| Abstract | Zgrade danas postaju kompleksni tehnički sustavi s velikim potencijalom za uštede
energije automatizacijom, umrežavanjem i koordinacijom različitih podsustava.
Automatizacija u zgradama stvara nove mogućnosti nadzora rada sustava koje su ključne
za informiranje korisnika ili uočavanje nepravilnosti u radu.
Neočekivano ponašanje korisnika u zgradama jedan je od ključnih faktora u stvaranju
nesklada između modela potrošnje energije i stvarne manifestacije. Ponašanje kao takvo
nije nužno moguće zaustaviti, ali je svakako poželjno podići razinu svijesti o problemu te
kvantificirati njegove stvarne posljedice. Jednom dobiven, model omogućuje lakše
predviđanje, praćenje i interveniranje u različitim aspektima energetske potrošnje. U
konačnici, ove komponente pružaju temelje za energetsku učinkovitost zgrade.
U ovom radu, model klasifikacije otvorenosti prozora dobiven je upotrebom heuristički
vođenog klasificiranja podataka, parametrizacijom te u konačnici strojnim učenjem.
Tok formalizacije obrade podataka (za izlučivanje heuristika) bio je sljedeći:
1) Definiranje sezonskog konteksta
2) Promatranje sezonskog konteksta i opaske
3) Zoniranje na vremenski kontekst
4) Analiza čestih uzoraka
5) Parametrizacija
6) Ovjera rezultata
Znanje dobiveno vizualnom inspekcijom pretočeno je u parametrizirani oblik, pogodan za
inspekciju pomoću algoritma. Parametri dobiveni na taj način služili su kao parametri
računalnog algoritma za označavanje. Jednom označeni podaci bili su pripremljeni za
algoritam strojnog učenja.
Kao algoritam strojnog učenja, odabran je algoritam slučajnih šuma (eng. Random Forest).
Ideja tog algoritma je sljedeća: korištenjem skupa slabih klasifikatora dobivamo snažan
klasifikator. Model tog algoritma sačinjen je od stabala odluka. Stablo odluke je
konceptualna struktura u kojoj se na temelju značajki podataka gradi stablo po kojem se
podaci kreću pri njihovoj klasifikaciji.
Potpuni skup podataka za algoritam bio je :
1) dan godine
2) dan tjedna (radni/neradni)
3) sat dana
4) unutrašnja temperatura prostorije
5) vanjska temperatura
6) brzina ventilokonvektora
7) relativna vlaga prostorije
8) direktna sunčeva dozračenost
9) difuzna sunčeva dozračenost
+ 10) presuda otvorenosti prozora
Konačni ostvareni model ima sljedeće efikasnosti:
1) 1.0 za sezonski kontekst grijanja
2) 0.79 za sezonski kontekst hlađenja
3) 0.99 za prijelazni sezonski kontekst
Prosječna efikasnost modela je: 0.92.
Iako taj rezultat na prvi pogled izgleda vrlo obećavajuće, u obzir valja uzeti ključne
aspekte modela. Model je treniran na podacima čije je znanje dobiveno heurističkim
algoritmom. To znanje nije u potpunom skladu sa stvarnim činjenicama.
Model kao takav u 92% slučajeva ispravno odgovara na pitanje je li prozor otvoren, ali ne
smije se zaboraviti da definiciju otvorenosti prozora (iz perspektive modela) ne definira
stvarna otvorenost prozora nego heuristički dobivena oznaka otvorenosti.
Dakle, model nema praktičnu efikasnost od 92%. To je jedna od mana rada s nepotpunim
skupom podataka. Kada bi od početka rada bili poznati i podaci o tome je li prozor otvoren
ili nije, cijela priprema podataka bila bi redundantna, a završna efikasnost modela bila bi u
praktičnom smislu potpuno ispravna.
Vizualizacija podataka koji se kontinuirano prikupljaju može dati bolji uvid u ponašanje
zgrade te potrošnju u pojedinim prostorijama. Ispravnom vizualizacijom velike količine
podataka moguće je vidjeti određene anomalije i odstupanja od očekivanih vrijednosti.
Uvid u takve anomalije pomaže nam da ih spriječimo te dodatno energetski optimiramo
promatranu zgradu. Iz tog razloga izrađena je web aplikacija FER Energy Expenditure
Visualization (FER EEV) kako bi se ovaj rad zaokružio i upotpunio.
|
| Abstract (english) | Today, buildings are becoming complex technical systems with great potential for energy
savings through automation, networking and coordination of different subsystems.
Implementing automation yields new possibilities for monitoring the operation of the
system, which is crucial for informing users or detecting irregularities.
Unexpected user behavior in buildings is one of the key factors in creating a mismatch
between the energy consumption model and the actual manifestation. Behavior as such is
not necessarily possible to stop, but it is certainly desirable to raise the level of awareness
about the problem and quantify its real consequences. Once obtained, the model makes it
easier to predict, monitor and intervene in various aspects of energy consumption.
Ultimately, these components provide the foundation for a building’s energy efficiency.
In this paper, the window openness classification model was obtained using heuristically
guided data classification, parameterization, and ultimately machine learning.
The course of formalization of data processing (for extracting heuristics) was as follows:
1) Defining the seasonal context
2) Observation of seasonal context and remarks
3) Zoning to a time context
4) Analysis of frequent samples
5) Parameterization
6) Verification of the results
The knowledge obtained by visual inspection is translated into a parameterized form,
suitable for inspection using an algorithm. The parameters obtained in this way served as
parameters of the computational notation algorithm. Once tagged, the data was prepared
for a machine learning algorithm.
As a machine learning algorithm, the Random Forest algorithm was chosen. The idea of
this algorithm is as follows: by using a set of weak classifiers we get a strong classifier.
The model of this algorithm is made up of decision trees. The decision tree is a conceptual
structure in which, based on the characteristics of the data, a logical tree is built according
to which the data moves during its classification.
The complete data set for the algorithm was:
1) day of the year
2) day of the week (working / non-working)
3) hour of the day
4) internal room temperature
5) outdoor temperature
6) fan coil speed
7) relative humidity of the room
8) direct sunlight
9) diffuse solar radiation
+ 10) label of window openness
The final model has the following efficiencies:
1) 1.0 for seasonal heating context
2) 0.79 for the seasonal cooling context
3) 0.99 for a transitional seasonal context
The average efficiency of the model is: 0.92.
Although this result seems very promising at first glance, key aspects of the model should
be considered. The model is trained on data whose knowledge is obtained by a heuristic
algorithm. That knowledge is not in complete agreement with the real facts.
The model as such in 92% of cases correctly answers the question of whether the window
is open, but it must not be forgotten that the definition of window openness (from the
perspective of the model) is not defined by the actual window opening but by a
heuristically derived openness label.
Thus, the model has no practical efficiency of 92%. This is one of the disadvantages of
working with an incomplete data set. If the data on whether the window was open or not
were known from the beginning of the work, the whole preparation of the data would be
redundant, and the final efficiency of the model would be completely correct in practical
terms.
Visualization of data that is continuously collected can give a better insight into the
behavior of the building and consumption in individual rooms. By correctly visualizing a
large amount of data, it is possible to see certain anomalies and deviations from the
expected values. Insight into such anomalies helps us to prevent them and additionally
energetically optimize the observed building. For this reason, the web application FER
Energy Expenditure Visualization (FER EEV) was created to complete this
work. |
