基于贝叶斯网络的复杂装备D-S诊断  

D-S Diagnosis Based on Bayes Network

作　　者：王波[1] 吴华丽[1] 王灿林[1] 

机构地区：[1]海军航空工程学院,山东烟台264001

出　　处：《计算机仿真》2009年第4期34-37,共4页Computer Simulation

摘　　要：利用D-S证据理论进行故障诊断时,首先需要确定出故障识别框架和基本置信分配。但在导弹武器等复杂装备中,存在着故障原因与故障征兆之间关联关系不确定的现象,增加了求取故障识别框架和基本置信分配的困难。文中提出了利用贝叶斯网络理论来解决难题的方法,使D-S证据理论能在复杂装备故障诊断中得到应用。文中首先给出了诊断贝叶斯网络的适应度函数,并利用改进的粒子群优化算法求解了诊断贝叶斯网络。然后分析了诊断贝叶斯网络中故障原因与故障征兆间的关联关系,给出了征兆集合、异因征兆集合概念,提出了求解故障识别框架和基本置信分配的方法。最后,通过实例的D-S诊断,验证了所提方法的正确性。In fault diagnosis with D - S evidence theory, defining exactly the Basic Belief Assignment and the identification frame of diagnosis is the first step. In complex equipments of weapons , the relation between causations of fault and symptoms of fault is uncertain, which increases the difficulty of gaining Basic Belief Assignment and the identification frame of diagnosis. So a kind of method based on Bayesian network for diagnosis is put forward. First, this paper gives fitness function of Bayesian network which is solved with an improved particle swarm optimization. Then a method for solving Basic Belief Assignment and the identification frame of diagnosis is given. In the end of this paper, an example of D - S diagnosis is given to validate the correctness of this method.

关 键 词：诊断 复杂装备 诊断贝叶斯网络 识别框架 基本置信分配 

分 类 号：TJ765.43[兵器科学与技术—武器系统与运用工程]