Player is loading...
Urban morphology and building PV energy production
Kin Ho Poon, National University of Singapore, SG
Tuesday, May 7, 2019 · 10:30 a.m. · 13m 33s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
good morning everyone uh you can call me i even i'm
00:00:03
doing my p. h. d. at the national university of singapore
00:00:07
and uh this is part of my p. h. d. study the urban mythology and building peewee energy production
00:00:16
uh i just want to give a brief the introduction of the research lab i'm working
00:00:21
for a i mean apart from i'm doing uh my p. h. d. i'm also working forty
00:00:27
solar energy research institute of singapore um uh it is uh
00:00:32
the singapore's national institute for fights all the energy research in hats
00:00:38
for cool workgroups um sour was materials peewee more to
00:00:44
send more application that would that is the solar energy systems
00:00:48
so the group i'm belongs to is actually does or
00:00:52
been so it worked with work which works on the um
00:00:56
city scout free free the mapping choose optimisation all thought building
00:01:01
typology an urban mythology is full maximising buildings all the energy potential
00:01:06
so um what i'm working on these actually this part of the of the focus
00:01:12
and um they also support some kaufman initiated like this all the nova programs as well
00:01:20
so singapore emphasise the importance of sustainable development and one
00:01:25
of the approaches is to increase the reliance on renewable energies
00:01:29
however singapore's highly urban noise to we've all limited we need new opal options so
00:01:35
the energy is the most promising a
00:01:37
renewable resources ops so in two thousand fourteen
00:01:43
the singapore government uh announced plans to increase that option of
00:01:47
solar power to three hundred fifty mega will peak by twenty twenty
00:01:51
which is a a row you could for then uh to around
00:01:54
like five percent of the right project a peak electricity team on
00:01:59
and you can see in the picture they are some of my
00:02:01
in initiated it's like i'm not there's a project costs all the nova
00:02:07
so so no place actually hopping or uh many most of the uh
00:02:11
public ass taste in two is taught this solar panels on the roof tops
00:02:17
and now i'm the project i'm working on this and other
00:02:20
uh government project they they want us to develop some choose
00:02:23
to estimate the the soul the potential and try to uh
00:02:28
so that they can uh deployed a p. v. panels effectively
00:02:33
so effective deployment of solar panels is not
00:02:37
solely dependent on so that that building design
00:02:41
um but it is actually a significantly affected by the contextual to the
00:02:46
amp armand how which is still a bit over all of the morphology
00:02:52
so i'll come due to the urban oops shading by
00:02:56
neighbourhood buildings not act actually not or the building services are
00:03:00
suitable for people installation just lesion choose nowadays can can find
00:03:05
out the solar energy potential for existing building already funded community
00:03:10
but at the very beginning at if a really urban planning stage there's still a lack of choose and
00:03:18
also the knowledge for the urban planners and government officials
00:03:21
to estimate the solar energy potential often new planning so
00:03:26
so it is real um some researchers has lower our study the
00:03:32
relationship between urban mythology and the urban environmental performance of the urban temperament
00:03:39
an uh shows that that by the morphology has a notable impact on the
00:03:45
i'm all for uh energy consumption and also diesel the energy potential of buildings
00:03:53
it's an uh that's why i pay go urban morphology to uh
00:03:58
as the protectors to to to estimate the solar energy potential and the cooling load in singapore
00:04:06
and uh this other fall common types of approach to develop that the straight energy model
00:04:14
the uh one is actually the bottom up approach and the rest free out top down approach um
00:04:22
and for the morphological pays mortal and
00:04:26
ah well them the billing information you
00:04:30
require these but at the same time the accuracy of the quantification of the um
00:04:38
of the more full of the existing morphological based models ah it's probably a little uh
00:04:45
