DANAOS Shipping is one of the largest managers of modern, large-size containerships voyaging through the open seas, planning trips that last many weeks. EO4EU aims to limit the solution space and increase the throughput reducing the computational time. Such an approach will have a positive impact on ships travelling, towards their destination, as it will limit the time of the optimised route. 

Shipping deployed in long distance liner services will be engaged. The whole process is described in the following steps/phases:

1. Phase 1: Before vessel departure the navigational Officer on board plots a voyage plan. Then synchronisation of the on-board version with the main optimization engine ashore allows continuous monitoring and update of Master’s route plotting decisions along the voyage and any deviation from the initial plan. Routing advice is sent back to the vessel where the Master is able to visualise the optimised route with respect to the forecasted weather en route.
2. Phase 2: Post voyage route assessment. At the end of the voyage performance results will be displayed (travel time, consumption, average speed, etc.) while a voyage performance comparison will be triggered between the actual route, driven by the master’s decision, and the system’s route advice baseline. Both routes will be fed with real weather data and will be assessed against actual voyage conditions.  

Challenge

Optimal Weather Routing is an NP-hard problem. The International Maritime Organization (IMO) recognizes that weather routing must be available to shipping in the form of recommended “optimum routes” for individual crossings of the oceans. Practice of weather routing has proved of benefit to ship operations and safety as well as to their crews and cargoes. Hence, for substantially large routes (e.g., trans-Atlantic), the computational time may even reach 20-30 mins per run. Hence, even a small increase in granularity (i.e., one point every 0.5 degrees) may lead to computational times that exceed an hour, which is unacceptable for the standards set for the service.

Solution

EO4EU aims to enhance the quality of information and the optimization method so as to produce better analysis results and more efficient routes in terms of fuel consumption, safety and arrival time precision. DANAOS plans on integrating the EO4EU capability of handling extreme volumes of data by fusing the meteorological data collected from EO data sources and a vessel to perform route optimization during the voyage of the ship. A common model will integrate a fine grained weather report as close as the ship is en route to produce more accurate results. The existing data-driven methods will benefit from the modular and powerful architecture of EO4EU delivering extreme analytics over the collected meteorological data to reveal the hidden patterns and correlations between meteorological features/parameters. DANAOS envisages to define complex analytical queries based on the possible combinations of meteorological data adopting a high number of features, thus, to fully expose their effects in designing alternative routes.

Current status

Through the EO4EU platform UC2 and DAN has implemented an ETL pipeline that continuously harvests weather data and integrates it with operational data from vessels to optimize fuel consumption and routing. This pipeline helps train models that accurately predict fuel consumption (FOC). These predictions are utilized in real-time to suggest alternative, environmentally friendly routes, ensuring compliance with regulatory standards. Additionally, a Convolutional Neural Network (CNN) has been developed to monitor ocean conditions and predict ambient weather phenomena in real-time. The current work benchmarks weather data with different spatio-temporal granularities, improving the reliability of route optimization models. 

The proposed framework in the context of UC2 and its main components is visualized in Figure 1 below. 

Leader:   DANAOS Shipping (Leader for the UC2 "Ocean Monitoring" Use Case)

• Meteorological Data: Collected from various Earth Observation (EO) sources, offering high granularity. 

• Vessel Operational Data: Data from on-board sensors, capturing real-time information like speed, fuel consumption, and environmental conditions. 

• Weather Routing Data: Used to improve route optimization, focusing on safety, fuel consumption, and timely arrival.