1876-6102 © 2015 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL
doi: 10.1016/j.egyp o.2015.11.662
Ene gy P ocedia 78 ( 2015 ) 2899 – 2904
ScienceDi ec
6 h In e na ional Building Physics Con e ence, IBPC 2015
An app oach o u ban qua e design using building gene a i e
design and he mal pe o mance op imiza ion
Eugénio Rod iguesa,b,*, Ana Ri a Ama ala,
Adélio Rod igues Gaspa a, Ál a o Gomesb,c
aADAI, LAETA, Depa men o Mechanical Enginee ing, Uni e si y o Coimb a, Rua Luís Reis San os, Pólo II, 3030-788 Coimb a, Po ugal
bINESC Coimb a – Ins i u e o Sys ems and Compu e Enginee ing a Coimb a, Rua An e o de Quen al 199, 3000-033 Coimb a, Po ugal
cDepa men o Elec ical and Compu e Enginee ing, Uni e si y o Coimb a, Rua Luís Reis San os, Pólo II, 3030-290 Coimb a, Po ugal
Abs ac
Buildings he mal pe o mance is in luenced by he u ban con ex , such as adjacen buildings shadows, block wind pa hs, o sola
adia ion e lec ion. Fo his eason, in his pape , an au oma ed p ocedu e is used o gene a e and op imize buildings’ he mal
pe o mance in a closed O-shape u ban qua e wi h en building blocks o de e mine he impo ance o including he su oundings
in es ima ing he buildings’ he mal beha io . The o e all shape o he u ban qua e is p e-designed, being mainly esiden ial,
wi h se e al s o es in he g ound loo and limi ed o h ee le els. Each building will ha e ou al e na i e designs c ea ed using
a hyb id e olu iona y s a egy echnique ha gene a es building’s loo plans acco ding o p ac i ione ’s p e e ences and
equi emen s. Then, a sequen ial a iable op imiza ion p ocedu e coupled wi h dynamic simula ion engine is used o explo e
he imp o emen po en ial o each solu ion by changing and adding se e al building elemen s. The inal qua e design is
de e mined by combining he bes o all buildings’ solu ions om he mal pe o mance c i e ia.
The esul s demons a e he in luence o u ban con ex in he esul ing building’s he mal pe o mance. Despi e he building’s
shape is simila in he ou solu ions, hese ha e signi ican he mal beha io di e ence due o hei in e io o ganiza ion and
posi ion in he u ban qua e . A compa ison analysis is ca ied ou be ween all building block designs. The buildings, which ha e
ex e io walls wi h openings acing sou h, ha e almos hal deg ee-hou s o he mal discom o due o he combina ion o la ge
openings and shading o e hangs. I is also possible o conclude ha gene a i e ools, enhanced wi h op imiza ion p ocedu es,
may help p ac i ione s in designing mo e ene gy e icien buildings.
© 2015 The Au ho s. Published by Else ie L d.
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL.
* Co esponding au ho . Tel.: +351 962719008.
E-mail add ess: eugenio. od [email protected] (Eugénio Rod igues)
A ailable online a www.sciencedi ec .com
© 2015 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL
2900 Eugénio Rod igues e al. / Ene gy P ocedia 78 ( 2015 ) 2899 – 2904
Keywo ds: u ban gene a i e design; building he mal pe o mance; dynamic simula ion; geome y op imiza ion
1.
In oduc ion
The buil en i onmen is esponsible o 31% o he global ene gy consump ion, wi h space hea ing and cooling
ep esen ing 30% (in cold clima es up o 60%) [1]. The building pe o mance depends signi ican ly om i s
su oundings, hus i is c ucial o assess and imp o e he u ban o m. The bes way o de e mine he building
pe o mance aking in o accoun all he elemen s b ough by he u ban con ex ha may in luence is o use dynamic
simula ion. Mo eo e , se e al app oaches ha e p esen ed his dynamic simula ion coupled o pe o mance
op imiza ion echniques as a pa h o op imize he solu ions achie ed, in an ea ly s age o he design p ocess. Each
app oach in ends o ocus on di e en c i e ia.