that and the other three approaches so so making sure that
00:04:49
uh you your morphological protective more though is accurate is very important
00:04:56
so there are some gaps to be filled with um uh the
00:05:01
first one is each city is identical for multiple logical studies um
00:05:06
uh you cannot replicate the are the previous study to to another city
00:05:13
so and there's a lack of study on finding the relationship
00:05:16
between the uh the morphology and solar energy potential in singapore
00:05:20
and the second thing is that the more for visual basic energy model is rare the for a up thing is that the
00:05:28
some architects and planners a sceptical about the usefulness of parametric pays a design supporting choose
00:05:34
and the four and get is that some study still bill ah
00:05:40
well the after feel studies that are top like a non parametric
00:05:44
machine than a for building the energy uh the straight energy model
00:05:51
so in the overall work for all of my project
00:05:55
has free stages to address the three objectives as well
00:05:59
um the first objective is to understand the relationship between the uh
00:06:03
morphology including solar energy potential and uh the cooling load in singapore
00:06:09
the second objective is to build a morphological pays up doing solar g. and a cooling
00:06:14
no protective model by machine learning and compare
00:06:18
it with the model developed by statistical regression
00:06:22
and uh for objective is to recommend an
00:06:24
optimist uh urban all morphological form that can maximise
00:06:29
all of the solar energy comes contribution to buildings ah energy consumption of me but when singapore
00:06:37
so the first stages to do the simulations g. for data collection
00:06:41
and based on the data collected by simulation will build a predictive model by both
00:06:47
the question and machine learning and the first stages to do to upload optimisation model
00:06:54
so for simulation no no first of all we need to pick
00:06:58
up some of the uh urban morphological parameters so based on the
00:07:05
literature we will i pay nine indicators for the first say uh
00:07:11
kiwi panels and five indicators for rooftop ah the
00:07:15
installation of up kiwi panels on the roof top
00:07:20
and i also shows like got the five uh indicators for for building energy consumption
00:07:26
um yeah i'm not the final protector as because i asked do need to
00:07:30
lie out tacit uh after brooding will um the the more though and and
00:07:35
see where the the the the protectors ah ah have a significant impact or
00:07:40
not but these are the the the parameters that i chose to be tested
00:07:48
and then ah
00:07:51
these items will be ah seem late at
00:07:55
the sole insulation simulation on roof top so
00:08:00
the installation uh on the side and also this guy will factor and also to in the
00:08:05
the peewee energy generation and also the cooling low uh um
00:08:11
a simulation so so i'd curving that shows a lady
00:08:16
part and also the high maybe to to do the simulations
00:08:19
but sometimes act i i've fun find that it is actually quite a time
00:08:23
consuming to to use this to to use to it is in the nation
00:08:28
so now actually i'm also that that testing whether i i can reduce
00:08:32
uh the the time for simulation by using a city seem as well
00:08:39
so uh this is the sample case of of the simulations
00:08:43
uh basically you can see that um t. v. can those also
00:08:50
served as a a shaving device on on um
00:08:55
on the buildings save you install a so set
00:08:59
different threshold and different p. week after age you
00:09:02
will get that you will also affect the energy consumption this well not only the purely energy generation
00:09:10
so the second stage of my project used to do the protective model building
00:09:16
and uh i will use 'em both the you re question more
00:09:21
though not as the most such additional techniques used by most previous studies
00:09:26
and i also will which is some machine learning techniques like new network
00:09:31
and also gaussian process and compare the results with
00:09:35
those models with the traditional ah multi re question model
00:09:43
um the first stage is optimisation um
00:09:48
optimisation is a bit tricky because um nah
00:09:53
doing simulation optimisation requires quite well uh sometime
00:09:59
and uh but uh i will watch has to
