Comment prédire les risques de roulis ?

The objective of this project is to design an algorithm capable of predicting roll risks specific to each type of ship in order to guarantee the safety of the ship, its crew and its cargo. Discover this project in partnership with GTT.

context

In partnership with OSE, GTT is developing innovative software for ship weather routing. This program integrates various data, such as weather conditions, marine currents, and navigation constraints, in order to generate optimal trajectories. The objective is to reduce, according to needs, the distance travelled, fuel consumption or polluting emissions. As part of this project, our contribution consists in designing a complementary module that specifically addresses the risk of ship stability at sea, with an emphasis on the risks of parametric and synchronous roll.

White wave created by the bow of a boat moving in a deep blue sea.

stakes

The phenomena of parametric (asynchronous) roll and synchronous roll represent dangers, especially for container ships. These phenomena have the ability to trigger intense rolling movements, increasing the risk of cargo deterioration or loss. Their unpredictable nature lies in the resonance between the wave frequencies perceived by the ship and those of its own rolling motion. In order to solve this challenge, GTT collaborates with OSE to develop behavioral models specific to each ship category. These models will make it possible to produce risk maps specific to each ship, thus improving forecasting and prevention capacity.

enhancing maritime safety through intelligent Reduced-Order Models

Using a machine learning–based approach, OSE designs reduced-order models capable of anticipating the behaviour of different vessels and limiting rolling events, thereby improving the safety of both crew and cargo. Does this solution align with your needs?

Let's talk

missions

ADVANCED PHYSICAL SIMULATIONS

Development of a method based on fine simulations and statistical analyses to assess the seaworthiness of the ship according to different loads, speeds and wave states.

ROLL RISK DATABASE

Realization of a large number of simulations in order to constitute a database dedicated to the risks of roll on various categories of vessels.

REDUCED MODEL USING MACHINE LEARNING

Construction of a rapid model based on simulation data, using learning techniques to estimate the roll risks specific to each ship.

GENERALIZATION TO THE VARIOUS VESSELS

Variation and adaptation of these reduced models for the various families of ships in order to broaden their field of application.

INTEGRATION WITH WEATHER ROUTING SOFTWARE

Implementation of these models within the weather routing software for operational use in navigation.

solution

We have developed in-depth expertise in a complex field that remains an unresolved concern in the maritime sector. Research is still ongoing, with specific theses aimed at finding predictive solutions to alleviate the roll problem and at developing appropriate countermeasures.

In this perspective, we have developed an initial model that offers the ability to assess roll risks according to the characteristics specific to a ship and its load, without requiring dedicated simulations. This approach represents a significant innovation in the current context of the maritime industry.

In addition, we have developed a solution that will provide the routing software with a real-time assessment of the risks associated with stability at sea. This advance has the potential to contribute significantly to improving the safety and efficiency of maritime operations.

How to predict rolling risks?