摘要
运用嵌入约束方法,对故障诊断中的多指标信息融合问题作了统一的数学描述,系统地研究多指标诊断的信息理论、故障的可诊断性检验、诊断指标的筛选以及基于Bayes方法的诊断模型等,并将其应用于内燃机活塞缸套磨损故障的诊断,取得了预期的效果。
Using information theory and multivariate statistical theory, the theory and method of faults diagnosis based on multi index fusion are described mathematically, including the information theory of multi index diagnosis, test on possibility of fault diagnosis, selection of diagnosis index, Bayesion diagnosis modeling and so on. The methods are applicable not only in single sensor but also in multi sensor case. As an application, the diagnosis of I.C.engine piston liner wear condition is studied, including the test on possibility of diagnosis, selection of index, diagnosis modeling and so on. A practical diagnosis example is given, and the predictable effectiveness is achieved.
出处
《振动工程学报》
EI
CSCD
1999年第1期55-63,共9页
Journal of Vibration Engineering
基金
交通部科技进步"通达计划"资助
关键词
故障诊断
多元统计分析
活塞-缸套磨损
内燃机
faults diagnosis
data fusion
multivariate statistical analysis
Bayesion method
piston liner wear condition