一种基于SVM的航舵故障诊断方法  

A Method of Nautical Steer Fault Diagnose Based on SVM

作　　者：周晶[1] 宋辉[1] 李铁[1] 李明海[1] 

机构地区：[1]海军大连舰艇学院训练部模拟训练中心,辽宁大连116018

出　　处：《兵工自动化》2012年第4期56-58,共3页Ordnance Industry Automation

摘　　要：针对舵机故障样本数量不足、诊断知识不完备的实际情况,提出一种基于支持向量机(support vectormachine,SVM)的故障诊断方法。根据航舵故障输入输出映射非线性的特点,分析SVM的分类机理,对基于SVM的故障诊断步骤进行介绍,解决了小样本模式的分类问题,并通过仿真对该方法的有效性进行验证。仿真结果表明:该方法对舵机故障分类准确性可达92%。Aiming at lack of fault samples and diagnosis knowledge in nautical steer, introduce a method based on support vector machine （SVM）. According to nautical steer nonlinear input-output mapping feature, analyze SVM classification mechanism, introduce the fault diagnosis steps based on SVM, and solve the classification problems of small sample mode. Then, use simulation to verify the validity. The simulation result shows that the method fault classification correctness can reach 92%.

关 键 词：计算智能 支持向量机 故障诊断 

分 类 号：TJ03[兵器科学与技术—兵器发射理论与技术]