As Ve meulen e al. [2] show, he combina ion o simula ion ools wi h nume ical op imiza ion me hods appea s
o be an e icien way o ob ain he bes op ion o each case, acco ding o de ined objec i es and equi emen s. This
esea ch eam coupled a hyb id e olu iona y algo i hm o an u ban ene gy simula o (Ci ySim) ocused on adia ion
and buildings ene gy lows, pu suing an e alua ion o annual ene gy needs, de ined as he objec i e unc ion o be
minimized in he op imiza ion p ocess. Koenig [3] used simula ion echniques based on cellula au oma a o
gene a e
complex u ban pa e ns, combining he six which he conside ed he mos common. In his way, al e na i e
solu ions
we e c ea ed o he same design p oblem in a sho ime and in an easy way, as he expe imen s we e
c ea ed in
CAD en i onmen . Kämp and Robinson [4] used he sola i adia ion c i e ion o apply an e olu iona y
algo i hm as
a building op imiza ion o he p og am RADIANCE, in o de o ob ain he bes building and u ban
o m o p edic
u ban sola po en ial o use, namely h ough sola he mal collec o s o pho o ol aic panels.
Summing up, and acco ding o Sanaienan e al. [5], an ex ensi e e iew on ecen esea ch can lead he p edic ion
o he e ec o u ban o m on buildings pe o mance o he agg ega ion in o h ee ca ego ies: he he mal beha io ,
sola access o passi e hea ing and dayligh pu poses, and indoo and ou doo na u al en ila ion. Howe e , his
s udy showed how di icul i is o analyze he impac o he su oundings on buildings he mal pe o mance, leading
esea che s o ocus on he pa ame e wi h mo e me hods a ailable o es ima e i s po en ial: sola adia ion.
In his sense, p e ious s udies show ha building and u ban o m usually esul om shaping he building shell
wi hou any conce n o accommoda e he building design p og am, as he in e io con igu a ion and ooms
dis ibu ion
a ec s he o m and he a ailable ex e io wall o place he openings.
This pape p esen s an app oach o u ban qua e design by using building gene a i e design and pe o mance
op imiza ion echniques coupled o dynamic simula ion. In he p esen ed app oach, he building’s design p og am is
he s a ing poin o gene a e, op imize, and in o m he use on he he mal pe o mance o each building block in
he u ban qua e . This documen is made up o ou sec ions. In he i s pa is p esen ed he p oblem and i s
backg ound. In he ollowing sec ion, he me hodology is desc ibed and he me hods p esen ed. In he hi d sec ion,
a showcase is ca ied ou and analyzed. Finally, in he las sec ion, conclusions a e made.
2.
Me hodology
2.1.
Gene a i e design echnique
The gene a i e echnique used in his u ban design exe cise is an e olu iona y s a egy enhanced wi h a
s ochas ic
hill climbing me hod [6–8]. This hyb id popula ion-based app oach gene a es al e na i e loo plans by
alloca ing
spaces and e ical ci cula ions, acco ding o he use ’s p e e ences and equi emen s, de ined in he
design
p og am. The echnique aims o minimize a g oup o objec i es, such as connec i i y, adjacency, o ien a ion,
leng hs
and minimum a eas, maximum g oss and cons uc ion a eas, o e low o spaces, e c. The app oach uses
adap i e
ans o ma ion ope a o s o he indi iduals acco ding o andom geome ic ac ions, such as ansla ion,
o a ion,
e lec ion, e c. These ope a o s may a ec single elemen , such as an opening o a space, o a clus e o
spaces,
le els, o he whole loo plan.
