Jou nal o Cleane P oduc ion 349 (2022) 131491
A ailable online 22 Ma ch 2022
0959-6526/© 2022 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC license (h p://c ea i ecommons.o g/licenses/by-
nc/4.0/).
Design o dis ic hea ing ne wo ks in buil en i onmen s using GIS: A case
s udy in Vi o ia-Gas eiz, Spain
Mikel Lumb e as
a
,
*
, Gonzalo Dia ce
a
, Koldobika Ma in-Escude o
a
, Al a o Campos-Celado
b
,
Pello La inaga
a
a
ENEDI Resea ch G oup, Ene gy Enginee ing Depa men , Facul y o Enginee ing o Bilbao, Uni e si y o he Basque Coun y (UPV/EHU), Pza. Ingenie o To es Que edo
1, Bilbao, 48013, Spain
b
ENEDI Resea ch G oup, Ene gy Enginee ing Depa men , Facul y o Enginee ing o Bilbao, Uni e si y o he Basque Coun y (UPV/EHU), A da. O aola 26, Eiba ,
20600, Spain
ARTICLE INFO
Handling Edi o : Mingzhou Jin
Keywo ds:
GIS
Indus ial was e hea
Da a-d i en model
LiDAR
Dis ic hea ing
ABSTRACT
The e icien in eg a ion o high le els o indus ial was e hea in low empe a u e dis ic -hea ing ne wo ks is a
p omising echnique ha equi es speci ic me hodologies o i s sa is ac o y implemen a ion. This pape p esen s
a no el me hodology o assessing he ene gy and economic easibili y o new dis ic -hea ing ne wo ks in
exis ing u ban a eas o he in eg a ion o indus ial was e hea sou ces. The me hodology consis s in an inno-
a i e mul is ep p ocedu e using geog aphic in o ma ion sys ems and da a analysis ools, combining geo e e -
enced da a abou buildings, indus ies and oads. The spa ial dis ibu ion o he analysis a ea is di ided in o
smalle bu e s and g ids, as a esul , he ou ing design o he pipelines ha makes up he dis ic -hea ing o-
pology is ob ained unde se e al assump ions. The me hodology p o ides he mos sui able a ea choice o he
deploymen o a dis ic -hea ing, also implemen ed wi h a mul i-s ep algo i hm o ou ing he pipelines o he
ne wo k. This me hodology is applied o a pa icula case s udy loca ed in Vi o ia-Gas eiz (no he n Spain).
Di e en con igu a ions o he dis ic hea ing ne wo k a e ob ained wi h leng hs o he ne wo k a ying om 8
o 27 km. Payback alues nea o six yea s a e achie ed in mos o he dis ic -hea ing ne wo k con igu a ions.
The maximum payback pe iod ob ained wi hin he con igu a ions is 8.5 yea s. An economic sensi i i y analysis is
p esen ed o he p oposed op imal dis ic -hea ing ne wo k con igu a ion. The p oposed me hodology could be
eplica ed o di e en case s udies as long as he inpu da a is a ailable o he use .
1. In oduc ion
Ene gy consump ion in buildings cu en ly accoun s o a ound 40%
o he o al ene gy consump ion in he Eu opean Union (EU)
(P´
e ez-Lomba d e al., 2008). In he pa icula case o esiden ial
buildings, 57% o he o al inal ene gy consump ion is used o space
hea ing and 25% o domes ic ho wa e (DHW) (Bala as e al., 2005).
Mo e han 50% o his ene gy consump ion is nowadays ul illed wi h
na u al gas and elec ici y (Eu opean Commission, 2019). The e o e, he
implemen a ion o al e na i e ene gy sou ces in buildings is i al o
main ain a sus ainable en i onmen in ci ies and achie e he objec i es
o ca bon neu al en i onmen o 2050 (Eu opean Commission, 2017;
Zhang e al., 2020) in he EU.
Dis ic Hea ing (DH) ne wo ks can p o ide an e icien al e na i e
o indi idual ins alla ions in densely popula ed u ban a eas (Ch is ian
Holms ed Hansen, 2018). Today, DH ne wo ks p o ide mo e han 13%
o he hea ing ene gy o buildings in he EU (Lund, 2007). Besides, new
DH ne wo ks enable he connec ion o low g ade and decen alized
enewable ene gy sou ces (Lund e al., 2018; Wen e al., 2021; Yılmaz
Balaman and Selim, 2016), which o e s he possibili y o employ was e
hea s eams as a sus ainable hea sou ce.
The e a e di e en sou ces o was e hea a ailable in ci ies o nea
hem. Indus ial esidual he mal ene gy —commonly e e ed o as
indus ial was e hea (IWH) — is one o he mos common. Recen
s udies show ha he amoun o IWH cu en ly ejec ed o he en i-
onmen accoun s o 20–50% o he indus ial ene gy consump ion
ac oss EU (B ueckne e al., 2014). F om ha , 18–30% o i could be
e-used in a echnically easible way (B ueckne e al., 2014; Vance
e al., 2019). Since he emission empe a u e o a ound 65% o his IWH
emains below 200 ◦C (Anku Kapil, Igo Bula o , Robin Smi h, 2017),
i s eu iliza ion o elec ici y p oduc ion is hinde ed; howe e , hese
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (M. Lumb e as).
Con en s lis s a ailable a ScienceDi ec
Jou nal o Cleane P oduc ion
jou nal homepage: www.else ie .com/loca e/jclep o
h ps://doi.o g/10.1016/j.jclep o.2022.131491
Recei ed 18 Augus 2021; Recei ed in e ised o m 8 Feb ua y 2022; Accep ed 20 Ma ch 2022
Jou nal o Cleane P oduc ion 349 (2022) 131491
2
empe a u es a e easible o DH sys ems (Fi ´
o e al., 2020; Mose and
Lassache , 2020), as demons a ed in he li e a u e. Fo ins ance, in a
case s udy (Ziemele e al., 2018), he use o IWH was able o co e om
4% o 12% o he sys em load in he hea ing season o a DH ne wo k
loca ed in La ia. Simila ly, he po en ial o in eg a ing IWH (and sola
he mal sys ems) in o he DH o Ge many was s udied (Pelda e al.,
2020). A g ea po en ial o exploi a ion was iden i ied.
