Research on the fire source location of aircraft cargo compartment based on deep learning

In view of the high false alarm rate of fire detection and lack of early efficient prevention and control methods, this paper based on deep learning algorithm for a fire area location and local application method of fire extinguishing, using fire simulation software Pyrosim full size B707 aircraft hold fire model, simulate 18 fire location, 8 kinds of fire power (20 KW~800 KW) a total of 144 fire scenarios, obtain the key characteristics of cargo fire smoke parameters as a fire data set. A neural network-based localization model of ignition source region was developed to deeply analyze the correlation between parameters such as fire temperature and gas concentration and fire source location. Find the best hyperparameter combination to improve the accuracy and robustness of the model. The results show that the model can reveal the smoke in the cargo hold canopy jet fire dynamics law, can accurately predict the cargo hold fire area location, accuracy of 95.5%, the fire area location is conducive to strengthen the fire extinguishing effect, research results to provide useful reference for aircraft cargo hold intelligent airborne fire prevention design.