Neural networks based model predictive control of a hybrid diesel-electric marine propulsion plant

Σωτηρόπουλος, Παναγιώτης; Sotiropoulos, Panagiotis

dc.contributor.author	Σωτηρόπουλος, Παναγιώτης	el
dc.contributor.author	Sotiropoulos, Panagiotis	en
dc.date.accessioned	2025-09-08T07:25:51Z
dc.date.available	2025-09-08T07:25:51Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/62374
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.30070
dc.rights	Default License
dc.subject	Predictive control	en
dc.subject	Neural-Networks	en
dc.subject	Diesel-Electric	en
dc.subject	Hybrid propulsion	en
dc.title	Neural networks based model predictive control of a hybrid diesel-electric marine propulsion plant	en
heal.type	bachelorThesis
heal.classification	Hybrid propulsion	en
heal.classification	Predictive control	en
heal.classification	Diesel-Electric	en
heal.classification	Battery	en
heal.classification	Neural-Networks	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2025-03-06
heal.abstract	This thesis investigates power-split strategies for a hybrid marine propulsion plant for optimal energy management and emissions, using neural networks-based model predictive control. Initially, each component of the power plant is been presented and analyzed. Then both non-linear and neural networks-based models were developed for the main engine, while several non-linear dynamic models were examined for the other components. This procedure developed accurate and efficient models, which were then implemented into the controller. The controller utilized Model Predictive Control (MPC) to solve the system’s optimization problem in real time. MPC achieved handling this nonlinear multi-variable problem with constraints by minimizing an objective function over a finite horizon. Its aim was to reduce fuel consumption and NOx emissions while following a speed reference, paying respect in predefined constraints. Finally, the developed controller was tested and validated with simulations by utilizing a virtual model of the HIPPO-2 hybrid diesel-electric power plant and a propeller loading model in Simulink. These tests were conducted for various different loads into realistic marine scenarios and the results were compared with these of a conventional diesel engine power plant utilizing PID control.	el
heal.advisorName	Παπαλάμπρου, Γεώργιος	el
heal.committeeMemberName	Παπαδόπουλος, Χρήστος	el
heal.committeeMemberName	Δημόπουλος, Γεώργιος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ναυπηγών Μηχανολόγων Μηχανικών. Τομέας Ναυτικής Μηχανολογίας	el
heal.academicPublisherID	ntua
heal.numberOfPages	98 σ.	el
heal.fullTextAvailability	false