摘要
针对歼击机的起飞、爬升阶段 ,数据量大且复杂 ,对故障诊断要求精度高 ,实时性好 ,设计出粗集和神经网络相结合 ,分层诊断的方法 ,先定故障的类型 ,然后定故障的度 .其中在粗集诊断部分 ,提出了适合该研究对象的离散和简约方法 ,从而得到了少量但足够用的决策规则 ,使得实时诊断程序结构简单 ,实时性好 .包含诊断和报警模块的实时程序证明 ,此方法可以达到精度和实时性要求 .
A diagnosis method based on rough set theory and neural network is proposed for take-off and climbing of fighter. In this phase, the data are large and complex and high precision and real-time diagnosis are required, therefore a scheme is formed to diagnose fault style first with rough set theory and then fault degre e with neural network. Appropriate discretization method and reduction method are chos en and a few but enough decision rules are obtained, which can make the real-ti me diagnosis program simple. Real-time program with diagnosis modu le and alarm module proves that this method can satisfy the requirements.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2004年第B11期131-134,共4页
Journal of Southeast University:Natural Science Edition
基金
国家自然科学基金重点资助项目 (60 2 3 40 10 )
航空科学基金资助项目 (0 2E5 2 0 2 5 )
国防基础科研资助项目 (K160 3 0 60 3 18) .
关键词
故障诊断
粗糙集
神经网络
起飞
爬升
fault diagnosis
rough set
neural network
take-off
climbing