| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font
First page > Show document

Show document Help

Title:Napovedovalni model proizvodnje energije v digitalnem dvojčku sončne elektrarne
Authors:ID Zajc, Lea (Author)
ID Bokal, Drago (Mentor) More about this mentor... New window
ID Pihler, Jože (Comentor)
Files:.pdf MAG_Zajc_Lea_2018.pdf (1,80 MB)
MD5: 4566361E4F75C9A664AA5BD9DED46330
PID: 20.500.12556/dkum/f1f9efc1-a6e8-41b0-8cfc-c7b54928b8ad
 
Language:Slovenian
Work type:Master's thesis/paper
Typology:2.09 - Master's Thesis
Organization:FNM - Faculty of Natural Sciences and Mathematics
Abstract:Magistrsko delo obravnava koncept digitalnega dvojčka in predstavi implementacijo digitalnega dvojčka sončne elektrarne s tehnologijo napovedovanja proizvodnje električne energije. Implementiranih je pet metod napovedovanja s poudarkom na metodi naključnih gozdov. Delo je razdeljeno v šest poglavij. Prvo poglavje definira koncept digitalnega dvojčka in opiše njegove glavne procese. Navedejo se namen in primeri uporabe. Drugo poglavje predstavi napovedovanje proizvodnje sončne elektrarne s fotonapetostnimi celicami. Opisan je proces pretvorbe sončne energije v električno. Našteti in opisani so vplivi na proizvodnjo sončne elektrarne, definirani sta atmosferska in električna formula za izračun električne energije v nekem času t. Opisane so metode za napovedovanje proizvodnje električne energije: posplošena linearna regresija, metoda podpornih vektorjev in avtoregresija. Definirana je formula za napoved proizvodnje električne energije z vključenimi lastnostmi sončne elektrarne in nekaterimi vremenskimi podatki, kar predstavlja fizikalni model. V tretjem poglavju je podrobneje predstavljeno delovanje naključnih gozdov in primeri uporabe, kjer se na začetku omeni odločitvena drevesa, s pomočjo katerih je izpeljana definicija naključnih gozdov. Četrto poglavje predstavi taksonomijo problemov in metod napovedovanja elektroenergetskih količin, v tabelo so dodane nove taksonomske enote, primerjave med tehnikami napovedovanja so uvrščene v izdelano taksonomijo, razširi se graf primerjav. Peto poglavje prikaže implementacijo digitalnega dvojčka sončne elektrarne FERI, predstavijo se vhodni parametri senzorjev, ostali vhodni podatki ter podatki, ki bi jih bilo smiselno pridobiti in vključiti. Opisana je integracija digitalnega dvojčka s sončno elektrarno. V analitiki digitalnega dvojčka je, s pomočjo zgledov, predstavljena uporaba metod napovedovanja. V zadnjem poglavju so definirani pogreški za primerjavo modelov, sledijo razlage rezultatov posameznih modelov polurne napovedi. Na koncu so modeli med sabo primerjani glede na novo definiran pogrešek E-MAPE, ki je konsistenten s standardno uporabljenim pogreškom M AE, razen pri primerjavi rezultatov avtoregresije in posplošene linearne regresije.
Keywords:digitalni dvojček, sončna elektrarna, napovedovanje, naključni gozd, podporni vektorji, posplošena linearna regresija, avtoregresija, fizikalni model, taksonomija
Place of publishing:Maribor
Publisher:[L. Zajc]
Year of publishing:2018
PID:20.500.12556/DKUM-71981 New window
UDC:519.172.1:621.311.243(043.2)
COBISS.SI-ID:24056840 New window
NUK URN:URN:SI:UM:DK:U0VWSSPE
Publication date in DKUM:04.10.2018
Views:1430
Downloads:236
Metadata:XML DC-XML DC-RDF
Categories:FNM
:
Copy citation
  
Average score:(0 votes)
Your score:Voting is allowed only for logged in users.
Share:Bookmark and Share



Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
Link:http://creativecommons.org/licenses/by-nc-nd/4.0/
Description:The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.
Licensing start date:05.09.2018

Secondary language

Language:English
Title:Energy forecasting model of energy generation in a digital twin of a solar power plant
Abstract:This master thesis examines the concept of a digital twin and demonstrates the implementation of a digital model of a solar power plant with the digital twin technology and intention of forecasting the production of electrical energy. Implemented are five methods of forecasting with the emphasis on the random forest method. The work is divided into six chapters. The first chapter defines the concept of a digital twin and explains its primary functions. The purpose and uses of digital twin are presented. The second chapter presents the forecasting of the production of the solar power plant. The processing of solar energy into electrical energy is described. The effects on the production of solar power plants are described, atmospheric and electrical formulas for calculating electrical energy are defined. The methods for forecasting and production of electrical energy are presented: generalized linear regression, support vector machine and autoregression. The formula for forecasting the production of electrical energy with the incorporated properties of the solar power plant and some weather data is defined, which follows the physical model. The third chapter presents in detail the methodology of random forests and states examples of use where, at the beginning, the decision trees are mentioned, with the help of which the definition of random forests is given. The fourth chapter presents the taxonomy of problems and methods of forecasting the electric energy, new taxonomic units are added to the taxonomic table, comparisons between forecasting techniques are included in the generated taxonomy, the graph of comparisons is expanded. Chapter five shows the implementation of the digital twin of the FERI solar power plant, the input parameters of the sensors and the other relevant input data are introduced. The integration of the digital twin with the solar power plant is described. With the use of examples the use of methods of forecasting in the digital twin analytics is defined. The last chapter defines errors for the comparison of models, followed by the interpretation of the results of individual models of half-yearly forecasts. Finally, the models are compared to each other according to the newly defined E-MAPE error that is consistent with the standard M AE error, except when comparing results of autoregression and generalized linear model.
Keywords:digital twin, solar power plant, forecasting, random forest, support vector machine, generalized linear model, autoregression, physical model, taxonomy


Comments

Leave comment

You must log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica