Player is loading...
Dr. Kavan Javanroodi - Extending the concept of energy hub to facilitate sector and spatial coupling
Dr. Kavan Javanroodi, Solar Energy and Building Physics Laboratory (LESO-PB) at EPFL
Monday, Nov. 15, 2021 · 11:59 a.m. · 16m 22s
Understanding the complex interactions within different sectors (sectoral) and localities (spatial) of cities plays a crucial role in improving efficiency and sustainability, which is extremely challenging due to the complex urban morphology. We introduce the Urban Cell concept to address this bottleneck. Optimal sizing of the distributed energy system (including renewables, energy storage, and dispatchable sources) and optimal urban morphology is determined within a modular unit. A game-theoretic approach is used to model the interactions between urban cells (modular units).
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:01
just sell from the store in physics and the tree so he's uh also with the professors countries
00:00:08
in egypt and his green should jog j. both
00:00:12
energy homes and sector copy yeah special copping actually
00:00:20
sector and special come with it
00:00:29
yes
00:00:37
thank you very much uh so um comments on the pro as well
00:00:41
and today i think for the next fifty minutes i'm not talking about production and technically machine learning for now
00:00:48
and i'm going to present now propose concept urban cell which is
00:00:53
focus on enabling sector and special copying design optimisation of energy system and benefits
00:00:59
so why would a sector copping so we know that we need to reduce greenhouse gas emissions by
00:01:06
i'm more than eighty to ninety five per se it's compared to lightning levels
00:01:11
european countries and put doing data we need to be capitalised he would be really system
00:01:17
and of course so we need to prove the efficiency and flexibility of energy system
00:01:22
prior when we going to design tools and main challenges reducing the cost of the combination actually
00:01:29
so uh i'll watch we can do is focusing on four major sectors inside every area
00:01:36
which is uh energy supply industry transport and of course energy demand from the buildings
00:01:41
and try to reduce that the conversation cost in dos sectors so
00:01:46
sector coupling is technically recognise that the key the energy transition for future cities
00:01:53
so by into connecting the energy consuming sector like
00:01:56
building transport industry leafy every neighbourhood in the street
00:02:00
uh we can do we have so originally developed in germany i think the
00:02:05
jen what is sector could plot hopefully it doesn't light and the main
00:02:10
cause it was electrifying a new sector so that was the main focus of
00:02:14
sector copy and of course uh the main definition that we have a from a report is to
00:02:21
provide greater for the flexibility to energy system and
00:02:24
uh to be capitalised always with cost effective way
00:02:28
we also have a several recent reports and national international
00:02:33
reports concerning sector probably out of course special coupling which is
00:02:38
technically doze inter connectivity between different energy systems in our uh
00:02:43
real urban area so uh this think that art art in considering
00:02:49
a special and sector cop link there are several concept like smart greed
00:02:54
we have energy heart which is technically an extension to hybrid energy system concept
00:03:00
and of course the goal of those are uh to improve the privilege integration urban areas
00:03:06
and uh uh we have also other metaphors like you never systems like
00:03:11
uh every metabolism or industrial ecology which is more a holistic
00:03:16
and uh if you want to summarise the problem cause
00:03:19
word is on a present concept in this regard so
00:03:24
none of them can bold uh consider sector and special copying the same time so
00:03:30
s. y. agreed energy hop a hybrid energy system was uh mostly focus or you sector
00:03:35
while the everyman thousand concept can consider those bytes can that
00:03:40
consider system dynamics and that's the main problem so
00:03:43
it's not specifically designed for energy system optimisation actually
00:03:48
so we want to go to our special sector coupling so we present
00:03:52
it um every cell concept so if we consider c. t. v. a.
