HYPERGRYD. This p ojec has ecei ed unding om he Eu opean Union’s Ho izon 2020 esea ch and
inno a ion p og amme unde g an ag eemen No 101036656
WP3 – ICT Modules and Simula ion Tools
Task 3.3 – De elopmen o no el machine
lea ning echniques d i en by IoT o LEC-
in eg a ed DHC ne wo ks
D3.4 – No el IoT-d i en edge compu ing
me hodologies o LEC-in eg a ed DHC
ne wo k applica ions
Re . A es(2024)6909419 - 30/09/2024
D3.4 No el IoT-d i en edge compu ing me hodologies
o LEC-in eg a ed DHC ne wo k applica ions 2
DISCLAIMER
The opinion s a ed in his epo e lec s he opinion o he au ho s and no he opinion o he
Eu opean Commission.
All in ellec ual p ope y igh s a e owned by HYPERGRYD conso ium membe s and a e p o ec ed by
he applicable laws. Rep oduc ion is no au ho ized wi hou p io w i en ag eemen .
The comme cial use o any in o ma ion con ained in his documen may equi e a license om he
owne o ha in o ma ion.
ACKNOWLEDGEMENT
This p ojec has ecei ed unding om he Eu opean Union’s Ho izon 2020 esea ch and inno a ion
p og amme unde g an ag eemen Nº 101036656.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 3
P ojec
P ojec Ac onym
HYPERGRYD
P ojec Ti le
Hyb id coupled ne wo ks o he mal-elec ic in eg a ed Sma Ene gy Dis ic s
G an Ag eemen
numbe
101036656
Call iden i ie
H2020-LC-GD-2020
Topic iden i ie
LC-GD-2-1-2020
Inno a i e land-based and o sho e enewable ene gy echnologies and hei
in eg a ion in o he ene gy sys em
Funding Scheme
Resea ch and Inno a ion Ac ion
P ojec du a ion
42 mon hs (F om 1 Oc obe 2021)
Coo dina o
ARCbcn
Websi e
h p://hype g yd.eu
Deli e able
Deli e able No.
3.4
Deli e able i le
Me hodologies and amewo ks o op imize demand esponse and peak load sha ing
o 4 h-5 h DHC in h ee ypical Eu opean clima es
Desc ip ion
D3.4 (No el IoT-d i en edge compu ing me hodologies o LEC-in eg a ed DHC
ne wo k applica ions) summa izes he main ac i i ies in Task 3.3 (De elopmen o
no el machine lea ning echniques d i en by IoT o LEC-in eg a ed DHC ne wo ks)
Ene gy communi ies, combined wi h IoT and ene gy s o age echnologies, a e key
enable s o enewable ene gy in eg a ion in he 4 h and 5 h gene a ion dis ic
hea ing and cooling sys ems. In his con ex , edge compu ing is c ucial o op imizing
hese echnologies wi h low cos s and high lexibili y. In Task 3.3, KTH ocuses on
de eloping a s andalone, secu e, and scalable edge compu ing concep o dis ic
hea ing-powe ed communi ies, ensu ing local da a p ocessing and enhanced
eliabili y. This ask also includes edge-based machine lea ning me hods, allowing
join op imiza ion o ene gy communi ies and enewables wi hou cloud da a sha ing.
WP No.