The u ban qua e is de e mined by de ining sui able dep h and wid h o each building block acco ding o he
desi ed building design p og am. Al e na i e building plans a e gene a ed and he dimensions a e calcula ed o each
Eugénio Rod igues e al. / Ene gy P ocedia 78 ( 2015 ) 2899 – 2904 2901
block o he O-shaped qua e , made up o 10 buildings. This qua e is su ounded by o he buildings wi h simila
heigh and dis anced by 12 me e s ha co espond o he s ee . Fo each building block o he u ban qua e , 18
al e na i e designs a e gene a ed and he ou i es designs a e used o he mal e alua ion and op imiza ion, hus
o alizing 40 c ea ed buildings. A e he he mal pe o mance op imiza ion is ca ied ou o all o hese designs, he
bes pe o mance solu ion in each building block is chosen. Du ing he gene a ion p ocess and op imiza ion
p ocedu e
o each building block, he su ounding buildings in he u ban qua e a e dummy buildings wi h simila
heigh . In he
end o he o e all p ocess, he en bes buildings a e me ged in a single u ban design solu ion.
2.2.
The mal pe o mance op imiza ion
A e al e na i e buildings a e gene a ed, he he mal pe o mance is ca ied ou using a sequen ial a iable
op imiza ion p ocedu e [9,10]. The building pe o mance is assessed using dynamic simula ion (Ene gyPlus 8.1.0),
which is coupled o he algo i hm. All simula ion speci ica ions, wea he da a, cons uc i e sys em, occupancy and
ligh ing/equipmen usage schedules, which o his s udy a e speci ied in Tables 1 o 3, a e se in a da abase o
au oma ic model o maliza ion. Thus, he he mal pe o mance algo i hm changes a se o building geome y
a iables,
besides o he s. In his app oach he a iables we e he openings posi ion and size, o e hang dep h, la e al
size, and
heigh o he window, and he ins dep h. The pu pose is o minimize he weigh ed sum o deg ee-hou s o
he mal
discom o in all spaces on he loo plan. The deg ee-hou s a e calcula ed acco ding o he adap i e
ope a i e
empe a u e limi s o spaces na u ally en ila ed de ined by he Eu opean No m 15251:2007. The
sequence o he
op imiza ion p ocedu e is acco ding o he use ’s design s a egy.
3.
Showcase
3.1.
Building’s design p og am and speci ica ions
The u ban qua e is loca ed in Coimb a, Po ugal. I has an O-shape con igu a ion wi h en building blocks o
h ee le els. Fou s ee s delimi he u ban qua e . Th ee buildings a e placed in he No h side and o he h ee in he
Sou h side. E e y building o he u ban qua e has he same design p og am as speci ica ions. Howe e , he
gene a ed designs a e di e en , due o he loca ion o each building in he u ban qua e and he ex e io openings in
he g ound loo ha mus ace s ee side. In he case o he building placed in he No hwes co ne , i s ex e io
openings may be o ien ed No h and Wes . Simila ly, he emaining buildings sa is y his equi emen . The design
p og am o each building is de ined as ha ing h ee le els. In he g ound loo , he e is a s o e and an en ance hall.
In he i s loo , he e a e he co ido , he ki chen, he li ing oom, one bed oom, and one ba h oom. In he las
le el, he e a e h ee bed ooms and a ba h oom. One s ai se es all le els.
Besides he building’s geome y, he dynamic simula ion was ca ied ou using common cons uc i e and usage
speci ica ions o his kind o design p og am. Table 1 lis s he occupancy, schedules o use, ac i i y, and numbe o
occupan s in each space. Also, in Table 2 a e lis ed he a i icial ligh ing and equipmen in each space. Table 3
depic s he cons uc i e sys em o each building elemen and he ma e ials physical p ope ies. Besides hese
speci ica ions, each space ha has an ex e io opening wi h an in il a ion a e o 0.4 ACH. I a a speci ic ime, a
space is occupied and he empe a u e di e ence o he ex e io is one deg ee o supe io , a en ila ion a e o 1.5
ACH is summed up. The wea he da a is o Coimb a, a ailable in he US Depa men o Ene gy websi e.