None heless, due o he spa ial dimension o he p oblem, he use o
geo e e enced da a is essen ial o he assessmen and p elimina y
design o DH solu ions. The spa ial con igu a ion o he buildings con-
nec ed o he DH and he oad ne wo k will de ine he leng h and,
consequen ly, he in es men cos o he DH deploymen . These speci ic
design equi emen s can be aced by Geog aphical In o ma ion Sys ems
(GIS), which has been p o en o be e icien o di e en applica ions a
u ban scale (Ali e al., 2018; Ma ques-Pe ez e al., 2020; To abi Mog-
hadam e al., 2018), including DH ne wo k deploymen . As ep esen-
a i e examples, he po en ial o ex ension o DH ne wo ks in Uni ed
S a es based on densi y demand analysis and high- esolu ion GIS da a
was analyzed (Gils e al., 2013). This s udy concluded ha hea dis i-
bu ion cos s and, consequen ly, economic easibili y o hese sys em, a e
s ongly dependen on he hea demand densi y applied. Ano he
e e ence (Chiche in e al., 2018) p oposed a GIS-based model o
combine annual cos analysis o DH sys ems wi h geo e e enced da a in
o de o help in he decision-making p ocess o DH ne wo k planning.
The me hod was applied in he DH ne wo k in Omsk (Russia) and he
pape highligh ed he ad an ages o using GIS o u ban planning.
Ano he s udy (Nielsen and M¨
olle , 2013) de eloped a GIS-based anal-
ysis o examine he po en ial o expanding DH in Denma k by de el-
oping a de ailed cos analysis. The pape concluded ha some
high-densi y a eas in Denma k a e sui able o ins alling DH ne wo ks.
Finally, in he las e e ence analyzed (Un e n¨
ah e e al., 2017), a GIS
based me hodology was p oposed o spa ially assess he in eg a ion o
DH ne wo ks in u ban ene gy sys ems, based on geo e e enced da a o
he geo he mal ene gy esou ce and he oad ne wo k, among o he s. All
hese s udies showed he ad an ages o combining GIS ools wi h
adi ional cos analysis in DH ne wo ks. None heless, no s udies a e
ound ha p esen he comple e me hodology o an economic easi-
bili y s udy o a new DH sys em.
When he ela ed li e a u e is assessed, i can be obse ed ha mos
o he a icles a e de o ed o non-cons uc ed a eas. To da e, DH sys ems
ha e been usually designed o new u ban de elopmen s (Nguyen e al.,
2020; Tol, 2015), whe e he buildings ha e low hea ing ene gy con-
sump ion, and he dwellings a e s uc u ed in an o de ly manne . In
hese cases, he ou ing o he pipelines o a new DH ne wo k is usually
e y in ui i e and manageable. Howe e , he cu en a e o inc ease o
new buildings’ cons uc ion is a ound 2% pe yea in he EU (Eu os a ,
2020). This means ha mo e han hal o he building s ock in a
mid- e m ange is al eady cons uc ed. Acco dingly, ene gy supply
scena ios o he ollowing yea s need o conside ene gy e iciency
ac ions in al eady occupied a eas and, in u n, simple design p ocedu es
ha a e capable o ake hose aspec s in o accoun a e also equi ed.
As a esul , his s udy p esen s a GIS-based mul is ep me hodology
sui able o design new DH ne wo ks wi hin a buil en i onmen . An
indus ial was e hea sou ce is in oduced as a ixed spa ial cons ain .
The me hod conside s he a ailable oad ne wo k o ou ing he DH
pipes, a wo king scale ha conside s small sub-a eas and a no el
mul is ep me hodology o g aph c ea ion and algo i hm ou ing. The
p ocedu e is hen applied in a pa icula case s udy in Vi o ia-Gas eiz
(Basque Coun y, no he n Spain). I is in ended o p o ide a simple
me hod o p elimina y assess he economic easibili y o new DH sys-
ems ed wi h IWH ha con ibu e o mo e sus ainable ci ies.
2. Me hodology
This pape p esen s a no el me hodology o he op imiza ion o he
design o ou ing p ocess o he pipelines in a new DH ne wo k. The
p oposed me hodology s a s wi h he de e mina ion o he wo
ollowing bounda y condi ions: (i) Iden i ica ion o he IWH a ailable
nea o he new DH ne wo k and (ii) de ini ion o he hea ing ene gy
demand in he buildings o be co e ed by he DH ne wo k. The e a e
a ious op ions o he calcula ion o each o he bounda y condi ions
and he me hodologies chosen o he case-s udy p esen ed in his pape
will be shown in Sec ions 3.1 and 3.2. I he bounda y condi ions a e p e-
de ined, he me hodology p esen ed in he ollowing lines would be
alid o any o he case-s udy. Thus, his sec ion will conside ha he
IWH a ailable is al eady cha ac e ized and ha he demand in all he
buildings in he egion a e also a ailable. Then, Sec ion 3 will explain
he me hodologies chosen o his case-s udy.
The me hodology s a s wi h he de ini ion o he a eas in which he
algo i hms o he ou ing o he pipelines will be applied. This i s s ep
is de ailed in Sec ion 2.1. Then, Sec ion 2.2 explains he mul is ep
combina ion o he algo i hms employed o de ine he op imal ou e o
he ne wo k’s pipelines and inally, he economic me ics used o e al-
ua e he di e en DH con igu a ions a e ou lined in Sec ion 2.3.
Two main so wa e we e used in he s udy. An open-sou ce GIS
so wa e, QGIS (QGIS De elopmen - eam, 2020), was employed o he
ac i i ies ela ed wi h geo e e enced a iables, while R S udio was
applied o da a p ocessing and cos calcula ions (Co e-Team, 2013).