WP3
Rela ed ask
Task 3.3 - De elopmen o no el machine lea ning echniques d i en by IoT o LEC-
in eg a ed DHC ne wo ks
Lead Bene icia y
3 - KTH
Au ho (s)
Mus apha Habib (KTH), Qian Wang (KTH)
Con ibu o (s)
Vale ia Palomba (CNR)
Type
R
Dissemina ion
PU
Language
English – GB
Due
30/09/2024
Submission da e
30/09/2024
Ve sion
Da e
Au ho s
Desc ip ion
V.0.1
19/09/2024
Mus apha Habib (KTH)
The i s e sion o in e nal e iewing
V.0.2
23/09/2024
Vale ia Palomba (CNR)
Re iew
V.1.0
28/09/2024
Mus apha Habib (KTH)
Final e sion a e e iew
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 4
Table o Con en s
1. Execu i e Summa y ............................................................................................... 7
2. In oduc ion .......................................................................................................... 9
2.1. Scope ................................................................................................................. 9
2.2. Audience ........................................................................................................... 9
2.3. Abb e ia ions .................................................................................................... 9
2.4. Con ibu ions o pa ne s .................................................................................. 10
2.5. Rela ion o o he ac i i ies ................................................................................ 10
2.6. S uc u e .......................................................................................................... 11
3. P oposed edge compu ing solu ions o sec o coupling ........................................ 12
3.1. Edge solu ion o coo dina ed DSM o DH-powe ed communi ies ...................... 12
3.1.1. So wa e a chi ec u e ....................................................................................... 12
3.1.2. F om ene gy me e s o he edge ....................................................................... 13
3.1.3. F om he edge o he DH communi y membe ................................................... 14
3.2. Edge solu ion o upg ading he building managemen sys em .......................... 15
3.2.1. Limi a ions o he cu en solu ions ................................................................... 15
3.2.1.1. BMS limi a ions ............................................................................................. 15
3.2.1.2. Cloud-based solu ion limi a ions ................................................................... 17
3.2.2. Edge compu ing as a local op imiza ion-based EMS ........................................... 17
3.2.3. Edge panel – ha dwa e componen s .................................................................. 18
3.2.4. Edge panel – so wa e componen s ................................................................... 19
4. De elopmen and deploymen s a egies .............................................................. 20
4.1. Edge solu ion o DH communi ies managemen ............................................... 20
4.1.1. Da a s eaming o ma om he ene gy me e s ................................................ 20
4.1.2. Da a se ices om he DSM-se e ................................................................... 21
4.1.3. Da a exploi a ion in HYPERGRYD ....................................................................... 23
4.2. Edge o BMS upg ading .................................................................................... 25
4.2.1. De elopmen phase a KTH ............................................................................... 25
4.2.2. Deploymen phase a KEZO ............................................................................... 26
4.2.3. Tes phase p ocedu e ........................................................................................ 28
4.2.4. Nex s eps ......................................................................................................... 31
Conclusions .................................................................................................................... 33
Rela ion o con inued de elopmen s ............................................................................. 33
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 5
Re e ences ..................................................................................................................... 34
Lis o Figu es
Figu e 1 The so wa e a chi ec u e o he VM-based edge solu ion .................................................. 13
Figu e 2 Da a collec ion p ocedu e (da a jou ney om he DH communi y o he edge) ................. 13
Figu e 3 Coo dina ed con ol as DSM s a egy (da a jou ney om he edge o he DH communi y) 15
Figu e 4 Typical a chi ec u e o building managemen sys em .......................................................... 16
Figu e 5 da a ci cula ion in classical a building managemen sys em ................................................ 17
Figu e 6 edge-BMS connec ion poin .................................................................................................. 18
Figu e 7 ha dwa e componen s o he edge panel ............................................................................. 19
Figu e 8 so wa e componen s o he edge panel .............................................................................. 19
Figu e 9 Da a o ma being sen by he MQTT b oke ........................................................................ 21
Figu e 10 DSM-se e accessible se ices ........................................................................................... 22
Figu e 11 DSM con ol execu ion mechanism: on he le -hand side, di ec con ol ia API o IoT ( eal-