Table 1. People occupancy, ac i i y, and schedule by space.
Time
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 {W} Pp
Hall/Co . (3x) 190 2
S ai 190 2
S o e 190 2
Ki chen 190 2
Li ing oom 110 5
Ba h oom (2x) 207 1
Bed oom (4x) 72 2
2902 Eugénio Rod igues e al. / Ene gy P ocedia 78 ( 2015 ) 2899 – 2904
Table 2. Ligh ing/equipmen gains and schedule by space.
Time
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 {W/m2}
Hall/Co . (3x) 7.0
S ai 7.0
S o e 20.0
Ki chen 10.0
Li ing oom 10.0
Ba h oom (2x) 7.0
Bed oom (4x) 7.0
Table 3. Ma e ials’ physical p ope ies by elemen .
Elemen
U- alue
Laye
T{cm}
λ{W/m-K}
σ{kg/m3}
Sp{J/kg-K}
TA
SA
VA
Ceiling/Slab
2.60
High weigh conc e e
20.0
1.73
2242.60
836.80
0.90
0.65
0.65
Ha dwood
3.00
0.20
825.00
2385.00
0.90
0.78
0.78
Ex e io doo
2.86
Insula ion
1.00
0.04
32.00
836.80
0.90
0.50
0.50
Ha dwood
1.00
0.20
825.00
2385.00
0.90
0.78
0.78
Plas e
2.00
0.43
1250.00
1088.00
0.90
0.60
0.60
Dense b ick
11.00
1.25
2082.40
920.50
0.90
0.93
0.93
Ex e io wall
0.43
Insula ion
8.00
0.04
32.00
836.80
0.90
0.50
0.50
Conc e e block
15.00
1.73
2242.60
836.80
0.90
0.65
0.65
Plas e (gypsum)
2.00
0.22
950.00
840.00
0.90
0.60
0.60
High weigh conc e e
20.00
1.73
2242.60
836.80
0.90
0.65
0.65
Floo 0.45
Insula ion
8.00
0.04
32.00
836.80
0.90
0.50
0.50
Lime plas e
2.00
0.80
1600.00
840.00
0.90
0.50
0.50
Ha dwood
1.50
0.20
825.00
2385.00
0.90
0.78
0.78
Ha dwood
0.50
0.16
720.80
1255.20
0.90
0.78
0.78
In e io doo
1.36
Chipboa d
3.00
0.07
430.00
1260.00
0.90
0.78
0.78
Ha dwood
0.50
0.16
720.80
1255.20
0.90
0.78
0.78
Plas e (gypsum)
2.00
0.22
950.00
840.00
0.90
0.60
0.60
In e io wall
2.17
Conc e e block
7.00
1.73
2242.60
836.80
0.90
0.65
0.65
Plas e (gypsum)
2.00
0.22
950.00
840.00
0.90
0.60
0.60
Slag
1.50
1.44
881.00
1673.60
0.90
0.55
0.55
Fel and memb ane
1.00
0.19
1121.30
1673.60
0.90
0.75
0.75
Roo
0.37
Dense insula ion
10.00
0.04
91.30
836.80
0.90
0.50
0.50
High weigh conc e e
20.00
1.73
2242.60
836.80
0.90
0.65
0.65
Plas e (gypsum)
2.00
0.22
950.00
840.00
0.90
0.60
0.60
Elemen
U- alue
Type
SHGC
VT
Window
2.60
Double Glazed Win.
0.63
0.70
T - laye hickness; λ - conduc i i y; σ - densi y; Sp - speci ic hea ; TA - he mal abso p ance;
SA - sola abso p ance; VA - isible abso p ance; SHGC – sola hea gain coe icien ; and VT – isible ansmi ance.
3.2.