2.1. Su ace lay-ou : bu e s and g ids
Once he hea supply sou ce ( he indus ial acili y) was selec ed, a
ci cula a ea o 1.5 km adius a ound he ac o y was se , o ensu e ha
he hea demand wi hin ha a ea was la ge han he was e hea om
he ac o y, enabling all was e hea a ailable o be exploi ed. This a ea
was di ided in o smalle sub-a eas in o de o calcula e he DH ne wo k.
The eason o his di ision is ha , al hough DH iabili y s udies a e
usually made a egional o municipal scale, he design o he pipeline
Nomencla u e
Ac onyms
EU Eu opean Union
DHW Domes ic Ho Wa e
DH Dis ic -Hea ing
RES Renewable Ene gy Sou ces
WH Was e Hea
IWH Indus ial Was e Hea
GIS Geog aphical In o ma ion Sys em
DSM Digi al Su ace Model
DTM Digi al Te ain Model
PB Payback
MST Minimum Spanning T ee
Pa ame e s
Q
n
Building demand [MWh/yea ]
dNPV Discoun ed ne p esen alue [EUR]
C
i
Cash- low
Discoun - a e [%]
DEM Nominal demand densi y [kWh/m
2
]
F Numbe o loo s in a building [-]
S Ho izon al su ace o a building[m
2
]
P Ra io be ween hea ed su ace and he o al su ace
LN Linea Hea Densi y [MWh/m]
Ø Pipeline Diame e [m]
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
3
ou ing o he ne wo k is de eloped a a lowe scale. Howe e , educing
he wo king scale o indi idual buildings would equi e an excessi e
compu a ional cos , so a middle g ound solu ion was aken de ining he
men ioned sub-a eas wi hin he a ea o s udy. In o de o p oduce hem,
wo di e en ypes o pa i ion we e es ed: bu e s and g ids (Fig. 1).
Rega ding bu e dis ibu ion, he ope a ion consis ed o es ima ing
he hea ing demand o ci cula bu e s, wi h he IWH sou ce a he
cen e o he bu e (g een iangle in Fig. 1a). Bu e s o a ious sizes
we e calcula ed, in o de o op imize he esul . The size (diame e ) o
hese bu e s di e ed in 100 m be ween hem. The use o hese bu e s
implied ha he esul ing DH ne wo k was e enly dis ibu ed a ound he
IWH sou ce, co e ing all he esiden ial buildings inside he ci cula
bu e . This app oach in ends o ind co ela ions be ween he economic
esul s and he geog aphical a iables in he sys em bu is no ocused on
inding he op imal con igu a ion o he DH ne wo k.
Fo g id dis ibu ion ypology (Fig. 1b), he main a ea o 1.5 km was
di ided in o a g id o med by squa ed cells. Each cell in he g id com-
p ises a building o g oup o buildings. The hea demand o e e y cell
was es ima ed by Eq. (2) (Sec ion 3.1), in o de o subsequen ly de ine
he ou e o he DH ne wo k. Since he esul depends on he employed
cell size, di e en alues we e es ed: om 50 ×50 m o 500 ×500 m,
wi h s eps o 50 m. Fo each g id size, a DH ne wo k con igu a ion was
p oposed. E e y cell was linked wi h he adjacen cells and he pa h o
he ne wo k only connec ed hose buildings showing he bes economic
esul s. A g id dis ibu ion o 150 ×150 m is shown in Fig. 1b.
Fo he g id dis ibu ion case, di e en adjacency le els we e
conside ed, in o de o ind he op imal pa h o he DH ne wo k.
S a ing om a pa icula cell, he pa h ha he DH ne wo k migh
ollow (i.e., he nex cell) depends no only on he cha ac e is ics on he
immedia e su ounding cells, bu also on cells ha a e a om i .
Howe e , his implies ha , in o de o pe o m he calcula ions o each
speci ic cell, all he emaining cells ha comp ise he o al su ace
should be assessed, which is no easible due o compu a ional easons.
The e o e, calcula ions wi h di e en adjacency le els we e pe o med.
This idea is illus a ed in Fig. 2 and explained nex .
The s a ing cell ep esen s he indus y loca ion ( ed cell in Fig. 2).
The i s adjacency le el includes he eigh cells ha di ec ly su ound i .
The second adjacency le el is o med by he cells ha su ound he i s
le el. The subsequen le els a e de ined in he same way. Now, in o de
o de e mine he nex cell ha he pa h o he DH will ollow, he i s
adjacency le el assesses eigh di e en possible op ions. The inco po-
a ion o a second adjacency le el analyzes 32 di e en op ions and so
on. No e ha hose cells ha ha e been co e ed once by he ne wo k a e
conside ed o be no useable o he nex s eps. Acco ding o pe o med
p elimina y a emp s, ou le els o adjacency we e included in he
p esen s udy; he inco po a ion o highe le els en ailed an excessi e
compu a ional cos .
As a esul o his sec ion, he buildings ha may be connec ed o he
new DH ne wo k a e es ablished. In he case o bu e dis ibu ion, his
means all he buildings wi hin hese bu e a eas, and, o he g id dis-
ibu ion, i means all he buildings included in he op imal pa h ha is
de ined by cells’ algo i hm. Di e en con igu a ions will esul in
di e en buildings. The nex sec ion explains he mul is ep me hodology
ollowed o he ou ing o he pipelines o he ne wo k.
2.2. DH ne wo k de ini ion using ou ing algo i hms
The goal o his sec ion is o selec he op imal DH ne wo k con ig-
u a ion. Fo his pu pose, he leng h o he ne wo k is es ima ed using a
mul i-s ep me hodology combining di e en ou ing algo i hms. Bo h
p ima y and seconda y sides o he DH ne wo k a e de ined o each
con igu a ion p oposed in he p e ious sec ion. Thus, he leng hs a e
compu ed based on g aph heo y me hods (Be ge, 2001). The objec i e
o his ou ing s ep is o de e mine he minimum dis ance connec ing all
he buildings selec ed om p e ious sec ion and using he exis ing oad
ne wo k. Unde his amewo k, he buildings unde s udy a e ans-
o med in o e ices co esponding o he cen oid o he su ace o each
building. The pipelines o he p ima y side o he DH a e usually con-
s ained by he oad ne wo k, as pipe di ching is hus acili a ed. The
used algo i hm comp ises he ollowing h ee s eps ha a e sequen ially
applied:
Fig. 1. Selec ed a ea o s udy di ided in (a) bu e s; (b) in a g id o 150 ×150m.