ime), on he igh -hand side, manual con ol (p edic i e) ................................................................ 22
Figu e 12 Addi ional compu ing ha dwa e o ensu ing eal- ime emo e con ol ............................ 23
Figu e 13 DSM simula ion (communi y-cen ic), op: HP op imal condense empe a u e se poin ,
bo om: TES SOC a ia ion (bo om igu e) in a 2-day simula ion)..................................................... 24
Figu e 14 Di e en pu poses o connec ing o he DSM-se e da abase in he amewo k o
HYPERGRYD ......................................................................................................................................... 24
Figu e 15 Displaying senso da a o each building in Sonnenpla z communi y ia he p ojec pla o m
............................................................................................................................................................. 25
Figu e 16 Edge panel du ing assembling and es a KTH campus ...................................................... 26
Figu e 17 The a ge ed hea ing and cooling p ocesses in he KEZO HVAC sys em (SCADA iew) ...... 27
Figu e 18 Edge panel du ing commissioning a KEZO esea ch cen e ............................................... 28
Figu e 19 Da a low du ing es phase o edge compu ing (AI-enabled PC is no included a his s age)
............................................................................................................................................................. 29
Figu e 20 Web dashboa d o an ope a ion scena io o he CO2 hea pump o simul aneous hea ing
and cooling .......................................................................................................................................... 29
Figu e 21 Web dashboa d o he same ope a ion scena io gene a ed locally by he edge panel .... 30
Figu e 22 Sc eensho om he KEZO SCADA sys em o he main s a us da a o HP du ing he
ope a ion es ...................................................................................................................................... 30
Figu e 23 s andalone EMS ole o he edge panel (nex con igu a ion) ............................................. 31
Figu e 24 Simple ep esen a ion o he ein o cemen lea ning echnique o he hea pump con ol
............................................................................................................................................................. 32
Figu e 25 So p ion s o age du ing he design phase .......................................................................... 32
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 6
Lis o Tables
No able o igu es en ies ound.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 7
1. Execu i e Summa y
The goal o he HYPERGRYD p ojec is he de elopmen o a se o eplicable and scalable cos -
e ec i e echnical solu ions o allow he in eg a ion o Renewable Ene gy Sou ces (RES) wi h
di e en dispa chabili y and in insic a iabili y inside The mal G ids as well as hei link wi h he
Elec ical G ids, including he de elopmen o inno a i e key componen s, in pa allel wi h inno a i e
and in eg a ed ICT se ices o med by a scalable sui e o ools o he p ope handling o he inc eased
complexi y o he sys ems om building o Local Ene gy Communi y (LEC) le els and beyond, and
accele a e he sus ainable ans o ma ion, planning, and mode niza ion o Dis ic Hea ing and
Cooling (DHC) owa ds 4 h and 5 h gene a ion.
HYPERGRYD also aims o de elop eal- ime managemen o elec ical and he mal ene gy lows in he
coupled ene gy ne wo k complex, including he syne gies be ween hem. The e o e, HYPERGRYD
aims a h ee o e -a ching Gene al Objec i es:
• To p o e ha sma ene gy ne wo ks a e he u u e o e icien ene gy managemen in DHC
and a e in syne gy wi h he elec ical g ids in LEC/sma ci ies o he u u e.
• To de ine he oadmap o design and plan u u e DHC as well as he mode niza ion o he
exis ing ones in di e en clima es and RES pene a ion le els owa d 4 h-5 h gene a ion,
• To demons a e HYPERGRYD RES-based Enabling Technologies, Sma Ene gy G id Solu ions
empowe ed by new ICT ools and se ices as he key o his e olu ion.
Du ing he p ojec , HYPERGRYD’s solu ions will be implemen ed ac oss ou Li e-In-Labs cases in
h ee ep esen a i e clima es, wi h special conside a ion o hei cos e ec i eness and po en ial
eplicabili y o achie e hese h ee main objec i es. One o he a ge ed miles ones in he
HYPERGRYD p ojec is he exploi a ion o edge compu ing o hos ing da a moni o ing and con ol
unc ionali ies dedica ed o ene gy ne wo k managemen , ei he a a building o a communi y le el.
The e o e, he edge solu ion amewo ks a e designed o be e ec i e on h ee di e en le els:
• Sys em le el: he e he edge o e sees he iden i ica ion o he dynamic o complex ene gy
sys ems (e.g., indus ial hea pumps (HPs), so p ion s o age), which migh unlock any
po en iali y o op imal con ol and managemen unc ionali ies. Addi ional se ices such as
p edic i e main enance and anomaly de ec ion can be inse ed.
• Ne wo k le el: he e he edge upg ades he compu ing and communica ion capabili y o he
building managemen sys em (BMS) by le e aging deep da a analy ics and op imiza ion as an
ad anced decision-making ool. He e, he edge can include online se ices (cloud-based) o
be e pe o mance as wea he o ecas and ene gy ma ke ends.
• Communi y le el: in some si ua ions, pa icula ly in esiden ial communi ies, he end-use s
migh c ea e an ag eemen o coo dina e hei ene gy u iliza ion by sha ing hei da a and
gi ing access o hei H&C sys ems. He e, he edge solu ion can enable communica ion ac oss
he communi y and hos his coo dina ed demand-side managemen (DSM).