Resul s and Discussion
A e gene a ion o he ou al e na i e designs o each building block, he geome y was changed aiming o
imp o e he he mal pe o mance o all solu ions. The bes solu ion o each building block is hen selec ed o he
inal u ban qua e design. The esul s o his showcase a e p esen ed in Fig. 1. In i a e depic ed ou iews
co esponding o ou ca dinal poin s (NW, NE, SW, and SE). The buildings’ heigh s a y due o he s ai ype and
i s
dimensions ha esul ed in he gene a ion p ocess. I is possible o obse e ha building blocks acing sou h ha e
o e hangs in he ex e io openings ha ace his o ien a ion. Also, he windows a e la ge han in he o he
buildings.
I is also no iceable he same esul in he middle building in he no h side o he u ban qua e , as he
ex e io
openings in he uppe le els a e ee o be o ien ed sou h.
Eugénio Rod igues e al. / Ene gy P ocedia 78 ( 2015 ) 2899 – 2904 2903
Fig. 1. (a) View om No hwes ; (b) iew om No heas ; (c) iew om Sou hwes ; and (d) iew om Sou heas .
Fig. 2. (a) The mal pe o mance esul s o he ou building designs in each building block (weigh ed sum o deg ee-hou s); (b) The mal
pe o mance esul s o each building. G ading g een- o- ed ep esen s bes - o-wo s o e all building he mal pe o mance. C4 has he bes
pe o mance and C1 he wo s . Top ow aces no h and bo om ow aces sou h.
Fig. 2 illus a es he he mal pe o mance esul s o he ou al e na i e designs in wo g aphics, ep esen ed by
he weigh ed sum o deg ee-hou s o he mal discom o . In he igh is ep esen ed a schema ic o he u ban qua e
whe e a colo g ading, om g een- o- ed, ep esen s he bes - o-wo s he mal pe o mance, espec i ely. I is
possible
o obse e ha buildings ha ha e walls wi h ex e io openings acing sou h ha e he bes pe o mance
(g een
colo ). The wo s pe o mance solu ions a e placed in he eas side o he qua e ( ed colo ). Thus, he bes
building
pe o mance solu ion is placed in C4 (sou heas co ne o he u ban qua e ). In he le side o he igu e is
depic ed
he so ed pe o mance o he ou design solu ions in each building block. As i is possible o obse e, he
esul s a e
no jus ui o andomness o design, as he al e na i e designs ha e hei pe o mance wi hin he same
ange o
alues. e.g. building block A1 has he ou designs anging be ween 37045 and 52875 deg ee-hou s and C4
be ween
14653 and 21652.

2904 Eugénio Rod igues e al. / Ene gy P ocedia 78 ( 2015 ) 2899 – 2904
The e o e, i he u ban qua e had o he con igu a ion, o ins ance a ho izon al O-shape con igu a ion wi h he
majo axis o ien ed o Eas -Wes , he numbe o building blocks wi h be e he mal pe o mance would be g ea e .
Thus, he use o his kind o app oaches is e y aluable in he u ban and a chi ec u al design p ocess, no only o
gene a e designs o be u he de eloped in ollowing s ages, bu also o explo e u ban con igu a ions ha aim o
sa is y unc ional and pe o mance objec i es. This app oach may be ex ended o include o he decision a iables
such as o cons uc i e and ac i e sys ems.
4.
