Τεχνικές μηχανικής και βαθιάς μάθησης για βραχυπρόθεσμη 
πρόβλεψη ζήτησης ηλεκτρικής ενέργειας:  
μια συγκριτική ανάλυση σε χρονοσειρές Ευρωπαϊκών 
διαχειριστών συστημάτων μεταφοράς

Σεισόπουλος, Ιωάννης-Κωνσταντίνος; Seisopoulos, Ioannis-Konstantinos

dc.contributor.author	Σεισόπουλος, Ιωάννης-Κωνσταντίνος	el
dc.contributor.author	Seisopoulos, Ioannis-Konstantinos	en
dc.date.accessioned	2022-11-14T08:49:25Z
dc.date.available	2022-11-14T08:49:25Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/56122
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.23820
dc.rights	Default License
dc.subject	Μηχανική Μάθηση (ML)	el
dc.subject	Βαθιά Μάθηση (DL)	el
dc.subject	Πρόβλεψη Χρονοσειρών	el
dc.subject	Έξυπνα Δίκτυα	el
dc.subject	Πρόβλεψη Φορτίου Βραχυπρόθεσμου Χρόνου (STLF)	el
dc.subject	N-Beats	en
dc.subject	LightGBM	en
dc.subject	Χρονικά Συνελικτικά Δίκτυα (TCN)	el
dc.subject	Νευρωνικά Δίκτυα	el
dc.subject	Ensemble	en
dc.title	Τεχνικές μηχανικής και βαθιάς μάθησης για βραχυπρόθεσμη πρόβλεψη ζήτησης ηλεκτρικής ενέργειας: μια συγκριτική ανάλυση σε χρονοσειρές Ευρωπαϊκών διαχειριστών συστημάτων μεταφοράς	el
heal.type	bachelorThesis
heal.classification	Νευρωνικά Δίκτυα	el
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-07-09
heal.abstract	Electrical short-term load forecasting (STLF) has emerged as one of the most essential fields of research for efficient and reliable operation of power systems in the last few decades. It plays a very significant role in the field of scheduling, contingency analysis, load flow analysis, planning and maintenance of power systems, let alone in the participation of energy companies in the energy markets. Especially with the advent of smart grids, the need for fairly precise and highly reliable estimation of electricity load is greater than ever. Moreover, the aforementioned task (STLF) has become nowadays more popular for European countries as the energy crisis has reached an unprecedented peak with geopolitical extensions. Machine learning methods are well adapted to the nature of the electrical load, as they can model complicated nonlinear connections in Time Series Forecasting through a learning process containing historical data patterns. The presented Thesis conducts a comparative study of state-of-the-art Machine Learning (ML) and Deep Learning (DL) techniques, namely Light Gradient Boosting Machine (LightGBM), Neural Basis Expansion Analysis for Time Series forecasting (N-Beats) and Temporal Convolutional Networks (TCN), while investigating the effect of various external variables in the process. Four categories of features, historical loads, temporal covariates, energy price and weather factors, are considered and utilized leveraging various encoding mechanisms. In the experimental studies, hourly-resolution load datasets from Portugal, Spain and Greece and the corresponding prices of load records, as well as the provincial weather data from Portugal are utilized, trying to effectively forecast the day-ahead load. Finally, a hybrid model is proposed, combining Machine (ML) and Deep Learning (DL) models, applying an Ensemble method.	en
heal.advisorName	Ψαρράς, Ιωάννης	el
heal.committeeMemberName	Ψαρράς, Ιωάννης	el
heal.committeeMemberName	Ασκούνης, Δημήτριος	el
heal.committeeMemberName	Δούκας, Χρυσόστομος (Χάρης)	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών. Τομέας Ηλεκτρικών Βιομηχανικών Διατάξεων και Συστημάτων Αποφάσεων	el
heal.academicPublisherID	ntua
heal.numberOfPages	71 σ.	el
heal.fullTextAvailability	false