摘要
以船用核动力装置为研究对象 ,建立了运行在仿真机上的全工况核动力装置运行仿真系统。利用智能专家控制理论 ,建立了能够管理整个装置运行的 ,能对典型运行故障进行检测与诊断的管理运行专家系统。此系统在出现故障时能及时调用知识库专家知识进行专家推理 ,分析核动力装置控制实际运行中典型故障产生的原因及解决方法、故障诊断处理具有实时性 ;利用神经网络理论 ,建立了神经网络故障检测与诊断专家系统 ,此系统将不断检测核动力装置各系统 ,并根据检测数据给出故障诊断结果。结果表明 ,在核动力装置三层智能控制结构下 ,利用神经网络故障检测与诊断专家系统对船舶核动力装置可能的运行故障进行险测与诊断 ,利用运行管理专家系统对核动力装置进行运行管理 ,提高了船用核动力装置的运行性能。
A marine nuclear power plant simulation system is founded, and a management expert system, which can administer and diagnose the typical faults, is constituted by the intelligent expert theory. This real time expert system can analyze the reason of the typical fault caused by the nuclear power plant practically running and give the best solvent by the expert knowledge reasoning in the repository; a neural network fault inspection and diagnosis expert system which can inspect every equipment continually and give the faults diagnosis result based on the inspective dates is established. Based on the three hierarchal architecture, the operation performance is improved very much by using of the neural network fault inspection and diagnosis expert system to inspect and diagnose the nuclear power plant faults and the management expert system to supervise the nuclear power plant running. The expert system research can give guidance for the marine nuclear power plant safety operation.
出处
《核科学与工程》
CAS
CSCD
北大核心
2002年第3期220-228,共9页
Nuclear Science and Engineering
关键词
船用核动力装置
仿真系统
专家系统
神经网络
故障诊断
运行管理
nuclear power plant
simulation system
expert system
neural network fault diagnosis
running management