Fig. 2. Adjacency le els concep : g id showing 4 le els o adjacency.
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
4
1. G aph C ea ion (Delaunay iangula ion). This algo i hm enables an
op imal connec ion o all he buildings/poin s.
2. Rou ing (Johnson algo i hm) o he ne wo k connec ing he build-
ings. This algo i hm enables he iden i ica ion o oad segmen s o
join all he connec ions om he iangula ion p ocess.
3. Calcula ion o he Minimum Spanning T ee (MST), enabling he
calcula ion o he minimum DH ne wo k leng hs.
4. Calcula ion o Economically Op imal DH Con igu a ion
The combina ion o hese ou s eps is illus a ed in Fig. 3. Each s ep
is de ailed in he ollowing sec ions.
2.2.1. G aph c ea ion
The buildings ma ked as op imal o DH deploymen in Sec ion 2.1
will se e as inpu o his poin . These buildings a e con e ed in o
poin s o e ices, using he cen oid o he shape o each building. In
his i s s ep, he objec i e is o c ea e a g aph in which all he buildings
a e in e -co ela ed. Fo his pu pose, he Delaunay iangula ion
(Delaunay B., 1934) is used o connec he buildings in each o he cases,
o ming a se o e ices and edges ha make up a se o iangles.
Delaunay iangula ions maximize he minimum angle o all he angles
o he iangles in he iangula ion, a oiding iangles wi h one o wo
ex emely acu e angles ha a e no desi able du ing some in e pola ions
and ma hema ical p ocesses. The edges o hese iangles join each pai
o poin s ha co espond o he cen oids o he buildings, which mus
be analyzed in he ollowing s eps. The se o e ices and edges in he
sys em o ms a plana g aph, which ensu es ha buildings ha a e a
away om each o he a e excluded. Thus, each iangle connec s h ee
buildings, o ming h ee pai s o buildings. The ou pu om his ian-
gula ion p ocess is mul iple pai s o buildings connec ed by he Delau-
nay iangula ion ha will be used as inpu o he ollowing p ocess.
Fig. 3. Rou ing Algo i hm p ocess scheme: G aph C ea ion (a), ou ing be ween buildings (b) and calcula ion o he minimum spanning ee (c).
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
5
2.2.2. Calcula ion o ou ing be ween buildings
The objec i e o his sec ion is o de ine he eal dis ance by he oad
ne wo k o all he pai s o buildings connec ed by he Delaunay ian-
gula ion o Sec ion 2.2.1. The weigh o he edges om he g aph c ea ed
using he Delaunay iangula ion co esponds o he Euclidean dis ance
be ween he nodes o buildings. Howe e , his dis ance does no
co espond o he ac ual oad ne wo k. Acco dingly, Johnson’s algo-
i hm (Johnson, 1977) is used o ind he sho es pa h be ween all pai s
o buildings in he a ec ed a ea, aking in o accoun he oad ne wo k.
In eal ci cums ances and aking in o accoun he ac ha he dis i-
bu ion pipelines (p ima y side) o he DH ne wo k will couple wi h he
cu en oad ne wo k, he dis ance o ge om one poin o ano he
could be much longe han he Euclidean dis ance. The eal ou ing
dis ance be ween wo buildings will be he sum o he oad segmen s
be ween he wo poin s and he dis ance be ween he building cen oid
o he closes oad. Thus, a new weigh is ob ained o each pai o
buildings in he plana g aph, wi h he Euclidean dis ance be ween
hem. This will be he inpu o he ollowing s ep.
2.2.3. Calcula ion o minimum spanning ee
The objec i e o his inal s ep is o de ine he con igu a ion o he DH
ne wo k by he calcula ion o he Minimum Spanning T ee (MST) om
he se o e ices and edges esul ing om Sec ion 2.2.2. Fo his pu -
pose, K uskal’s algo i hm (K uskal, 1956) is applied, which enables he
connec ion o all he buildings ( e ices) wi h he sho es pa h weigh ed
by he eal dis ance be ween he di e en buildings connec ed by eal
oads. The calcula ion o he MST ensu es ha he DH con igu a ion
ob ained om his mul is ep algo i hm is he sho es ne wo k and,
consequen ly, he op imal one in economic e ms.
The p ima y ne wo k connec s he hea sou ce wi h he subs a ions
o he building, whe eas he seconda y ne wo k in a DH connec s he
subs a ions wi h he acili ies o he building. The leng h o he p ima y
side o he ne wo k (L
PRIM
) is made up o he sum o he leng hs o he
di e en a cs in he MST. Fu he mo e, he sum o he dis ance o he
ansla ion o he cen oids co esponds o he leng h o he seconda y
side (L
SEC
) o he ne wo k. This connec ion can be made using an algo-
i hm in (QGIS De elopmen - eam, 2020) in oduced wi h he plugin
om (Con ad e al., 2015).
2.2.4. Calcula ion o Economically Op imal DH con igu a ion
This inal sec ion aims o iden i y he op imal DH con igu a ion
among he di e en scena ios p oposed in p e ious sec ion. Fo his
pu pose, a simpli ied economic assessmen is p oposed in which only he
la ges expenses and cash lows a e included. Acco ding o (ETI, 2017),
he la ges cos s a e eached o ci il wo ks, wi h 36% o he o al cos s
(including he planning and de elopmen o he p ojec ). Mo eo e , he
pipelines ha need o be in oduced in he ench a e he ci il wo ks
also ake up an impo an pa o he o al cos s; bo h adding up o o e
50%. Since his analysis does no include ope a ion cos s, he esul s
should be conside ed as p elimina y. Howe e , his app oach is accu a e
enough o an ini ial easibili y s udy o any new DH ne wo k. Fo he
compa ison o all he cases, a simpli ied payback (PB) pe iod is p o-
posed. Acco ding o (Ha is, 2018), he PB pe iod o his kind o sys em
is calcula ed as shown in Eq. (1):
Payback =Ini ial In es men
Ene gy sa ings Eq. (1)
whe e he ene gy sa ings ep esen he cos s sa ed by no using he hea
supply sys em ha is cu en ly in use in hose buildings. As he e is no
exac in o ma ion o wha ype o hea supply echnology exis s in each
building, i is assumed ha all he buildings used a na u al gas
condensing boile , wi h a nominal he mal e iciency o 90%. This
app oxima ion alue o he e iciency is a s anda d alue o boile s
wo king a hei nominal alue.
2.3. Economic sensi i i y analysis o selec ed DH ne wo k
In his inal s ep and wi h he objec i e o analyzing he economic
sensi i i y o he selec ed ne wo k, an ex ended economic assessmen is
ca ied ou o he op imal ne wo k. This economic assessmen is based
on he gene al economic me ic, discoun ed ne p esen alue (dNPV)
which is calcula ed by he ollowing equa ion (Eq. (2)).
dNPV =∑
T
=1
Ci
(1+ )iEq. (2)
whe e F
i
is he ne cash low du ing pe iod , and e e s o he discoun -
a e.
The sensi i i y analysis is based on he a ia ion o some o he mos
luc ua ing ope a ional a iables and s udying he e ec s o hese
changes in he dNPV. Posi i e dNPV indica es ha he discoun ed eco-
nomic e u ns a e g ea e ha he in es men equi ed. So, o his
economic sensi i i y analysis, discoun - a e and na u al-gas p ice a e
pa ame ized o he i s 20 yea s. The alues used o he case s udy a e
p esen ed in Sec ion 3.3 Thus, he selec ed con igu a ion o he
deploymen o he ne wo k is economically analyzed in o de o a oid
in e p e a ion e o s and o co e a wide ange o scena ios.
3. Desc ip ion o he case s udy
The me hodology p oposed along he a icle was applied o es ima e
he basic design o a DH ne wo k in he ci y o Vi o ia-Gas eiz (249,176
inhabi an s in 2018), loca ed in he adminis a i e egion o he Basque
Coun y (no he n Spain). Acco ding o K¨
oppen-Geige classi ica ion,
his loca ion co esponds wi h a C
b
clima e, e e ing o oceanic clima e.
The case s udy was applied o an a ea u banized o e a ious decades,
wi h a p edominance o buildings cons uc ed in he 1970s. The s udy
a ea comp ises a 1.5 km adius ci cula bu e a ound he indus y unde
s udy (speci ied in Sec ion 2.1). This a ea is chosen o ensu e ha he
hea demand is la ge han he was e hea om he ac o y. As de ailed
in Sec ion 2.1, his a ea was di ided in o wo ypes o smalle sub-a eas,
bu e s and g ids.
3.1. Indus ial was e-hea sou ces
The i s s ep in he p esen ed case-s udy was he cha ac e iza ion o
he was e hea sou ces a ailable in he ac o ies and he iden i ica ion o
a sui able indus ial emplacemen . The cha ac e iza ion o he IWH
a ailable in he amewo k o his case s udy was pe o med by using he
esul s om a p e ious wo k done by he au ho s in he Basque Coun y
(Spain) (La inaga e al., 2021). This wo k p esen s he cha ac e iza ion
o he mos impo an ac o ies in he egion by he applica ion o a
bo om-up me hodology using commonly a ailable measu es, such as
hea sou ced bu ned in he ac o y.
A ac o y is app op ia e o he exploi a ion o i s was e hea by a DH
ne wo k i some condi ions a e me . Fi s , high was e hea ene gy has o
be a ailable and high empe a u es p o ide a highe -g ade ene gy. The
ac o y may be loca ed nea an u ban co e wi h high densi y demand.
The low dis ance be ween he ac o y and he buildings connec ed o he
DH ne wo k educe he hea losses in he dis ibu ion pipelines and
op imizes he use o he IWH. Finally, no la ge physical ba ie s ha e o
be ound be ween he ac o y and he buildings. These ba ie s would
d as ically inc ease he ini ial in es men equi ed o he ne wo k.
Examples o physical ba ie s could be moun ains o i e s.
The chosen indus y was a ound y (Fundici´
on Olazabal y Hua e,
2020) ha has been in ope a ion o mo e han i y yea s. I is loca ed
e y nea o he ci y cen e , which shows a la ge ene gy densi y demand,
and cu en ly employs 54 wo ke s. The es ima ed echnical was e hea
a ailable om he di e en manu ac u ing p ocesses is a ound 158
GWh/yea (La inaga e al., 2021). The empe a u e dis ibu ion o he
hea a ailable in he ound y is shown in Table 1. F om all he ac o ies
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
6
analyzed, he ac o y chosen o his case s udy p esen ed mee s all he
condi ions men ioned abo e.
A ound 60% o he was e hea a ailable in he ound y is a high o
e y high empe a u e ange (>500 ◦C); whe eas he es o he ene gy is
a ailable a low-medium empe a u e. E en hough high empe a u e
ange IWH is ypically used o elec ici y p oduc ion due o i s highe
quali y, all he hea a ailable in he ac o y is adequa e o injec ion in o
DH ne wo ks. Since his is a p elimina y s udy, i was op ed o include
also he high empe a u e IWH in o he a ailable esidual hea o he
DH ne wo k.
3.2. Buildings ene gy demand es ima ion
The hea demand (Q
n
) in each building was calcula ed ough Eq. (3):
Qn[kWh] = DEMn⋅Fn⋅Sn⋅PEq. (3)
The e, DEM
n
is he nominal hea demand in each building [kWh/
m
2
]; F
n
, he numbe o loo s [-]; S
n
, he o al su ace [m
2
] and P [-] is he
a io be ween he hea ed su ace and he o al su ace o he building.
De ailed in o ma ion o calcula e he pa ame e s in Eq. (3) is p o ided
nex .
