magistrsko delo magistrskega študijskega programa II. stopnje Strojništvo
Anže Murko (Author), Primož Podržaj (Mentor)

Abstract

Proizvodnja električne energije sončnih elektrarn je zaznamovana z veliko spremenljivostjo, kar predstavlja izziv pri upravljanju energetskih omrežij. Z namenom izboljšanja napovednih rezultatov proizvodnje energije sončnih elektrarn je bil razvit nov pristop. Slednji pri napovedovanju proizvodnje centralne elektrarne uporablja različno število vključenih sosednjih elektran v napoved. Odvisnost napovednih rezultatov modelov je bolj odvisna od števila vključenih sosednjih elektrarn kot pa od same topologije mreže. Izvedli smo optimizacijo hiperparametrov modelov in napovedne rezultate primerjali z obstoječimi raziskavami. Ugotovili smo, da neuporaba meteoroloških podatkov rezultira v slabših napovednih rezultatih.

Keywords

magistrske naloge;električna energija;sončne elektrarne;napovedovanje;strojno učenje;električno omrežje;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UL FS - Faculty of Mechanical Engineering
Publisher: [A. Murko]
UDC: 502.21:523.9:620.9(043.2)
COBISS: 167625219 Link will open in a new window
Views: 44
Downloads: 9
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Forecasting the electricity generation of grid-connected solar power plants
Secondary abstract: Electricity production from solar power plants is characterised by volatility, which presents a challenge in the management of electrical grids. A new approach was developed in order to improve the predictive results of the energy production of solar power plants. The latter uses a different number of included neighbouring power plants in the energy forecast of the central power plant. The dependence of the predictive results depends more on the number of neighbouring power plants included, than on the topology of the network. We performed the optimisation of the model's hyperparameters and compared the predictive results with the existing research. We found out, that not using meteorological data results in worse forecasting results.
Secondary keywords: master thesis;electricity;solar power plants;forecasting;machine learning;electrical grid;
Type (COBISS): Master's thesis/paper
Study programme: 0
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. Ljubljana, Fak. za strojništvo
Pages: XXII, 61 str.
ID: 19872171