Εφαρμογές μηχανικής μάθησης σε ηλεκτρικά οχήματα
Machine learning applications in electric vehicles
View/
Keywords
Ενεργειακή ζήτηση ; Μηχανική μάθηση ; ΠρόβλεψηAbstract
This thesis focuses on the field of deep learning. More specifically, it aims to address the issue
of accurately predicting the energy demand of chargers in an electric vehicle (EV) charging
network. Initially, a methodology is proposed in order to solve this particular problem. The
methodology suggests the use of deep learning technologies with the ultimate goal of
examining, evaluating and comparing three models based on convolutional neural networks for
short-term energy demand prediction. The application of this methodology utilizes an opensource dataset containing data about a charger network in the city of Palo Alto, California. After
their creation and training, the models were evaluated using RMSE on the forecast for the total
system. Based on the lowest RMSE scores, it was conducted that the T-GCN model is the ideal
model for energy demand predictions.
Department
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
48Language
GreekCollections
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