In o de o es ima e he numbe o loo s (Fn), he heigh o each
building had o be i s calcula ed. Typically, GIS in o ma ion o build-
ings is a ailable in shape iles p o ided by he local au ho i ies. They
include loca ion, o al su ace and polygonal shape o buildings; how-
e e , hey do no usually p o ide in o ma ion on he use ul a ea o ne
heigh o buildings, as equi ed in Eq. (3). To ob ain hem, LiDAR (Lase
Imaging De ec ion and Ranging) da a we e he ein employed (Sha ma
e al., 2021). LiDAR is an op ical emo e sensing echnique ha uses
lase ligh o ob ain a dense sample o disc e e alues o he heigh o
ea h’s su ace.
Amongs he di e en LiDAR da a ypically a ailable, he Digi al
Su ace Model (DSM) and he Digi al Te ain Model (DTM) we e he ein
applied. DSM p esen s eal ele a ions on he su ace, including man-
made ea u es, such as buildings o in as uc u e, as well as na u al
ea u es, such as ees. DTM is a model de i ed om DSM, whe e all he
abo e-men ioned su ace ea u es ha e been emo ed, lea ing only he
ba e e ain. S a ing om hese wo da ase s, he heigh o he in ol ed
buildings can be es ima ed wi h a as e calcula o by means o Eq. (4):
Ne Heigh s =DSM −DTM Eq. (4)
Since LiDAR da a a e usually a ailable as *.las o *.laz iles, hey
need o be con e ed in o VRT iles ( as e iles), which enable he as e
ope a ions be ween he DSM and DMT as e iles. To do so, i is equi ed
ha he dis ance be ween poin s (o cells in as e ) is he same in bo h
DSM and DTM da a. This ende s a new as e laye ha includes he ne
heigh o all he ea u es abo e he e ain. In he p esen case, he ob-
ained g id size in bo h as e laye s was 5 m. Then, a plug-in om SAGA
o QGIS (Con ad e al., 2015) was used o assign loca ion a ibu es
om he shape ile o he buildings o he heigh s ile o he ea u es. An
a e age heigh was ob ained o each building.
Fig. 4 shows he main di e ence be ween hese wo da ase s o he
in ol ed loca ion ( he speci ic de ails o he buildings included in he
p esen case-s udy a e p o ided in Table 1). GIS in o ma ion o he
buildings o he case-s udy was ob ained om (Gobie no Vasco, 2020),
while LiDAR da a we e ob ained om CNIG (CNIG, 2020).
The numbe o loo s o each building (F
n
in Eq. (3)) we e calcula ed
di iding he ne heigh ob ained om LiDAR da a by he dis ance be-
ween loo s. The dis ance be ween loo s was he ein ixed o 2.7 m,
sligh ly highe han he minimum dis ance be ween slabs ixed o es-
iden ial buildings (2.5 m) (CTE, 2019). The esul om he di ision was
hen ounded o i s lowes alue.
Finally, he ho izon al p ojec ion o each building (S
n
in Eq. (1)) was
ob ained om he in o ma ion ga he ed in he GIS iles. The ein,
buildings a e a ailable as polygonal elemen s; hus, S
n
was ob ained by
calcula ing he a ea o hose polygons. This ho izon al p ojec ion was
mul iplied by a educ ion ac o (P)o 0.83, in o de o ob ain he use ul
hea ing a ea in each building (D’Alonzo e al., 2020). This way, ex e nal
o in e nal walls and non-hea ed spaces such as ele a o s a e conside ed.
The building laye om (Gobie no Vasco, 2020) was clipped by he
1.5 km bu e , iden i ying 4555 buildings in he egion. The buildings in
he loca ion included a ious ypes o cons uc ions, which a e de ailed
in Table 2.
Table 1
IWH dis ibu ion by empe a u e ange in he selec ed ound y, e ie ed om (La inaga e al., 2021).
Tempe a u e[◦C] 200–300 ◦C 300–400 ◦C 500–1000 ◦C >1000 ◦C To al
IWH [GWh] 48.77 14.10 73.14 21.57 157.55
IWH [%] 30.95 8.94 46.42 13.69 100
Fig. 4. (a) DSM and (b) DTM o Vi o ia-Gas eiz (Spain).
Table 2
Dis ibu ion o he buildings in he a ea o s udy di ided by he ype o buildings.
Type o Building Numbe o Buildings Pe cen age [%]
Gene ic Cons uc ion 3733 81.95
Ligh weigh Cons uc ion 150 3.29
G eenhouse 38 0.83
Wa ehouse 484 10.63
Buildings in uin 2 0.04
Sac ed Building 71 1.56
Canopy 68 1.49
Singula Building 9 0.20
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
7
The o al hea demand o he buildings wi hin he a ea was calcu-
la ed using LiDAR da a and based on building heigh s, ollowing he
me hodology explained in Sec ion 3.1. The p oposed me hodology is
only alid o esiden ial buildings. The es o he ypes o building
shown in Table 2 migh ha e di e en hea ing demand p o iles.
Rega ding he nominal hea demand (DEM
n
) o he buildings unde
s udy, wo al e na i es we e employed o es ima e i :
1. DEM1: Nominal densi y alues p o ided by he Spanish Ins i u e o
he Di e si ica ion and Sa ing o Ene gy (IDAE) (IDAE, 2011). The
da a a e based on a ious simula ions o a e age building ypes in
di e en clima es; hus, di e en demand densi ies o space hea ing
and DHW a e achie ed. Acco ding o (IDAE, 2011), he nominal
ene gy demand o space-hea ing is 163.6 kWh/m
2
and 13.5
kWh/m
2
o DHW.
2. DEM2: In o de o include hea consump ion o newe buildings, a
lowe nominal hea ing demand densi y was also employed, acco d-
ing o (Te ´
es-Zubiaga e al., 2015). The used alue was cons an : 60
kWh/m
2
(40 kWh/m
2
o space hea ing and he es o DHW).
Thus, he ob ained o al hea demand o he esiden ial buildings in
he a ea eaches 2528.82 GWh/yea wi h DEM1 hea densi y and
1273.97 GWh/yea wi h DEM2 densi y. These esul s ep esen a leas
10 imes he o al IWH a ailable in he conside ed ac o y. This excess
demand ea i ms he idea ha he buildings ha migh be connec ed o
he DH ne wo k ha e o be adequa ely chosen in o de o ob ain op imal
economic assessmen .
