ABSTRACT
This paper presents the results of the implementation of a virtualized model of the behavior of the firefighting (FiFi) system that allows autonomous detection and response of the system to failures. The developed system which is fed by this virtual model is based on Artificial Intellligence (AI) and will be implemented in the ship’s digital twin, providing it with intelligence. For the elaboration of this virtualized model, we have started with data obtained in the tests carried out on board, as well as the results of simulations of the FiFi system. In these on-board tests/simulations, a double characterization of the system has been carried out: obtaining, on the one hand, data that represents normal operation, and on the other, the rupture and the collapse (total or partial) of the pipes. The combination of both has allowed to obtain a sufficiently broad and representative data collection to carry out the training of an AI - based system.
