Εφαρμογή τεχνικών συσταδοποίησης σε δεδομένα έξυπνων μετρητών για τον σχεδιασμό προγράμματος απόκρισης ζήτησης

Πιπεργιάς, Άγγελος; Pipergias, Angelos

dc.contributor.author	Πιπεργιάς, Άγγελος	el
dc.contributor.author	Pipergias, Angelos	en
dc.date.accessioned	2024-09-24T07:57:14Z
dc.date.available	2024-09-24T07:57:14Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/60258
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.27954
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Ααπόκριση ζήτησης	el
dc.subject	Έξυπνα δίκτυα	el
dc.subject	Συσταδοποίηση	el
dc.subject	Έξυπνοι μετρητές	el
dc.subject	Ελαστικότητα	el
dc.subject	Smart grid	en
dc.subject	Clustering	en
dc.subject	Demand response	en
dc.subject	Smart meters	en
dc.subject	Flexibility	en
dc.title	Εφαρμογή τεχνικών συσταδοποίησης σε δεδομένα έξυπνων μετρητών για τον σχεδιασμό προγράμματος απόκρισης ζήτησης	el
dc.contributor.department	Decisions Support Systems Laboratory	el
heal.type	bachelorThesis
heal.classification	Ενέργεια	el
heal.classification	Τεχνητή Νοημοσύνη	el
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2023-03-03
heal.abstract	Recently, there’s been growing interest in demand response (DR) as a tool from the management of peak demand and the balance of generation and consumption in the electrical grid. However, the implementation of demand response programs is still quite limited, especially when it comes to residential consumers. As the cost of equipment that can help homes and businesses participate in demand response (i.e. smart meters, controllers and devices) decreases, there is a growing need for the design and implementation of programs to get interested consumers involved with DR. This thesis explores the use of clustering techniques to aid in the design and implementation of a Demand Response (DR) program for a network of commercial and residential prosumers. The goal of the program is to shift participant’s consumption behaviors to mitigate two issues with demand and generation timing in the electrical grid: a) reverse power flow, that occurs when generation from solar panels in the local grid exceeds consumption and b) system wide peak demand, that typically occurs during hours of the late afternoon. For the clustering stage, three popular algorithms for electrical load clustering, namely k-means, k-medoids and a hierarchical clustering algorithm, along with two different distance metrics, Euclidean and constrained Dynamic Time Warping (DTW), are evaluated using different validation metrics. The best configuration is employed to divide the dataset’s daily load profiles into clusters and each cluster is analyzed in terms of load shape, mean entropy and distribution of load profiles from each load type. These characteristics are then used to distinguish the clusters that would be most likely to aid with the DR program’s objectives and to select DR program structures that would fit each cluster. Finally, the thesis proposes the design of a DR system that uses forecasting, clustering and a demand projection engine based on price to produce daily, individualized DR recommendations and pricing structures for each costumer participating in the program. Apart from leading to the implementation of a possible DR program for the network analyzed, the methodology followed in this thesis along with the proposed digital system can hopefully help with the design and implementation of more DR programs in the future.	en
heal.advisorName	Ασκούνης, Δημήτριος	el
heal.advisorName	Askounis, Dimitrios	en
heal.committeeMemberName	Askounis, Dimitris
heal.committeeMemberName	Doukas, Haris
heal.committeeMemberName	Psarras, John
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών.	el
heal.academicPublisherID	ntua
heal.numberOfPages	77 σ.	el
heal.fullTextAvailability	false