Session: 08-04-01 Deep Learning based ROM for Wave Dynamics and Prediction in Marine Engineering

Paper Number: 126533

126533 - A Modern Data-Driven Approach for Accurately Assessing Wave-Induced Loads and Responses of Ships Using Ais and Weather Data 

Prescriptive load formulations used during the design process for most ships are based on a presumed design environment. For standard ships, the strength evaluation is performed according to the North Atlantic design environment, while the fatigue design is verified according to the Worldwide environment. However, it is well-known that some ships operate globally, while others operate in specific areas, being subjected to different environmental conditions than assumed during the design phase. Therefore, the prescriptive loads from design may not be relevant for the actual operation of certain ships. Consequently, within the last decade, significant attention has been given to digital twin models and hull monitoring systems, where the hull structure utilization (e.g., strength and fatigue) is continuously assessed by combining structural models with actual encountered weather conditions. Setting up advanced digital twin models and running the analyses takes time and effort, preventing such models from being used operationally on a large scale. Aiming to reduce the risks and increase the profitability of ships, as well as to make structural health monitoring solutions more accessible, an innovative hull condition monitoring system is proposed by combining advanced data-driven techniques with hindcast weather data. This paper presents the novel method developed for accurately approximating the wave-induced loads and responses for all ships without complex hydrodynamic models. The solution makes accurate predictive and preventive maintenance accessible for all.

Presenting Author: George Jagite DNV

Presenting Author Biography: TBD

Authors:

George Jagite DNV
Lars Holterud Aarsnes DNV
Gaute Storhaug DNV