One o he a iables ha bes desc ibes he ene gy needs o a speci ic
a ea is he ene gy densi y in kWh/m
2
. The hea densi y in each o he
bu e s is calcula ed as he sum o all he demand in he bu e di ided by
he su ace co e ed by he bu e . Wi hin buildings wi h simila ene gy
consump ion, he highe buildings will show highe ene gy densi y. In
ac , Fig. 5 shows he buildings laye classi ied by he ene gy densi y
(Fig. 5a) and a 3D iew (Fig. 5b) o he buildings in he a ea using a
plugin by (WebGL echnology and h ee.js Ja aSc ip lib a y, 2020),
also classi ied by he hea consump ion densi y.
Thus, o all he buildings in he loca ion, only hose o gene ic and
ligh weigh cons uc ion (Table 2) a e included in he analysis. O he
building ypologies would equi e an addi ional demand cha ac e iza-
ion due o hei singula hea consump ion p o iles. Rega ding hese
wo ypes o cons uc ions, Fig. 6 shows he dis ibu ion o he esi-
den ial buildings acco ding o he ene gy demand densi y.
3.3. Economic assessmen & sensi i i y analysis
On he one hand, he ini ial in es men comp ises he cos o
enching, pipelines and he connec ion o he ins alla ions o he
exis ing buildings (subs a ions). The equa ions Eq. (5) and Eq. (6) o
his economic app oach a e aken om (Pe sson and We ne , 2011):
In es men Pipelines =130 +2858⋅ ØEq. (5)
Ø=0.0486⋅LN +0.063 Eq. (6)
whe e Ø is he diame e o he pipeline in [m] and LN ep esen s he
linea hea densi y in [MWh/m]. In hese equa ions, he in es men
cos s a e gi en by pipe leng h uni in [EUR/m]. Mo eo e , he cos o
he DH subs a ions, including he equipmen , ins alla ion and connec-
ions is ixed a 2500
€
/subs a ion (Gudmundsson e al., 2013).
On he o he hand, he ope a ional a iables o he case s udy a e se
o he sensi i i y analysis o he selec ed ne wo k. Fi s , he na u al gas
p ice is expec ed o inc ease in he ollowing yea s (Gao e al., 2021) and
consequen ly alues in Table 3 a e p oposed as o household con-
sume s’ p ice. E en hough he cu en na u al gas p ice is a ec ed by
se e al ac o s, a mean alue o 0.05
€
/kWh is used as cu en e e ence.
The o he main pa ame ic a iable is he discoun a e. This pa ame e
is changed om 0% up o 10% using he alues shown in Table 3.
Thus, he selec ion o he op imal DH ne wo k con igu a ion is ca -
ied ou calcula ing he simpli ied PB pe iod and he sensi i i y analysis
is calcula ed only o he selec ed ne wo k con igu a ion.
4. Resul s
4.1. Bu e app oach
This sec ion will p esen he esul s om he applica ion o he
me hodology shown in Sec ion 2 o he case s udy p esen ed in Sec ion
0. The applica ion o ou ing algo i hms explained in Sec ion 2.1 allows
he eal pa h o he DH ne wo k con igu a ions in each o he a ian s o
be de ined. The s udy a ea o he case-s udy has been dis ibu ed and
di ided in wo di e en ways, ep esen ing wo di e en concep s o
ne wo k con igu a ion: ci cula bu e s and g ids.
The dis ibu ion o he s udy a ea in o ci cula bu e s aims o
de elop mul idi ec ional DH ne wo ks om he IWH sou ce as he ini ial
poin . Thus, all he esiden ial buildings wi hin he ci cula bu e a e
co e ed by he hea ing ne wo k, ega dless o he o e all hea demand
o hea demand densi y o he buildings. As an example o he con-
s uc ion o he DH ne wo k using he ou ing algo i hm p esen ed in
Sec ion 2.1, Fig. 7 p esen s wo ne wo k con igu a ions o wo di e en
bu e sizes: 300 and 1000 m.
The economic esul s ob ained o he bu e ’s me hod a e p esen ed
in Table 4. The payback pe iod o di e en bu e sizes and he wo
conside ed hea demand densi ies a e ou lined. Since he adial
Fig. 5. A ea unde s udy. (a) Building plana laye ; (b) 3D iew.
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
8
cha ac e o he ne wo k makes i s leng h inc ease exponen ially wi h
he size/ adius o he bu e , he cos s o he cons uc ion o a new DH
ne wo k a e comple ely in luenced by he size o he bu e .
A simpli ied payback pe iod below 10 yea s is achie ed wi hin a
bu e adius equal o o less han 600 m o he wo hea ing demand
densi y alues, DEM1 and DEM2, as i can be obse ed in Table 4. Wi h
he DEM1 hea ing densi y, all he IWH a ailable is consumed in he
400m adius bu e , whe eas, wi h he DEM2 densi y, he was e hea
om he indus y manages o co e all he buildings wi hin he 700 m
bu e . No mo e demand can be co e ed in la ge egions and, conse-
quen ly, he economic me ics exponen ially wo sen om hose bu e
size alues. F om he bu e sizes indica ed abo e, he payback a iable
lea es i s linea i y and begins o inc ease exponen ially, mainly due o
he la ge pipeline sys em necessa y o co e all he buildings. In bo h
cases, he payback shows a minimum in he 200 m bu e .
F om he analysis o he esul s o he bu e dis ibu ion, i can be
concluded ha he buildings ha migh be co e ed by he ne wo k ha e
o be ca e ully and adequa ely chosen. Al hough he analysis made in
he s udy based on bu e a eas is in e es ing om a heo e ical poin o
iew, he eal ne wo k would be cons uc ed by he esul s ob ained
om he g id analysis. This algo i hm allows di ec ional DH ne wo k
con igu a ions o be buil , wi h he economic op imiza ion included.
