In the last decade, the application of progressive flooding simulations improved the decision support available onboard in case of flooding. However, all these Decision Support Systems (DSS) rely on flooding sensors, thus cannot be adopted on the large majority of the existing fleet without a costly retro fi t of the flooding detection system. Here an alternative has been proposed to assess the main flooding consequences from the time evolution of the fl oating position, which can be recorded with a very basic set of sensors. Here the decision trees are employed to assess the final fate of the ship, the damaged compartments and estimate the time-to-flood. Decision trees are here trained by means of two types of databases of progressive flooding simulations: one based on Monte Carlo (MC) generation of damages according to SOLAS probability distributions and a parametric one. The method has been tested on barge geometry employing another MC database for validation purposes.
Application of decision trees to predict damage consequences during the progressive flooding
Braidotti, L.
;
2021-01-01
Abstract
In the last decade, the application of progressive flooding simulations improved the decision support available onboard in case of flooding. However, all these Decision Support Systems (DSS) rely on flooding sensors, thus cannot be adopted on the large majority of the existing fleet without a costly retro fi t of the flooding detection system. Here an alternative has been proposed to assess the main flooding consequences from the time evolution of the fl oating position, which can be recorded with a very basic set of sensors. Here the decision trees are employed to assess the final fate of the ship, the damaged compartments and estimate the time-to-flood. Decision trees are here trained by means of two types of databases of progressive flooding simulations: one based on Monte Carlo (MC) generation of damages according to SOLAS probability distributions and a parametric one. The method has been tested on barge geometry employing another MC database for validation purposes.File | Dimensione | Formato | |
---|---|---|---|
2021_MARSTRUCT_MachineLearning.pdf
Accesso chiuso
Descrizione: Articolo principale
Tipologia:
Documento in Versione Editoriale
Licenza:
Copyright Editore
Dimensione
806.08 kB
Formato
Adobe PDF
|
806.08 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
3024815_2021_MARSTRUCT_MachineLearning-Post_print.pdf
Open Access dal 30/12/2022
Tipologia:
Bozza finale post-referaggio (post-print)
Licenza:
Digital Rights Management non definito
Dimensione
1.31 MB
Formato
Adobe PDF
|
1.31 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.