A fuel consumption prediction model for ships based on historical voyages and meteorological data

Abstract

Predicting the fuel consumption of a ship during a voyage is a challenging task, given the internal and external factors that influence it. This challenge has gained crucial importance in light of the regulations imposed by the International Maritime Organization, which aim to reduce greenhouse gas emissions from ships. The objective of this study is to develop a fuel consumption prediction model using data collected from bulk carriers. These predictions will serve as input for a ship routing tool aimed at optimising routes while considering fuel consumption and, consequently, emissions. We propose a data-driven approach to develop a predictive model of fuel consumption for these bulk carriers using a multiple linear regression model considering the propeller rotational speed and with a particular focus on weather factors such as wind, waves and currents, each contributing to the overall speed loss. The results show that the estimated fuel consumption of the studied bulk carriers is strongly affected by the engine setting and the meteorological conditions. The developed model can predict fuel consumption accurately for more than 80% of the voyages of the dataset with a mean absolute error and a root of the mean squared error lower than 0.01 metric ton per nautical mile, and a mean absolute percentage error of less than 15%, making it useful for ship routing purposes.