enough in o ma ion o be
521 ob ained o gene a e ene gy simula ion models ha minimize he ‘pe o mance gap’ be ween eal and
522 es ima ed beha iou . This is essen ial i e o i ing decisions a e o be based on he esul s o ene gy
523 simula ions.
524
525 Acknowledgemen s
526 This esea ch was pa ially unded by he V In e nal Resea ch Plan o he Uni e sidad de Se illa and he
527 TEP-130 esea ch g oup.
528
529 Pa o his esea ch was ca ied ou wi hin he esea ch p ojec s: ‘In e ención en ba iadas esidenciales
530 obsole as: Manual de buenas p ác icas’ ( e . 2013-0000006939), unded by he Andalusian Regional
531 Go e nmen ; and ‘REFAVIV: Rehabili ación ene gé ica de las achadas de i iendas sociales
532 de e io adas aplicando p oduc os inno ado es nacionales (DIT) y eu opeos (DITE)’ ( e . BIA2012-