Χρήση μηχανικής μάθησης για την εκτίμηση της κατανάλωσης καυσίμου και των εκπομπών CO2 ενός πλοίου Newcastlemax σε πραγματικές συνθήκες

View/ Open
Keywords
Machine learning ; Random forest regression ; Fuel consumption ; CII ; PythonAbstract
In this thesis, Random Forest Regression, a machine learning technique, was utilised as the optimal method for the available data. This approach involved the utilisation of measurements from historical data concerning a Newcastlemax type vessel to develop a model for the estimation of main engine fuel consumption and CO2 emissions under realistic sailing conditions. In particular, the data were used for 14 voyages in different loading and sea states. In conclusion, using a Python environment, models were created to estimate consumption and hence CO2 emissions using both maximum and minimum possible data, i.e. data that could be obtained from a database (weather provider).