elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Impressum | Datenschutz | Kontakt | English
Schriftgröße: [-] Text [+]

Evaluation of the Utility of Remote Sensing Data for the Provision of Model Parameters for Energy System Modelling

Sternberg, Torben (2021) Evaluation of the Utility of Remote Sensing Data for the Provision of Model Parameters for Energy System Modelling. Masterarbeit, Otto-von-Guericke University Magdeburg.

[img] PDF
58MB

Kurzfassung

Due to an increase of extreme weather events, climate change is steadily moving to the center of public discourse. The consensus on the global challenges posed by a further increase in global warming is also reflected in political decisions such as the Paris Climate Agreement. However, the transition to an energy supply based on renewable energies is opposed by a globally increasing energy demand, the security of supply, as well as economic interests. In this area of tension, energy system modelling seeks solutions that lead to both, climate-neutral and cost-effective energy systems by evaluating modeled scenarios. However, the decentralized and fluctuating character of renewable energies places high demands on the temporal and spatial resolution of these models, which cannot be ensured with current databases. This work therefore explores the possibilities of an automated database generation through a combined approach of remote sensing data and machine learning by the example of coal-fired power plants and wind turbines. To determine the relevant modeling parameters, state-of-the-art models are compared and current databases analyzed. Since essential training data for the development of neural networks tailored to the chosen power generators is not available this work develops a framework that allows to generate training data from ger-referenced databases and thus enables different combinations of energy infrastructures and satellite data. Moreover, the framework provides a toolbox to integrate different neural network structures. On this basis, a converted classification and detection network is trained and evaluated using coal-fired power plants and wind turbines. The results show that detection of wind turbines is possible even on image resolutions of 10 meters. The trained neural networks are able to detect the majority of existing wind turbines on area-wide remote sensing images. With the presented approach it is possible to generate global datasets for power system analysis with a high degree of automation based on satellite data.

elib-URL des Eintrags:https://elib.dlr.de/147261/
Dokumentart:Hochschulschrift (Masterarbeit)
Titel:Evaluation of the Utility of Remote Sensing Data for the Provision of Model Parameters for Energy System Modelling
Autoren:
AutorenInstitution oder E-Mail-AdresseAutoren-ORCID-iDORCID Put Code
Sternberg, Torbensternberg.torben (at) googlemail.comNICHT SPEZIFIZIERTNICHT SPEZIFIZIERT
Datum:Oktober 2021
Referierte Publikation:Nein
Open Access:Ja
Seitenanzahl:120
Status:veröffentlicht
Stichwörter:Earth observation, deep learning, energy systems analysis, Sentinel-II, wind turbines, coal-fired power plants
Institution:Otto-von-Guericke University Magdeburg
Abteilung:Institute of Fluid Dynamics and Thermodynamics
HGF - Forschungsbereich:Energie
HGF - Programm:Energiesystemdesign
HGF - Programmthema:Energiesystemtransformation
DLR - Schwerpunkt:Energie
DLR - Forschungsgebiet:E SY - Energiesystemtechnologie und -analyse
DLR - Teilgebiet (Projekt, Vorhaben):E - Systemanalyse und Technologiebewertung, R - Optische Fernerkundung
Standort: Stuttgart
Institute & Einrichtungen:Institut für Vernetzte Energiesysteme > Energiesystemanalyse, ST
Deutsches Fernerkundungsdatenzentrum > Dynamik der Landoberfläche
Hinterlegt von: Cao, Dr.-Ing. Karl-Kien
Hinterlegt am:15 Dez 2021 18:19
Letzte Änderung:15 Dez 2021 18:19

Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags

Blättern
Suchen
Hilfe & Kontakt
Informationen
electronic library verwendet EPrints 3.3.12
Gestaltung Webseite und Datenbank: Copyright © Deutsches Zentrum für Luft- und Raumfahrt (DLR). Alle Rechte vorbehalten.