基于模拟退火算法的船用核动力装置故障诊断研究  

Study on Fault Diagnosis of Marine Nuclear Power Plant Based on Simulated Annealing

作　　者：朱尧[1] 周婷[1] 李伟[1] 孔衍[1] 

机构地区：[1]海军蚌埠士官学校机械系,安徽蚌埠233012

出　　处：《数学的实践与认识》2015年第13期187-191,共5页Mathematics in Practice and Theory

基　　金：总后勤部科研基金(AWS10B026)

摘　　要：针对核动力装置故障诊断存在的诊断精度低等问题,提出了一种基于模拟退火算法和概率因果模型相结合的故障诊断方法.首先根据故障样本集和概率因果理论建立动态多故障诊断模型,将复杂系统的多故障诊断转换成非线性规划问题.利用模拟退火算法对该问题进行求解,并建立了诊断测试系统.测试结果表明,方法能有效避免误诊、漏诊现象,可用于复杂核动力装置的动态多故障诊断.Aimed at the problems of low diagnosis accuracy in marine nuclear power plant, a new technology based on simulated annealing（SA） and probability causal model was pro- posed. Dynamic multiple fault diagnosis （DMFD） model was established according to fault example set and probability causal theory, which was used to convert DMFD to a non-linear programming. SA was used to solve the problem, and the fault diagnosis system was estab- lished. It is indicated from the testing result that with the algorithm, mistaken diagnosis is improved effectively, accuracy of fault diagnosis is enhanced, and it can be used for DMFD in complex nuclear power plant.

关 键 词：故障诊断 模拟退火算法 概率因果模型 

分 类 号：U672[交通运输工程—船舶及航道工程]