Abstract
Fuel oil consumption constitutes over 25% of a vessel’s overall running costs. Therefore, accurately forecasting, and optimising fuel costs majorly impacts a vessel’s operation sustainability and profitability. This paper presents data-driven, multivariate main engine fuel consumption models leveraging the vast amount of data currently being recorded onboard vessels. Different data-driven modelling methodologies, such as shallow neural networks, deep neural networks, support vector machines, and random forest regressors are presented and implemented, comparing results. The suggested multivariate modelling allows the uncovering of latent interconnections that increase the robustness of the model in varied operating conditions.
| Original language | English |
|---|---|
| Title of host publication | 17th Conference on Computer and IT Applications in the Maritime Industries |
| Place of Publication | Hamburg |
| Pages | 144-152 |
| Number of pages | 9 |
| Publication status | Published - 8 May 2018 |
Keywords
- fuel oil consumption
- fuel costs
- sustainability
- multivariate modelling