00:03:57
here uh we have different the streets and different
00:04:00
uh areas each of them uh consider off different neighbourhood
00:04:05
which we call them as sam and this is one sale in our concept which of course
00:04:10
there are several services in layers in that sail
00:04:12
from energy to building started transportation and other sectors
00:04:17
so that's the first step for sector competent but of course we need to go
00:04:21
beyond that and consider special coupling to so between each cell
00:04:26
uh we have also uh into connectivity problem and that we need
00:04:31
to consider that in the design optimisation of urban areas you system
00:04:35
so we try to come up with that uh
00:04:38
a concept and try to hopefully consider all of those
00:04:43
a positive a requirement to optimise energy system so
00:04:49
there was solar network is technically based on um i'm connecting the for
00:04:55
every cell in every uh every how could could be district for example
00:04:59
and create a distributed multi agent design problem so uh instead of focusing on one system
00:05:06
we're dealing with is the marxist and so most of the studies in design problem
00:05:11
a specially designed control the problem we also have that button design problem or
00:05:16
focusing on one unit want energy system try to optimise that but we're dealing with the system a system
00:05:21
so uh for the operation mode we're not focusing on control of course here but corporation mode we
00:05:28
uh in a proposal later father was strategy which uh it's a technically efficient way to
00:05:34
consider into connectivity between each cell uh it's a combination of flat fully cooperative and non cooperative
00:05:40
which up uh each of them has some negative points for
00:05:43
example fully cooperative operation mode uh we have a central authority intervals
00:05:47
which uh for shore can have a can lead to a more efficient energy system but
00:05:53
we have privacy issues they're not cooperative so we have some ages that compete with each other
00:05:59
and of course we have a good privacy there but sometimes
00:06:02
will have probably problem in having optimism more efficient every assistants
00:06:08
ah i'm trying to be fast uh talk about the optimisation for what we have a two level optimisation
00:06:16
uh here uh out for the first step we going to looking into the obscene mall
00:06:22
uh and cool optimisation off every system and every
00:06:25
morphology and of course the transportation sector that we consider
00:06:30
and um uh based on now a days after our optimisation a work flow
00:06:35
and then uh of course here we use like morse purity out with guns
00:06:40
you optimise our problem considering the complexity to the egg up the problem so
00:06:44
previously our team uh for example um worked on more uh when i say
00:06:50
um evolutionary out virtual machine learning wally deep blue ones learning for the dispatch problem
00:06:57
but sees this problem was quite complex at this stage we couldn't use that
00:07:01
and we're focusing on m. t. v. and gee i'd with integration so and t. v.'s
00:07:05
a life cycle of all the costs inside uh the uh design
00:07:10
distributive energy system that is based on t. v. panels winter turbines
00:07:14
we have feet palms and the dispatcher energy source of course
00:07:18
to uh dealing with image met a between supply and a demand
00:07:23
and we also have the greed integration uh as a function of
00:07:26
our optimisation which uh dealing with the autonomy of the neighbourhood so uh
00:07:32
uh at the presence of the second step uh we try to organise our absolution
00:07:37
uh from the first is that and make them ready for the third is that
00:07:41
uh which is energy network optimisation so um each
00:07:45
uh had a solution of a unique uh evan cell
00:07:49
and of course are organised in descending into the order to be using the third
00:07:54
a step in the service that we have a two stage of optimisation we optimise network configuration
00:08:00
connecting or been cells and the decision vector and then we compute the flexibility of each line in the network
00:08:08
and uh finally we can say that we're optimising the strength
00:08:12
of the network and system configuration which means design enough energy system
00:08:16
and the fourth and fifth uh let's say steps are of course
00:08:21
over into the for uh the first uh one two or three step
00:08:25
so there are a running iteratively until richie
00:08:29
i'm madly epsilon nash equilibrium and then we uh
00:08:33
have optimal episode configuration in the forces there
00:08:36
and finally optimal network connection using energy internet concept
00:08:40
uh so we as a each cell we have interconnected energy infrastructure so we talk about
00:08:47
it was hard to talk about building start remote talk about transportation for each cell which is
00:08:52
the neighbourhood and of course going to have like up to mall output for each of them
00:08:57
whatever morphology model uh well we using a these three
00:09:01
elements of every form function and structured so every form usually
00:09:05
uh the the because it is used uh uh instead of
00:09:10
a bum apology every informers focusing on job which we call part
00:09:14
and based on this for major a morphological parameters of course a range of ten
00:09:20
problem utters we dove lope and evan area consisted of night cells which
00:09:25
is cell is that neighbourhood and with a different like is a multi function
00:09:30
uh every every our hobbies building residential building low medium and high rise building
00:09:35
we we came up with and one actually muttering one
00:09:38
kilometre area we had to design this to have more control
00:09:42
on our uh problem so we couldn't use existing uh let's say uh areas
00:09:49
and of course we based on the function based on the fancy of each building
00:09:53
uh we did a double the model for the transportation i.
