MOLS-SVM在舰船火灾探测中的应用

Application of MOLS-SVM in the fire detection of naval vessel

采用多输出最小二乘支持向量机(MOLS-SVM)方法,基于舰船火灾初期CO含量、烟雾浓度和温度变化等特征,对舰船无火、阴燃火、明火的发生概率进行预测。仿真结果显示,采用MOLS-SVM方法能及时、准确地对三种不同类型火灾的发生概率进行预测,精度较高,且MOLS-SVM是一种智能算法,能提高探测的智能化水平。

Take the MOLS-SVM （the Multiple Output Least Square Support Vector Machine） method and select some early characters of Naval Vessel＇s fire, such as the CO volume frac- tion, the concentration of smoke, the change of temperature and so on, to predict naval vessel＇s fire occur probability, in- clude fire-free, smoldering fire and open fire. The simulations indicated that the MOLS-SVM method can make timely accu- rate prediction to the probability of three different types of fire, moreover, the degree of accuracy is high. Meanwhile, the MOLS-SVM is an intelligent algorithm, and it can improve the intelligent level of detection.