00:10:03
to uh optimisation techniques one is called the
00:10:06
simulation optimised station techniques which means that you um
00:10:11
laying your optimisation i agree from with your simulation program
00:10:16
so evolutionary algorithm is the most free frequently a type of algorithm used
00:10:22
and um there as i mentioned out very feel
00:10:26
studies are um have been explored for optimisation optimising the
00:10:33
the energy a generation peewee energy generation in districts gal
00:10:39
more optimisation studies have been performed on that building scale
00:10:44
so uh and it's do need to explore a different kinds
00:10:47
of uh optimisation algorithms like the the algorithm developed by jerome
00:10:53
another approach is actually call the model based optimisation ah which means actually you use a surrogate
00:10:59
gate more do instead of linking it to the simulation a pro as to your simulation outward from
00:11:06
mm theoretically it will run faster but you have to make sure that this
00:11:11
or okay more the acrid uh will actually pull white you have very accurate resell
00:11:18
so this is the first stage and uh hopefully
00:11:22
the potential contribution of my research is actually the
00:11:25
i'm all for that a logical pays at this rate energy model coding is very very rare
00:11:32
yeah when you compared to other other the straight energy models to you find that um
00:11:41
the existing choose nowadays off useful when your oven form has been confirmed
00:11:48
so before it uh what at the fair it uh or the urban planning stage um you and
00:11:55
probably need another to for estimating the energy consumption that could be energy potential
00:12:02
so um this the last light of my presentation
00:12:06
and uh up because i still have a little bit of time just wanted
00:12:09
um introduced this project is is caught the way choosing a pro project ups
00:12:16
i i i discord do you work to singapore projects
00:12:21
a single which is singapore project is actually um it contains
00:12:25
all the three d. information ah of just the whole entire
00:12:29
city on the on the and uh on a single platform
00:12:34
so basically um for validation purpose i
00:12:39
will use these flat fee the model too
00:12:43
put it in my in in my eye with from n. c. whether um
00:12:48
my model camper why the actor right uh we sell
00:12:55
so they contain or the
00:12:59
or the building information
00:13:04
in the same platform in the fruity fruity case well
00:13:08
the information is not love publicly available yet but i just
00:13:11
want to show them what was going on in singapore and and this is the um the the model of the um
00:13:18
the buildings that you can't get and is very convenient for the researchers
00:13:24
so thank you very much and uh if you have any questions uh
Conference Program
Intelligent feed-back on user comfort in low energy buildings
Rui Oliveira, University of Aveiro, PT
May 7, 2019 · 10:09 a.m.
169 views
Urban morphology and building PV energy production
Kin Ho Poon, National University of Singapore, SG
May 7, 2019 · 10:30 a.m.
220 views
Urban morphology, energy needs and artificial intelligence
Roberto Boghetti, University of Pisa, IT
May 7, 2019 · 10:51 a.m.
121 views
Distributed simulation applied to multi-networks urban energy systems design
Pablo Puerto, CREM / Mines d’Albi, F
May 7, 2019 · 11:13 a.m.
113 views
Les défis du domaine Energy Informatics
Jérôme Kämpf, Idiap Research Institute
May 7, 2019 · 11:32 a.m.
530 views
The Ark Energy
Laura Schwery, Fondation The Ark
May 7, 2019 · 1:38 p.m.
232 views
Industrie du futur
Richard Pasquier, HEIA-FR et Benedikt Ramsauer, Swiss-SDI
May 7, 2019 · 1:46 p.m.
281 views
Naïade : gestion optimisée d’un réseau d’eau
Benoît Golay, Institut Icare
May 7, 2019 · 2:07 p.m.
Open data des villes
Duccio Piovani, Research Lead - Data Science, nam.R
May 7, 2019 · 2:27 p.m.
IA, L’entreprise augmentée
Dominique Genoud, HES-SO Valais-Wallis
May 7, 2019 · 2:52 p.m.
Planifier le quartier de demain avec l’IA
Diane von Gunten, CREM
May 7, 2019 · 3:14 p.m.
Recommended talks
Workshop introduction by Jérôme Kämpf & Henning Müller + Keynote speech Pr. Dr. Guglielmina Mutani - Place-based energy models for sustainable cities
Pr. Dr. Guglielmina Mutani, Politecnico di Torino
Nov. 15, 2021 · 9 a.m.
333 views
Solar energy - part 2
Prof. François Maréchal, EPFL
May 21, 2012 · 5:19 p.m.
131 views