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 8
By pe o ming hese unc ionali ies, he designed edge compu ing s a egies open he pa h owa ds
a smoo h and e ec i e ansi ion o ac ual DH ne wo ks o 4 h and 5 h gene a ions whe e he ene gy
managemen is mo e da a-d i en.
The pu pose o his deli e able in sho is o:
• Explain he di e en p oposed edge compu ing a chi ec u es o di e en use cases.
• Desc ibe he so wa e and ha dwa e componen s used in each p oposed a chi ec u e.
• Go h ough he de elopmen , es , and deploymen s a egies o wo main edge-d i en
ene gy managemen applica ions.
On behal o Au ho s
Mus apha Habib, KTH
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 9
2. In oduc ion
2.1. Scope
This deli e able summa izes he wo king p inciple o he edge compu ing solu ions p oposed by KTH,
ou lining he de elopmen , es , and commissioning s ages. In his con ex , wo main a chi ec u es
we e p oposed:
The indus ial panel a chi ec u e: whe e he solu ion is designed o upg ade he BMS unc ionali ies
when dealing wi h complex ha dwa e solu ions (e.g., so p ion s o age).
The i ual machine (VM)-based se e a chi ec u e: whe e he solu ion is designed o manage
locally he powe and hea lows in DH communi ies by le e aging In e ne o Things (IoT)
communica ion.
These wo main a chi ec u es will make i possible o hos algo i hms o da a-d i en modeling,
combined wi h me aheu is ic and de e minis ic op imiza ion app oaches. The i s unc ionali y is
dedica ed o explo ing he compo men o mode n HVAC ha dwa e solu ions o enable op imal
con ol, while he second unc ionali y capi alizes on ene gy lexibili y wi hin DH communi ies,
o e ing dis ibu ed con ol ia IoT communica ion o achie e op imal DSM.
2.2. Audience
This deli e able may p o ide use ul inpu s o hose who a e conce ned wi h he da a-d i en con ol
p ocedu e o he new indus ial HVAC sys ems and d i e hem wi hin a sec o coupling ne wo k.
2.3. Abb e ia ions
AI : A i iciel in elligence
API: applica ion p og amming in e ace
COP: Coe icien o pe o mance
DH: dis ic hea ing
DL: Deep lea ning
DNS: Domain name sys em
DSM: demand-side managemen
EMS: ene gy managemen sys em
H&C: Hea ing and cooling
HP: Hea pump
HTTP: Hype Tex T ans e P o ocol
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 16
• Limi ed AI/ML capabili ies: lack o ad anced machine lea ning and AI echnologies.
• Da a quali y and consis ency: challenges in ensu ing high-quali y, eliable da a.
Figu e 4 Typical a chi ec u e o building managemen sys em
In addi ion o he limi a ions lis ed abo e, da a ci cula ion in classical BMS a chi ec u es uns in a
closed-loop pa h, om gene a ion a he iled le el, passing o he collec ion and p ocessing in he
con ol le el, ending up a he managemen le el o EMS se ices and epo ing (see Figu e 5). The e
is no s age wi hin his da a jou ney whe e deep da a analysis o op imiza ion-based ene gy
managemen sys ems (EMSs) can ake place.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 17
Figu e 5 da a ci cula ion in classical a building managemen sys em
3.2.1.2. Cloud-based solu ion limi a ions
When he e is a need o managing he local ene gy esou ces a a building le el op imally wi h a
connec ion o DH, LV, and RES, upg ading he BMS capabili ies o adop op imiza ion-based con ol
can be he solu ion. In his con ex , cloud-based ene gy managemen sys ems (EMS) ha e ecen ly
become a p omising s a egy due o he unlimi ed da a compu ing and s o age esou ces (J.C.M. Siluk
e al., 2023, T. Ja ied e al., 2019 and F. Condon e al., 2023). Howe e , in any online EMS pla o m,
secu i y and da a p i acy conce ns a e always aised. Mo eo e , he dependency on he in e ne
connec ion o ensu ing con inuous ope a ion o EMS se ices is manda o y, which lowe s he
eliabili y o he whole managemen sys em.
