基于D-S证据理论信息融合的故障诊断方法  

Fault Diagnosis Method Based on D-S Evidence Theory Information Fusion

作　　者：屈志宏[1] 杨传道[1] 李方[1] 

机构地区：[1]空军工程大学工程学院,陕西西安710038

出　　处：《火炮发射与控制学报》2008年第4期107-110,共4页Journal of Gun Launch & Control

摘　　要：对证据推理理论用于故障诊断识别进行了论述和仿真。介绍了证据理论的相关重要概念,给出了一种常用构造基本概率赋值函数的方法以及分类规则的选取。应用D-S证据理论方法对系统进行故障分析,首先要解决的问题是要分析系统识别框架和故障特征向量的构造。以航空电源某故障为例,进行了诊断识别仿真和对比实验,验证了证据推理用于故障诊断的可行性。仿真结果表明,将证据推理方法应用于航空电源状态监测和故障诊断中,可大幅度提高故障诊断的准确性和置信度。The application of D-S evidence theory in fault diagnosis and fault recognition were discussed and simulated. Interrelated conceptions of evidence theory were introduced, and a method of constructing basic probability evaluation function and selecting classification rule was presented. By use of D-S evidence theory method, the system fault was analyzed. The problem to be solved firstly was to analyze system recognition frame and construction of fault feature vector. Taking a fault of aviation power supply system as an example, the diagnosis, simulation, contrast and recognition experiments were performed, and the feasibility of application of D-S evidence theory in fault diagnosis was verified. The simulation results indicated that this method can be used in state monitor and fault diagnosis of aviation power supply system, and it can greatly improve accuracy, credibility, and reliability of fault diagnosis.

关 键 词：信息处理技术 证据理论 信息融合 故障诊断 决策规则 状态监测 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]