A two-step strategy for fuel consumption prediction and optimization of ocean-going ships

Abstract

The prediction and optimization of ship fuel consumption are the core techniques for the successful implementation of smart and green shipping. However, in previous studies, there were some drawbacks with poor prediction performance based on single models, as well as a few studies on data-driven trim optimization. To bridge these gaps, a two-step strategy for the prediction and optimization of ship fuel consumption was proposed in this study. In the first step, the collected fuel consumption data is processed, and a novel hybrid prediction model is then developed based on the stacking theory by fusing several state-of-the-art single models. Subsequently, the second step proposes a method based on the developed hybrid model in combination with the enumeration method to optimize the fuel consumption from the perspective of trim adjustment. To support the proposal, two real-world voyages from bulk carrier are taken as examples. Through experimentation, the proposed hybrid model has better accuracy and robustness than the other seven popular single models. Furthermore, trim optimization can effectively reduce the fuel consumption and carbon emissions by 0.69%–1.82%. This study provides fundamental theoretical, methodological, and technical support for fuel saving and emission reduction in ships.