3.2.2. Edge compu ing as a local op imiza ion-based EMS
To o e come he challenges aised p e iously, HYPERGRYD has come up wi h an inno a i e edge
compu ing solu ion ha upg ades he implemen ed EMS, ypically implemen ed in a building
managemen sys em (BMS), wi hou he need o a cloud connec ion. The solu ion a chi ec u e is
designed as an indus ial panel which acili a es i s deploymen and in eg a ion o BMS. In such
panels, da a enginee ing, communica ion, and compu ing equi emen s a e all handled locally.
The edge panel is designed o ha e bi-di ec ional da a lows wi h BMS wi h he con ol le el as Figu e
6 shows. This makes i possible o access a ious senso da a a one single poin , which is he
moni o ing p og ammable logic con olle (PLC) o he a ge ed HVAC sys em and ca ies ou con ol
unc ions h ough he same PLC on a sys em le el only.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 18
Figu e 6 edge-BMS connec ion poin
3.2.3. Edge panel – ha dwa e componen s
The edge panel ga he s a se o de ices ha a e indispensable o i s p esumed ole. These de ices
a e lis ed below wi h b ie explana ions o hei oles (see Figu e 7):
• PLC: which is he i s de ice ha communica es wi h he a ge ed BMS con olle since i
suppo s he mos common indus ial communica ions adop ed in buildings (e.g., Modbus,
BACne , e c.). In addi ion o he low-le el communica ion capabili y, he PLC can be coded o
inco po a e secu i y ins uc ions o deal wi h any “unexpec ed” o miss-communica ion wi h
he AI-enabled PC (see below), in such scena ios, he PLC akes he lead by eplacing he
ad anced EMS hos ed on he indus ial PC by a local basic one.
• Ga eway: his de ice is he s age o p ocessing da a coming om he PLC, and sending i o
any local o emo e da abase. I is also possible o inse online in o ma ion i needed (e.g.,
da a o dynamic ene gy cos ).
• Indus ial PC: wi h high compu ing capabili y, his Linux-based compac and anless PC is
esponsible o handling op imiza ion e o s associa ed wi h he ad anced EMS. I is also he
s age o hos ing a local da abase needed o aining ML and DL models.
• E he ne swi ch: needed o connec ing he de ice men ioned abo e all in one ne wo k along
wi h he BMS ne wo k.
• P o ec ion and powe supply: needed o p o iding di e en ol age le els o each de ice
and p o iding p o ec ion agains o e loading and sho ci cui s.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 19
Figu e 7 ha dwa e componen s o he edge panel
3.2.4. Edge panel – so wa e componen s
In he p oposed edge solu ion, he da a jou ney om he BMS con olle o he edge da abase goes
h ough a se o so wa e s ages ha ensu e da a in e ope abili y, p ocessing, and communica ion.
Figu e 8 gi es an illus a i e ep esen a ion o he selec ed so wa e solu ion o his pu pose. In
con as o da a p ocessing and s o age, whe e open-sou ce solu ions we e adop ed, he i s s age
o communica ion wi h BMS has o be wi h comme cial so wa e ha suppo s he IEC 61131-3
s anda d o PLC p og amming.
Figu e 8 so wa e componen s o he edge panel
• S7-1200 i mwa e: as an ope a ion sys em o SIMATIC S7-1200 o SIEMENS PLC, i makes i
possible o communica e wi h he BMS con olle hanks o he suppo ed indus ial
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 20
communica ion p o ocols. This i mwa e ensu es also he execu ion o he logic coded and
downloaded om Tia Po al so wa e, his logic mainly inco po a es he da a eques
equency, in e locks, ala ming, and op ionally a local ule-based EMS. Communica ion wi h
a highe managemen le el ia open pla o m communica ion (OPC) is also included.
• Node-RED: as an open-sou ce web pla o m, his so wa e solu ion is a powe ul ool o
handling da a communica ion, p ocessing, and con e sion. I ecei es da a om he PLC ia
an OPC clien , p ocesses i o handle Modbus limi a ions when i comes o cons uc ing loa
alues (32-bi based), and communica es wi h emo e da abases wi h w i e/ ead unc ions
using REST ul API. Some unc ionali ies had o be w i en na i ely in Ja aSc ip as a high-
p og amming language.
