船舶舱室的火灾实时安全预测

Real-time fire safety prediction of ship cabins

利用机器学习的方法,可以有效减少复杂或未知过程的计算。为了及时预测出船舶舱室火蔓延的方向,减少火灾伤害,提出了一种基于BP神经网络的客舱火势蔓延模型。利用MATLAB的神经网络工具箱,对数据进行学习和检验。模型通过输入火源的燃烧热释放率、舱内氧气浓度和温度,可以有效而准确地预测出空间内其余可燃物被蔓延的时间及发展趋势。经验证,该模型可以预测不同条件下的客舱火灾蔓延,具有火灾的预测和判断能力,为船上灭火和救援提供了可靠的依据。

Using machine learning methods can effectively reduce the calculation of complex or unknown processes.In order to predict the fire propagation direction of ship cabins in time and reduce fire damage,a model of cabin fire spread based on BP neural network was proposed.Use MATLAB neural network toolbox to learn and test the data.By inputting the combustion heat release rate,oxygen concentration in cabin and temperature,the model can effectively and accurately predict the time and trend of the spread of other combustibles in the space.It is verified that this model can predict the fire spread in cabin under different conditions and has the ability of fire prediction and judgment,which provides a reliable basis for fire-fighting and rescue on board.