Development of a virtual sensor using convolutional neural networks

Kylikas, Angelos; Κυλίκας, Άγγελος

dc.contributor.author	Kylikas, Angelos	en
dc.contributor.author	Κυλίκας, Άγγελος	el
dc.date.accessioned	2023-01-25T11:25:38Z
dc.date.available	2023-01-25T11:25:38Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/56908
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.24606
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	Convolutional Neural Networks	en
dc.subject	Virtual Sensor	en
dc.subject	Engine Modelling	en
dc.subject	Time Series Prediction	en
dc.title	Development of a virtual sensor using convolutional neural networks	en
heal.type	bachelorThesis
heal.classification	Marine Engineering	en
heal.classification	Neural Networks	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-10-31
heal.abstract	In the current thesis, the development of a virtual sensor with the use of deep convolutional neural networks is investigated. The application, concerns a hybrid marine engine and the embedding of the software on it. On the first part, the mathematical foundations of the neural networks are described, with particular focus on convolutional neural networks and the tools required. Following that, the fundamental parameters of the engine operation are presented, concentrating on the parameters important for the current application. The next chapter, concerns the presentation of the physical system itself as well as the data used pre-processing. Additionally, the software embedding on the engine with the use of a Raspberry Pi coupled with a CAN bus controller, is discussed . The following part, is dedicated to specifying the model architectures as well as the statistical description of the data and their relationships. After selecting the inputs of the models, their training and the corresponding accuracy results are presented. Lastly, concluding remarks are discussed. The whole thesis, is carried out with the use of Julia Programming Language, which is also used for the software embedding on the Raspberry Pi.	en
heal.advisorName	Papalambrou, George	en
heal.committeeMemberName	Lyridis, Dimitrios	en
heal.committeeMemberName	Papadopoulos, Christos	en
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ναυπηγών Μηχανολόγων Μηχανικών. Τομέας Ναυτικής Μηχανολογίας	el
heal.academicPublisherID	ntua
heal.numberOfPages	102 σ.	el
heal.fullTextAvailability	false