• In luxDB: The ime se ies da abase suppo s open-sou ce e sions o speci ic applica ions. I
ea u es an e icien mechanism o managing ime se ies da a and o e s high scalabili y,
making i a obus da a s o age solu ion. Designed o mul i-solu ion communica ion, i
p o ides in eg a ion h ough HTTP API, In luxQL, and Py hon clien s.
• Py hon: ins alled on a Linux ope a ion sys em, i ep esen s he main edge backend whe e
all da a-d i en modelling and op imiza ion (ad anced EMS) we e w i en wi h (no shown in
Figu e 8). Thanks o he a ailable open-sou ce lib a ies, Py hon handles communica ion wi h
he da abase o ML/DL model aining/upda ing, and employing di e en op imiza ion
echniques.
4. De elopmen and deploymen s a egies
In his sec ion, we co e de ails abou he KTH edge solu ions, om he de elopmen phase o he
ins alla ion and commissioning on he chosen p ojec pilo s. Due o he p o ound di e ence be ween
he applica ions and he speci ica ions o he a ge ed pilo s, he deploymen p ocedu e wi hin
HYPERGRYD o hese wo designed solu ions is undamen ally di e en .
4.1. Edge solu ion o DH communi ies managemen
Due o he di icul y o engaging he end-use s o he DH ene gy communi y, managed by Sonnenpla z
pilo , in he de eloped coo dina ed DSM, i was un o una ely no possible o deploy he solu ion
locally in G öschenau municipali y in Aus ia. The e o e, applying coo dina ed DSM was e alua ed on
he simula ion phase only, which is ex ensi ely explained and p esen ed in D3.3. Howe e , se ing up
he DSM-se e wi h all da a enginee ing equi emen s was comple ely commissioned a he KTH
campus wi h emo e communica ion o he me e s and senso s in he Sonnenpla z communi y.
4.1.1. Da a s eaming o ma om he ene gy me e s
When he DSM-se e subsc ibes o he MQTT b oke wi h he igh c eden ials using Py hon o any
hi d-pa y so wa e (e.g., Node-RED), da a s eaming s a s lowing in eal- ime om he ene gy
me e s o he edge da abase. The da a o ma in his case is an objec -based o ma whe e he
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 21
measu ed alues a e sen all oge he wi h some me ada a in o ma ion (e.g., senso ID, uni ,
measu emen s a ion ID).
Figu e 9 Da a o ma being sen by he MQTT b oke
4.1.2. Da a se ices om he DSM-se e
Figu e 10 shows how hese se poin s a e being sen o he a ge ed HPs using Ja aSc ip Objec
O ien ed (JSON) o ma . This JSON-based in o ma ion package can be also published o he MQTT
b oke o which he a ge ed HPs a e subsc ibing ( ypically employing a buil -in communica ion
ea u e o using addi ional ha dwa e). The JSON package con ains he a ge ed HP IDs, he condense
empe a u e se poin o simply he ON/OFF command, and he applicable imes amp.
Along wi h he con ol unc ion, Figu e 10 shows addi ional unc ionali ies ha a e accessible by he
DSM-se e hanks o he implemen ed local ime se ies da abase. Using web dashboa ds, DH da a
can be made public and accessible by all in ol ed s akeholde s o apid and easy pe o mance
e alua ion. Addi ionally, hi d-pa y comme cial so wa e (e.g., he digi al win pla o m om IDP and
he simula ion ool om ENCOORD) can ake ad an age o hese eal- ime da a and eed i o hei
so wa e backends. The e o e, we see in Figu e 10 ha is possible o que y da a using simple Linux
commands using HTTPS APIs o classical SQL que ying.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 22
Figu e 10 DSM-se e accessible se ices
Deploying he p oposed edge-based coo dina ed DSM in o eal DHN-powe ed communi ies may lead
o one common challenge which is he willingness o he end-use s o gi e access o hei HVAC
sys ems (pa icula ly HPs) o emo e con ol. The e is always he case whe e his is no possible due
o p i acy conce ns. To o e come his, he s udy comes up wi h an al e na i e solu ion: he end-use s
who ag eed o be in ol ed in his coo dina ed DSM will ecei e, on a daily basis, an op imal se poin
p o ile o implemen , which is applicable o he nex 24-hou pe iod. Since mos o he ecen HP
echnologies a e p og ammable, he end-use who ecei ed he se poin p o ile one day ahead
( ypically ia a sma de ice) will jus p og am he HP o ope a e acco dingly. Figu e 11 shows bo h
p oposed DSM con ol mechanisms.