00:09:57
e. v. uh charges station should be connected to energy system
00:10:02
and we're out of course the mold is on a better just immolation forty so
00:10:06
if we go back to that cell concept technically we're dealing with these nine cents
00:10:12
so first we uh try to consider and the access the impact
00:10:16
of every morphology and e. v. on the uh i well the resolution
00:10:20
and i'm passing this because i think i'm drawing of
00:10:23
time but uh out there willing to uh s. s. v.
00:10:28
uh you see and ever in a solar network impacts so we divide
00:10:33
some benchmarks never use your so we affix to ration without sector and
00:10:38
inter connectivity or special coupling we have uh i selected couplings infrastructure another scenario i see
00:10:44
i which only consider sector coupling and of course you see at our concept which consider both
00:10:49
so if we uh look at the pad to solutions we can see
00:10:52
up to a thirty seven percent reduction in system cost using u. c. n.
00:10:58
and that's um compared to uh actually yes but for all the sector cop link model
00:11:04
scenario i see i we only have like uh about seven percent reduction of the infrastructure cost
00:11:12
and uh another impact is still looking into yeah more detail so
00:11:17
here we have a few select three p. h. u. i. r.
00:11:21
hi to solutions which are minimise or minimal uh n. p. v. uh for each of the scenario if you look
00:11:28
into that we see different configuration for of apology the bachelor that of a roomful you has a great impact your
00:11:34
and we also can see with a like a a
00:11:37
more detailed uh inciting did those solutions we can see that
00:11:41
uh every cell can improve the readability potential by twenty five percent uh as
00:11:45
solution or is from u. c. and and solution choose from i. c. i.
00:11:49
and um what about the scalability of the proposed approach whether this can be scaled op
00:11:56
to be the problems so would you all uh allies
00:11:59
those three benchmark serious together from for sale to sit
00:12:03
cell and considering nice l. so we can see from when we increasing the number of the sales were neighbourhood
00:12:09
we'll see that uh the the uh i. c. i. n. f. g. s. was trying to pack a
00:12:13
solution which run past the reason is that they will need because there's no connection between each every system
00:12:19
they will need to have more distribute it reduces them to uh actually um but i
00:12:26
get the demands opted area and that will uh that will lead to higher cost of course
00:12:32
but to using d. n. t. v. of course uh we'll have to
00:12:36
be like a twenty five percent uh less uh m. t. v. reduction
00:12:41
and finally uh what about these uh impact of system what is the mafia architecture
00:12:47
so we have a game terrific approach to assess that so we have i. c. i. sector coupling
00:12:52
and we have game one again to which are two consecutive games in the
00:12:56
optimisation problem and finally you see and which is the twenty iteration of the games
00:13:02
uh here we see that in the first into game uh id using concept
00:13:06
somehow coincide with the uh so uh i. c. i. scenario which only consider sector coupling
00:13:12
but after the third iteration it will be more similar to do
00:13:15
what we see here which is the twenties a generation and a great
00:13:20
a reduction in t. v. and of course flour a grade interracial level which lead to higher
00:13:26
a neighbourhood uh attorney so if we summarise the main find it's we know that
00:13:33
we have the problem we know that we should consider sectors special coupling and this concept
00:13:37
showed us that a cost up in the integrated
00:13:40
infrastructure can be reduced by sarah thirty seven percent
00:13:45
and a renewable energy penetration can be uh improve by twenty five percent
00:13:51
yeah there are some uh ongoing work based on these uh concept
00:13:55
so we need to uh use and add more uh uncertainties on to a certain sources so
00:14:02
what about climate uncertainty climate change extremely been micro climate conditions that would all
00:14:06
of those could be important our model and of course you might response management
00:14:11
so uh instead of focusing on actually break the date of one neighbourhood what about
00:14:15
considering all uh buildings as agent and
00:14:19
a move toward the multi agent optimisation problem
00:14:23
