Reducing cargo ship emissions through energy demand optimized routing

A fleet of more than 50.000 cargo ships worldwide has an enormous demand for energy resulting in considerable emissions. According to the 4th International Maritime Organization (IMO) global greenhouse gas (GHG) study, maritime transport emitted around 1,056 million tonnes of CO₂ in 2018 and was responsible for about 2.9% of the global anthropogenic CO₂ emissions. While the emissions per tonne and nautical mile have been reduced by almost 30% in the last decade, the overall emissions of cargo ships increased by more than 10% (up to 30% in some models) due to the growing demand. In order to tackle this increase, the MariData project conducts research on how an improved hydrodynamical modeling and the use of detailed predictions and data of sea state and environmental conditions can reduce the energy demand and hence emissions of cargo ships. 

In this set-up, a cloud-based geo data platform takes on a central role, which combines different data sets from CMEMS, GFS, recorded ship trajectories and further data sources. The geo platform acts as a data broker and provider as well as a machine learning environment for data mining and route predictions. One of its use cases is a data driven machine learning approach where freely available records of AIS data are combined with sea state and weather information and serve as a training set for a random forest regression. This model is capable of predicting the expected speed of cargo ships (characterized by width, length and draught) based on the sea state and weather forecasts. Due to the lack of detailed data on fuel consumption or energy demand, we need to exploit a heuristic. Under the assumption of a constant engine load for free sailing areas, the achieved speed depends on the resistance due to environmental conditions. Pixels with a low resistance are then favored for an energy optimized route. The geo platform also collects and provides data that is used by partners of the research project in their own routing application or to enhance and test their hydrodynamical analysis. 

We will present the technical set-up combining the data sources and facilitating the subsequent data mining and data analysis. Preliminary results of models and optimized routes will be presented. Finally, limitations of the approach and the data availability will be discussed.