Figu e 11 DSM con ol execu ion mechanism: on he le -hand side, di ec con ol ia API o IoT ( eal- ime), on he igh -hand
side, manual con ol (p edic i e)
I he end-use is willing o expose his HP o ex e nal con ol emo ely, howe e , he HP echnology
migh no suppo emo e communica ion, in his case, he e is a solu ion ha consis s o inse ing
addi ional de ices in o he HVAC con ol and moni o ing panel, exis ing al eady, in he
household/building (see Figu e 12). Such ex a in es men s will be ei he cha ged o he end-use
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 23
himsel i he is sa is ied wi h he p esen ed payback pe iod, o o he DH o LV g id ope a o s since
he coo dina ed DSM is, indeed, designed o lowe he s ess on bo h ene gy ne wo ks elec ici y and
DH. Howe e , he s udy ou lined in his deli e able is solely echnical and does no add ess he
implemen a ion s a egy wi hin eal communi ies o he ole o he s akeholde s in ol ed in his
coo dina ed managemen ..
Figu e 12 Addi ional compu ing ha dwa e o ensu ing eal- ime emo e con ol
4.1.3. Da a exploi a ion in HYPERGRYD
The DSM o he Sonnenpla z DH ene gy communi y has been alida ed in a simula ion s udy, whe e
his o ical ene gy me e da a, s o ed in he edge da abase, has been used o his pu pose. The goal
is o simula e i possible o each he ene ge ic independence o he conce ned buildings (he e, ou
buildings we e picked up). The me hod is based on op imally sha ing he a ailable pho o ol aic (PV)
and he mal ene gy s o age (TES) o gi e highe lexibili y o he HP uni s o wo k in op imal
coo dina ion and independen ly om he g id u ili y. The HP empe a u e se poin s, as well as he
s a e o cha ge (SOC) e olu ion o each TES, a e displayed in Figu e 13. Mo e in o ma ion abou his
s udy was de ailed in D3.3.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 24
Figu e 13 DSM simula ion (communi y-cen ic), op: HP op imal condense empe a u e se poin , bo om: TES SOC
a ia ion (bo om igu e) in a 2-day simula ion)
In he amewo k o HYPERGRYD, he de eloped DSM-se e has also o e ed an e icien da a
enginee ing solu ion o he p ojec pa ne s who a e in e es ed in Sonnenpla z communi y da a. Fo
his pu pose, KTH c ea ed speci ic au hen ica ion c eden ials o each pa ne so hey can access he
se e da abase emo ely ia HTTP APIs o ge eal- ime and his o ical da a.
Figu e 14 Di e en pu poses o connec ing o he DSM-se e da abase in he amewo k o HYPERGRYD
This access enabled he in eg a ion o senso da a om he Sonnenpla z communi y in o he p ojec
pla o m, speci ically wi hin he digi al win so wa e (see Figu e 15). This app oach p o ides public
isualiza ion, no only o he s udied DH communi y bu also o anyone in e es ed in analyzing
ene gy consump ion pa e ns, whe he elec ical o he mal, in a ypical sec o -coupling sys em.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 25
Figu e 15 Displaying senso da a o each building in Sonnenpla z communi y ia he p ojec pla o m
4.2. Edge o BMS upg ading
Con a y o he i s solu ion, he ins alla ion and commissioning o edge compu ing o a sys em o
he building le el was a s aigh o wa d mission. The a ge ed pilo is he KEZO esea ch cen e in
Wa saw, Poland, o op imize hea and elec ici y consump ion wi hin an emula ed locally powe ed
DH sys em. This is h ough employing ad anced da a-d i en con ol o inno a i e ha dwa e solu ions,
speci ically he CO2 indus ial HP and he so p ion s o age sys em.
