基于神经网络与D-S证据理论的故障诊断  

Fault Diagnosis Based on Integrated Neural Network and D-S Evidential Reasoning

作　　者：王娜[1] 梁禹[1] 

机构地区：[1]沈阳化工学院信息工程学院,沈阳110142

出　　处：《仪器仪表学报》2005年第z1期773-774,共2页Chinese Journal of Scientific Instrument

摘　　要：在集成神经网络故障诊断的基础上,引入证据理论,提出将神经网络(NN)与D-S证据理论结合的故障诊断方法(NN-DS诊断方法)。多个征兆域分别用神经网络进行局部诊断,从每一个网络都可以得到各自的诊断结果,再用D-S证据理论对这些诊断结果进行全局融合,得到综合诊断结果。算例分析表明了该方法的正确性。On the base of integrated neural network, importing evidential reasoning, a fault diagnosis technique which combine neural network and D-S evidential reasoning (NN-DS diagnostic technique) is proposed. It uses neural network local diagnosis respectively from different symptom field, and each network receives respective result, then D-S evidential reasoning will be used for global diagnosis to gain a unified result. At last an exemple is given to indicate it's validity.

关 键 词：故障诊断 神经网络 D-S证据理论 

分 类 号：TH7-55[机械工程—仪器科学与技术]