4.2. G id algo i hm & de ini ion o DH ne wo k con igu a ion
The g id dis ibu ion o he loca ion aims o de elop a DH ne wo k
con igu a ion wi h a unique main di ec ionali y, inding egions and
buildings wi h op imal economic e u n. Unlike he bu e dis ibu ion,
his algo i hm inds he op imal di ec ion o he ne wo k in unc ion o
he ex e nal condi ions (hea demand, oad leng h, e c.), so ha only he
Fig. 6. Numbe o esiden ial buildings classi ied by he ene gy demand densi y o DEM1 and DEM2.
Table 3
Economic a iable Sensi i i y analysis.
Na u al Gas P ice [
€
/kWh] Discoun Ra e
0.03 0%
0.05 3%
0.10 5%
0.15 10%
Fig. 7. Con igu a ion o he ne wo k. (a) Bu e o 300 m adius and (b) Bu e o 1000 m adius.
Table 4
Payback pe iod analysis o di e en bu e - adius [m] leng hs.
Bu e [m] 100 200 300 400 500 600 700 800
DEM1 4.5 3.5 4.3 4.5 5.4 8.5 13.1 17.9
DEM2 5.2 3.9 5.6 6.0 6.8 8.5 10.1 14.1
Bu e [m] 900 1000 1100 1200 1300 1400 1500
DEM1 24.5 30.5 40.6 50.9 59.9 70.7 80.1
DEM2 19.9 25.1 34.3 43.4 51.6 61.6 70.2
M. Lumb e as e al.
Jou nal o Cleane P oduc ion 349 (2022) 131491
9
mos sui able buildings a e connec ed o he hea ing ne wo k. The al-
go i hm is de eloped in o de o maximize he economic iabili y o he
ne wo k, ollowing Eq. (3), Eq. (4) and Eq. (5). As an example o hese
ne wo ks’ concep , Fig. 8 shows he layou o wo DH con igu a ions o
wo cell sizes.
O e all, o y DH ne wo k con igu a ions a e modelled, including 10
g id sizes and 4 adjacency le els o each o he g id sizes. The dis i-
bu ion o he buildings a ies among he di e en dis ibu ion sizes, and,
in consequence, he ne wo k con igu a ion will be di e en in each case.
E en hough each o he DH con igu a ions is de eloped o be
economically op imal, he change o bounda y condi ions by changing
cell size will a ec he lay-ou o he ne wo k. Thus, simula ing 40
di e en cases wi h di e en bounda y condi ions allow he mos
app op ia e a eas o he deploymen o he ne wo k o be iden i ied in
o de o ma ch hem up wi h he mos epea ed a eas wi hin all he
simula ed cases. Fig. 9 shows he hea map o all he 40 cases o e lapped
in he same image, iden i ying he mos epea ed buildings/ egions wi h
ed ma ks.
Fig. 9 shows ha mos o he DH con igu a ions a e conduc ed o he
sou h and sou hwes o he indus ial si e. As i can be obse ed, some o
he buildings u n ou o be connec ed o he DH ne wo k in all he 40
con igu a ions, whe eas no he n egions a e no conside ed in any o
he 40 cases.
The ini ial dis ibu ion o he a ea o s udy in o di e en sized g ids
de ines he ini ial condi ions o he simula ions. Va ia ion o he cell
size a ec s he bounda y condi ions wi h di e en building dis ibu ions
along he cells o he g id. Thus, he esul ing pa h o he ne wo k may
sligh ly a y om case o case (See Fig. 9). The oads used will a y om
case o case and, as a esul , he leng hs and cha ac e is ics o he
pipelines will also a y. As concluded om bu e analysis, he pipelines
ollowing he oad ne wo k and hei leng hs and diame e s de e mine
he economic easibili y o he sys em. Thus, Fig. 10 p esen s he co -
ela ion be ween he leng hs o he p ima y and seconda y sides and he
linea mean ene gy densi y o he ne wo k in each con igu a ion. The
p ima y side o he ne wo k connec s he hea sou ce, he ac o y in his
case, and he subs a ions, whe eas seconda y side o he ne wo k con-
nec s he subs a ions wi h he inal use s/dwellings.
F om he s udy abo e was concluded ha mean linea ene gy densi y
is a c i ical pa ame e o economic easibili y o he sys em and his
pa ame e esul s o be exponen ially inc ease wi h long ne wo ks. O
he 40 con igu a ions, he la ges ne wo k would need a ound 27 km o
pipelines, including he p ima y and seconda y sides, whe eas he
sho es ne wo k would only esul in 8 km. Two main cu es a e
obse ed in Fig. 10 and some poin s p esen smalle ne wo ks o he
same ene gy densi y. This is caused by he be e oad con igu a ion o
he cases wi h smalle leng hs. Thus, hea ne wo k wi h la ge linea
densi y and sho ne wo ks a e he mos in e es ing om an economic
iew.
The ini ial in es men om he ins alla ion o he dis ibu ion
pipelines comple ely limi s he iabili y o he p oposed sys em. The
o he a iable a ec ing he payback pe iod is he economic sa ings om
no equi ing he ene gy co e ed by he IWH. Fig. 11 shows he e olu-
ion o he wo cash lows o Eq. (4), clus e ed by he g id size o he
ini ial dis ibu ion.
A clea di e gence be ween he wo linea endencies o he cash
lows conside ed in he s udy is obse ed. As long as he yea ly demand
co e ed by he ne wo k inc eases, he ini ial in es men equi ed o he
ins alla ion o he ne wo k also inc eases, wi h a g ea e slope han
economic sa ing unc ion. The ini ial in es men in he esul ing con-
igu a ions anges om a ound 20 M
€
o a ound 72 M
€
; whe eas he
economic sa ings om he p ima y ene gy sa ings ange om 2.5 M
€
o
a ound 10 M
€
.
Besides, Fig. 12 shows a hea -map o he payback pe iod, showing
he 40 con igu a ions unde analysis o he p oposed DH, dis ibu ed by
Fig. 8. Con igu a ion o he ne wo k o wo g id sizes: (a) 250m and (b) 400m. Le el 2 o adjacency in bo h cases.
Fig. 9. Hea Map o he mos co e ed a eas by he di e en DH ne wo k
con igu a ions.
M. Lumb e as e al.