4.2.1. De elopmen phase a KTH
Following a signi ican delay due o ha dwa e deli e y, which exceeded one yea , he assembly o he
panel de ices, including he equi ed coding and con igu a ions, ook app oxima ely wo mon hs
be o e being shipped o he pilo si e. Fo local es ing, he e was a need o employ Modbus se e
simula o s o oubleshoo ing Modbus communica ion.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 32
Figu e 24 Simple ep esen a ion o he ein o cemen lea ning echnique o he hea pump con ol
Ano he key objec i e o he edge panel is o analyze he ope a ional da a o he so p ion s o age,
o be ins alled a he same pilo as one o he inno a i e ha dwa e p oposed in HYPERGRYD, aiming
o unlock i s po en ial o op imal con ol. The app oach in ol es implemen ing a simple ule-based
EMS on he edge panel's PLC o ope a e he s o age in a ious modes and e alua e i s pe o mance
wi hin he building's o e all H&C ci cui s. This p ocess will gene a e da a ha can be used o de elop
DL models o he so p ion s o age, suppo ed by a ma hema ical ep esen a ion o i s physical
p ope ies, known as a physics-in o med neu al ne wo k (PINN). This, in u n, will pa e he way o
implemen ing ad anced op imal con ol s a egies, such as nonlinea model p edic i e con ol
(NMPC). Figu e 25 shows he so p ion s o age echnology, du ing he design phase.
Figu e 25 So p ion s o age du ing he design phase
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 33
Conclusions
One o he miles ones ha HYPERGRYD is aiming o achie e is o le e age edge compu ing wi h he
In e ne o Things po en ials o sec o coupling, demons a ing hei po en ial o e olu ionize
ene gy managemen in buildings. These solu ions, designed o op imize ene gy consump ion and
enhance building managemen sys ems, adop ad anced echnologies like machine lea ning, deep
lea ning, and op imiza ion algo i hms.
Key indings and achie emen s include:
• Success ul deploymen o edge solu ions in eal-wo ld ope a ion condi ions o e ed by he
p ojec pilo s, ei he comple ely as in he case o KEZO li e-in lab o pa ially as in he case o
Sonnenpla z.
• De ining he implemen a ion amewo ks and ul illing he da a enginee ing equi emen s
needed o da a-d i en con ol s a egies using machine lea ning and deep lea ning
echniques o op imizing hea pump ope a ions and so p ion s o age sys ems.
• In eg a ion wi h building managemen sys ems o enable mo e e icien and e ec i e ene gy
managemen .
• O e coming echnical challenges ela ed o ha dwa e, so wa e, and communica ion
p o ocols du ing deploymen .
Rela ion o con inued de elopmen s
As edge compu ing solu ions a e designed o le e age da a-d i en app oaches, his concep means
ha he e should be a es phase in he p ojec pilo s whe e he main goal is jus o collec , p ocess,
s o e da a, and build p edic i e models and op imiza ions on op o i . While wai ing o his p ocess
o comple e, KTH has de ined some po en ial ac i i ies o achie e:
• Assess and quan i y he KEZO building he mal mass and in eg a e his in o ma ion in o he
edge con ol amewo k. This has he po en ial o gi e mo e lexibili y o he op imiza ion
app oaches o deal wi h he ene gy p ice and wea he condi ion luc ua ions.
• Fu he de elop and e ine con ol s a egies, pa icula ly o so p ion s o age sys ems,
u ilizing ad anced echniques, pa icula ly, nonlinea model p edic i e con ol and
ein o cemen lea ning.
• Conce ning he coo dina ed edge-d i en demand-side managemen , expanding he
deploymen o a la ge numbe o buildings and communi ies o alida e he scalabili y and
gene alizabili y o he solu ions is a ge ed.
Add ess po en ial challenges ela ed o da a p i acy, secu i y, and use accep ance, especially o
esiden ial communi ies.
D3.4 Me hodologies and amewo ks o op imize demand esponse and peak load sha ing o 4 h-5 h DHC in
h ee ypical Eu opean clima es 34
Re e ences
J.C.M. Siluk , P.S. de Ca alho , V. Thomasi , C.A. de O. Pappis, J.L. Schae e . Cloud-based ene gy
managemen sys ems: Te minologies, concep s and de ini ions. Ene gy Resea ch & Social Science 106
(2023) 10331.
T. Ja ied e al. / IFAC Pape sOnLine 52-10 (2019) 171–175.
F. Condon e al. Design and Implemen a ion o a Cloud-IoT-Based Home Ene gy Managemen
Sys em. Senso s 2023, 23, 176.
