Ship energy consumption prediction with Gaussian process metamodel

Abstract

Shipping is a major contributor to the global CO2 emissions. It is important to improve the ship energy efficiency and reduce the ship emissions. Various emission reduction measures have been proposed to reduce the ship emissions. In order to evaluate the efficiency of these emission reduction measures, it is necessary to evaluate the reduction of fuel consumption for these measures. However, due to the complex of ship energy system, it is difficult to assess the ship fuel consumption under different scenarios considering both operational and weather conditions through physical systems or simulation models. In this paper, a Gaussian process metamodel is developed to predict the ship fuel consumption for different scenarios. This model not only considers the effects of operational conditions such as speed and trim, but also takes into account the impacts of weather conditions such as wind and wave effects. The effects of these factors on ship fuel consumption are further evaluated. The case study indicates the accuracy and efficiency of using Gaussian process metamodel for ship energy consumption prediction. Copyright � 2018 Elsevier Ltd. All rights reserved.