摘要
针对传统的间歇过程监控方法,在建模时只利用正常工况下的数据,其故障诊断能力并不令人满意的问题,提出了多向Fisher判据分析(MFDA:MultiwayFisherDiscriminantAnalysis)方法,用于间歇过程的监控。该算法同时利用正常工况和故障条件下的数据进行建模,其故障诊断能力要优于MPCA(MultiwayPrincipalComponentAnalysis),在故障检测的同时也实现了故障的诊断。通过对实际工业链霉素发酵过程数据分析,表明该算法是可行的,可以获得较满意的故障诊断结果。
MFDA (Multiway Fisher Discriminant Analysis) is proposed in order to monitor batch process. Although MPCA(Multiway Principal Component Analysis) contains certain optimality properties in terms of fault detection, it is not as well-suited for fault diagnosis because it does not use the information about fault. On the contrary, MFDA takes into account the information between the different fault classes and has advantage over MPCA for fault diagnosis. The method is illustrated by a case study of industrial streptomycin fermentation process.
出处
《吉林大学学报（信息科学版）》
CAS
2004年第4期384-387,共4页
Journal of Jilin University（Information Science Edition）
基金
国家高技术发展计划(863计划)基金资助项目(2001AA413110)
关键词
Fisher判据分析
多向主元分析
统计过程控制
间歇过程
模式识别
fisher discriminant analysis
multiway principal component analysis
statistical process control
batch process
pattern recognition