so we don somewhat in that regard for like thirty eight european cities uh in the concern too
00:14:29
climate change we saw great impact especially put according to many future
00:14:33
and those are something that should be considered future and uh oh
00:14:38
climate resilient interconnected or because it by structure we should uh they're improve
00:14:43
the resiliency of uh air bases system to the climate in extreme events
00:14:47
uh and so we also don uh some some working that record with the same team but
00:14:53
uh this is something that should be added to the every cell concept and i like the biggest there we only use
00:14:59
a t. n. y. file anymore spoil based on average about was the
00:15:02
question which lead us to the other problem which is micro climate data
00:15:07
uh we we work on that also and we uh relies
00:15:10
the draw some a major problem and micro climate impact the
00:15:13
u. b. m. evan building years you model is up to thirty five percent and can be even here in extreme events
00:15:20
problem is for animal is still analysis we usually use a been building a energy equations bottles equations
00:15:27
but we need more actually rate production so of course we have a under review wait uh which
00:15:32
she uh developed a hybrid model base and see if he and the plural network to pretty big
00:15:37
michael climate data alan and well the scale using c. f. c.'s which i think
00:15:42
that could also be some another source that could be added and finally
00:15:47
uh of course moving from grey back to black was motley going to badly mont response management
00:15:53
uh this is uh for for one unit of energy
00:15:55
system uh our a team actually did some analysis using deep
00:16:00
reforms learning and uh find out like a improving by sixty percent that's something we couldn't
00:16:06
do it we're our concert now because it was too complex to be honest to do that
00:16:10
i read yeah but we're working on that we're going to uh dad direction ah
Conference Program
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
Q&A (Keynote speech: Pr. Dr. Guglielmina Mutani)
Pr. Dr. Guglielmina Mutani, Politecnico di Torino
Nov. 15, 2021 · 9:49 a.m.
Roberto Boghetti - Simulating District Heating Networks’ hydraulics with residual physics neural networks
Roberto Boghetti, Idiap Research Institute
Nov. 15, 2021 · 10 a.m.
Q&A (Roberto Boghetti)
Roberto Boghetti, Idiap Research Institute
Nov. 15, 2021 · 10:13 a.m.
Amine Weibel - Machine learning forecast for photovoltaic power stations (PV) with spatial interpolation on the COSMO-7 model. A real ‘use-case’ on the Czech Republic
Amine Weibel, HES-SO
Nov. 15, 2021 · 10:16 a.m.
Q&A (Amine Weibel, HES-SO Valais-Wallis)
Amine Weibel, HES-SO
Nov. 15, 2021 · 10:27 a.m.
101 views
Dr. Giuseppe Peronato - HeatLearn: a convolutional neural network to assess the building energy consumption at regional scale from open land-use maps
Dr. Giuseppe Peronato, Idiap Research Institute
Nov. 15, 2021 · 10:30 a.m.
Q&A (Dr. Giuseppe Peronato)
Dr. Giuseppe Peronato, Idiap Research Institute
Nov. 15, 2021 · 10:43 a.m.
Pr. Dominique Genoud & Jérôme Treboux - Integrating machine learning and AI in practical energy applications : a review of these last 10 years activities at the Institute of Information Systems
Pr. Dominique Genoud & Jérôme Treboux, HES-SO Valais-Wallis
Nov. 15, 2021 · 10:46 a.m.
Q&A (Pr. Dominique Genoud & Jérôme Treboux)
Pr. Dominique Genoud & Jérôme Treboux, HES-SO Valais-Wallis
Nov. 15, 2021 · 10:59 a.m.
Dr. Marco Tognoli - Fuel-saving potential simulation and estimation when optimally predicting the supply set-point conditions of a District Heating Network with detailed system data-driven modelling
Dr. Marco Tognoli, Politecnico di Milano
Nov. 15, 2021 · 11:31 a.m.
Q&A (Dr. Marco Tognoli)
Dr. Marco Tognoli, Politecnico di Milano
Nov. 15, 2021 · 11:44 a.m.
Pr. David Wannier & Jean-Marie Allder - Research advance in Energy prediction for electrical vehicles and green mobility hotels, the results of the eVIP project
Pr. David Wannier & Jean-Marie Allder, HES-SO Valais-Wallis
Nov. 15, 2021 · 11:46 a.m.
Q&A (Pr. David Wannier & Jean-Marie Allder)
Pr. David Wannier & Jean-Marie Allder, HES-SO Valais-Wallis
Nov. 15, 2021 · 11:58 a.m.
Dr. Kavan Javanroodi - Extending the concept of energy hub to facilitate sector and spatial coupling
Dr. Kavan Javanroodi, Solar Energy and Building Physics Laboratory (LESO-PB) at EPFL
Nov. 15, 2021 · 11:59 a.m.
143 views
