Σύνοψη έξυπνου δικτύου και βραχυπρόθεσμη πρόβλεψη κατανάλωσης ηλεκτρικής ενέργειας σε οικιακούς χρήστες
An overview on smart grid and short-term residential load forecasting
View/
Keywords
Έξυπνο δίκτυο ; Smart grid ; AMI ; Χρονοσειρές ; Timeseries ; SARIMAX ; LSTM ; Encoder-Decoder LSTMAbstract
This research has been conducted as part of the postgraduate program at the University of Piraeus, Department of Digital Systems, Big Data & Analytics. The work is divided in two main sections: overview of smart grids and short-term load forecasting for residential users. Energy grid is on a transition phase from a conventional grid to a smart one. This work presents the main differences between those two and highlights the main components and the benefits of a smart grid. The deployment of such an infrastructure and more specifically the installation of smart meter, has increased massively the amount of collected data. Providers started diving into those data in order to identify patterns and improve their services and profit margins. Artificial intelligence and machine learning technics are currently used in large scale for load forecasting, which is necessary for planning, demand response, supply-demand equilibrium. The past five years there is a high interest in load forecasting for residential consumers; a task that is very challenging due to the volatility of such data. There many different consumption patterns depending on the area, type of house, demographics of the residents, weather, existence of solar panels, existence of an electric vehicle. The data, examined in this work, was collected during the GridFlex Heeten project [99]. The data was collected between August 2018 and August 2020 in 77 households all situated in Heeten (The Netherlands) and consists of electricity consumption per minute per household. After performing an exploratory data analysis, we created individual models, SARIMAX, Vanilla-LSTM, Encoder-Decoder LSTM, for three houses and compared the results.
Postgraduate Studies Programme
Πληροφοριακά Συστήματα και ΥπηρεσίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνNumber of pages
77Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα