Intelligent Identification of Fires in Ship Compartments Using a Bayesian Network

Fire is always a severe threat to ship safety and survival. To prevent the spread of a fire and eliminate serious accidental consequences, it is imperative for commanders to promptly identify the size and type of the fire so as to take rapid and effective firefighting action. In this study, the architectural design of an advanced ship fire identification system (SFIS) is presented that makes timely and critical decision support for selecting suitable suppression methods and firefighting tactics. Based on a Bayesian network (BN), a novel intelligent identification model that is capable of identifying small, medium or large fires and distinguishing between a solid fire and a fuel oil fire is proposed. The results indicate the effectiveness of the proposed model as well as its robustness during the failure of one fire sensor. The model can be integrated into damage control systems (DCSS) to further enhance the situational awareness of the damage and assist commanders in prompt decision-making by allocating the most efficient firefighting equipment and crew.