Research On Control Strategy Of Ship Fire Automatic Alarm System

In recent years, the ship has been concerned by more and more people as one of themajor incidents at sea. As an important part of the ship, if a cabin fire, it will trigger a chainreaction of fire when it spreads to several adjacent cabins, and bring hazards to people.According to the above we can see, it is important to study fire alarm system for earlydetection of cabin chain fire、the cause of the fire and discrimination of fire types、inquiries ofalarm level、trends of fire development and should start which linkage equipment toextinguish, the accurate fire discrimination and rationalization of fire fighting strategy caneffectively reduce the ship’s chain fire loss. In summary, the research of ship fire alarmsystem for the cabin chain alarm control strategy has several great significances.Based on the analysis of ship fire characteristics and developments of intelligent firealarm technology at home and abroad, this paper put forward a kind of three-stage fire alarmcontrol strategy, for the linkage control thought of the ship cabin fire alarm system,analyzedsome fire alarm discrminant elements of typical cabins, ship cabin fire chain alarm levelassessment model was established by quantized, and used LM algorithm to learn and train ofthe assessment model, verified the feasibility and accuracy of network evaluation modelthrough the test specimen. The method is helpful to improve the alarm system fire detectionaccuracy and adjacent cabins chain reliability, reduce the fire caused loss of ship cabincorrelation probability. Because of the designed cabin chain fire alarm level assessmentmodel based on BP algorithm is not perfect in some ways, the training parameters is morelikely to lead to the generalization capability of the network to drop lower classificationaccuracy, a direct impact on the accuracy of the prediction model. Therefore, established thefuzzy expert system of ship engine room fire alarm based on the multi-source informationfusion technology, this system taked the collected information into fuzzy inference andjudgment based on the expert system knowledge database of ship engine room fire alarmsystem, got the engine room fire alarm level information、the cause of fire、the trend of fire、fire linkage and expert advice information. In order to verify the feasibility of the system,using MATLAB and VC++mixed programming technology to realize the ship cabin firealarm expert system operation and display interface, and carried out a variety of engine roomfire simulation experiments, the effect is good.Finally, building the fire alarm control system hardware experiment environment, taking a ship as the experimental object, verificated the cabin fire chain alarm level assessmentmodel based on the neural network and engine room fire alarm fuzzy expert systemexperiments which have been created, such as simulated a single compartment fire alarmsituation or multiple compartments alarm situation, each compartment can be arranged at leasttwo detectors, heat detectors and smoke detectors, also can install a manual alarm button. Theuse of lighter, hot air guns and cigarettes simulated open flames or smoldering fires, the testshown that the realization of the two applied to ship compartment chain fire alarm control, notonly increased the ship fire alarm intelligent and humane, made alarm more timely, reliable,judgment fire more comprehensive, accurate, but also made the firefighters extinguishingrecommended to make more reasonable response in a timely manner, effectively prevent thespread of the fire of the ship, helps to reduce the cabin fire alarm false positives, leakagereported frequency.