Conclusions
In Po ugal, De ailed Plans a e e i o ial planning ins umen s p o ided by law, which embody he allo men
ope a ions [11]. Thus, d awing he lo comes i s , subjec ing in a iably he design o he building o he cons ain s
b ough by he p io design o he lo . Howe e , i he he mal pe o mance o he gene a ed solu ions a e assessed
and imp o ed, building p ac i ione s, u ban planne s o hose esponsible o he e i o y planning may ake
in o med
decisions, hus a oiding design co ec ions in u u e s ages o building design. The p esen ed showcase
demons a es ha each building pe o mance depends g ea ly om i s posi ion in he u ban qua e , and ha he
shape o he la e is dependen o he sa is ac ion o he in e io a angemen s o he design p og am. Resul s show
ha buildings wi h walls wi h ex e io openings acing sou h ha e he bes he mal pe o mance by ha ing la ge
window a eas p o ec ed by shading mechanisms. This kind o in o ma ion is e y impo an o p ac i ione s i hey
aim o design mo e ene gy e icien and li able buildings. So, new ools ha include pe o mance assessmen om
he ou se a e undamen al. Bo h algo i hms used in his case s udy a e now being de eloped as p o o ype
comme cial
ool unde he p ojec Ge APlanO [12].
Acknowledgemen s
This wo k is amed unde he Ene gy o Sus ainabili y Ini ia i e o he Uni e si y o Coimb a (UC) and has
been suppo ed by he p ojec Au oma ic Gene a ion o A chi ec u al Floo Plans wi h Ene gy Op imiza ion -
Ge APlanO - QREN 38922 P ojec (CENTRO-07-0402-FEDER-038922).
Re e ences
[1] IEA. T acking Clean Ene gy P og ess 2014. Pa is, F ance: 2014.
[2] Ve meulen T, Kämp JH, Becke s B. U ban o m op imiza ion o he ene gy pe o mance o buildings using Ci ySim 2013:915–20.
[3] Koenig R. Gene a ing u ban s uc u es: A me hod o u ban planning suppo ed by mul i-agen sys ems and cellula au oma a. ol. 3.
Bauhaus-Uni e si ä Weima ; 2011.
[4] Kämp JH, Robinson D. Op imisa ion o building o m o sola ene gy u ilisa ion using cons ained e olu iona y algo i hms. Ene gy Build
2010;42:807–14. doi:10.1016/j.enbuild.2009.11.019.
[5] Sanaieian H, Tenpie ik M, Linden K Van Den, Mehdizadeh Se aj F, Mo idi Shem ani SM. Re iew o he impac o u ban block o m on
he mal pe o mance, sola access and en ila ion. Renew Sus ain Ene gy Re 2014;38:551–60. doi:10.1016/j. se .2014.06.007.
[6] Rod igues E, Gaspa AR, Gomes Á. An e olu iona y s a egy enhanced wi h a local sea ch echnique o he space alloca ion p oblem in
a chi ec u e, Pa 1: Me hodology. Compu Des 2013;45:887–97. doi:10.1016/j.cad.2013.01.001.
[7] Rod igues E, Gaspa AR, Gomes Á. An e olu iona y s a egy enhanced wi h a local sea ch echnique o he space alloca ion p oblem in
a chi ec u e, Pa 2: Valida ion and pe o mance es s. Compu Des 2013;45:898–910. doi:10.1016/j.cad.2013.01.003.
[8] Rod igues E, Gaspa AR, Gomes Á. An app oach o he mul i-le el space alloca ion p oblem in a chi ec u e using a hyb id e olu iona y
echnique. Au om Cons 2013;35:482–98. doi:10.1016/j.au con.2013.06.005.
[9] Rod igues E, Gaspa AR, Gomes Á. Au oma ed app oach o design gene a ion and he mal assessmen o al e na i e loo plans. Ene gy
Build 2014;81:170–81. doi:10.1016/j.enbuild.2014.06.016.
[10] Rod igues E, Gaspa AR, Gomes Á. Imp o ing he mal pe o mance o au oma ically gene a ed loo plans using a geome ic a iable
sequen ial op imiza ion p ocedu e. Appl Ene gy 2014;132:200–15. doi:10.1016/j.apene gy.2014.06.068.
[11] Minis é io do Ambien e O do T e E. Dec e o-Lei n.o 136/2014 de 9 de se emb o 2014:4810–60.
[12] Ge APlanO P ojec 2013. h p://www.uc.p /en/e s/ esea ch/ge aplano.