Sensitivity Analysis of Data Sampling Frequency in a Multiple Linear Regression Model with application to the Fuel Oil Consumption Calculation

Φλουρής, Χριστόφορος; Flouris, Christoforos

dc.contributor.author	Φλουρής, Χριστόφορος	el
dc.contributor.author	Flouris, Christoforos	en
dc.date.accessioned	2022-04-11T09:13:15Z
dc.date.available	2022-04-11T09:13:15Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/55059
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.22757
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	Data sampling frequency	en
dc.subject	Noon Reports	en
dc.subject	Continuous Monitoring System	en
dc.subject	Multiple linear regression	en
dc.subject	Prediction models	en
dc.subject	Fuel oil consumption
dc.subject	Συχνότητα δειγματολειψίας
dc.subject	Καθημερινές Αναφορές
dc.subject	Συνεχές Σύστημα Παρακολούθησης
dc.subject	Πολλαπλή γραμμική παλινδρόμηση
dc.subject	Μοντέλο πρόβλεψης
dc.subject	Κατανάλωση καυσίμου
dc.title	Sensitivity Analysis of Data Sampling Frequency in a Multiple Linear Regression Model with application to the Fuel Oil Consumption Calculation	en
dc.title	Ανάλυση Ευαισθησίας της Συχνότητας Δειγματολειψίας Δεδομένων σε Μοντέλο Πολλαπλής Γραμμικής Παλινδρόμησης με Εφαρμογή στον Υπολογισμό της Κατανάλωσης Καυσίμου	el
heal.type	bachelorThesis
heal.classification	Μοντέλο πρόβλεψης	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2021-11-02
heal.abstract	In the last decades, the shipping companies aim to reduce the fuel oil consumption of their ships, both for financial and environmental reasons. By estimating the consumption of fuel oil in advance, more can be done to improve the performance of the vessels. The current diploma thesis is a sensitivity analysis of the data sampling frequency in a prediction model, with application in the calculation of the fuel oil consumption of a ship. The analysis includes operational data of a bulk carrier for a nine-month period, measured by specific onboard devices and recorded in datasets with different sampling rates. Therefore, the study includes data from the Continuous Monitoring System, which are logged automatically very often, and data from the Noon Reports of the ship, which are noted by the ship’s crew on a daily basis. Firstly, the data are processed and filtered for outlying data points, following a specific procedure of several stages, such as the application of threshold values and the detection of the outliers with a statistical process. The final filtered datasets, which are the basis for a reliable regression analysis, are compared with each other in tables, using basic statistical quantities. Subsequently, four multiple linear regression models are produced, one for each dataset, in order to calculate the fuel oil consumption, based on weather and sea travel-related parameters, such as the wind speed, the mean draft and the speed of the vessel. The occurring regression models and their results are compared in suitable graphs and tables, as for their fitting with the available data and between each other. Finally, an evaluation of the analysis is conducted, to find out which model gives the most trustworthy results for the prediction of the fuel oil consumption of the vessel, in combination with the advantages and disadvantages of the different sampling frequencies of the acquired data.	en
heal.advisorName	Themelis, Nikolaos	el
heal.advisorName	Θεμελής, Νικόλαος	en
heal.committeeMemberName	Σπύρου, Κωνσταντίνος	el
heal.committeeMemberName	Γρηγορόπουλος, Γρηγόριος	el
heal.committeeMemberName	Spyrou, Konstantinos	en
heal.committeeMemberName	Grigoropoulos, Grigorios	en
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ναυπηγών Μηχανολόγων Μηχανικών. Τομέας Μελέτης Πλοίου και Θαλάσσιων Μεταφορών	el
heal.academicPublisherID	ntua
heal.numberOfPages	146 σ.	el
heal.fullTextAvailability	false