摘要
针对基于主元分析的过程性能监控方法没有充分利用主元模型信息,以及基于SPE的信息重构故障诊断方法只利用了残差空间信息的局限性,通过定义故障子空间,对基于T2统计量的主元空间故障数据重构技术进行分析,得到故障可完全重构的条件及指标,从而计算出故障识别指标。将其应用于发酵过程的传感器故障识别,结果表明,该算法能够有效地找到故障源。
The limitations of PCA-based statistical monitoring approach omitting the principle model information,and that of the SPE-based fault diagnosis employing residual spatial information only were analyzed,together with the T2 statistic-based fault reconstruction technology by defining fault subspace,then,the theoretical conditions and index of reconstruction and identifiability were obtained.The results show that the approach applied to fermentation process can identify sensor＇s fault source effectively.
出处
《化工自动化及仪表》
CAS
北大核心
2011年第1期44-47,共4页
Control and Instruments in Chemical Industry
基金
国家科技部"863"计划项目(2007AA10Z241)
关键词
发酵过程
主元空间(PCS)
故障重构
故障识别
fermentation process
principal component space（PCS）
fault reconstruction
fault identification
