基于特征子空间的滑动窗PCA在批过程故障诊断中的应用

基于传统的多向主元分析MPCA(multiway principal component analysis)常会导致误诊断,且对批过程难以保证在线状态监测和故障诊断的实时性,提出了一种基于特征子空间的滑动窗主元分析方法。在实时故障监测与诊断时,该方法采用适当大小的滑动窗逐步更新当前子数据空间,对当前子数据空间故障的识别通过依次计算其与基底库中各故障的匹配度来进行。这种方法克服了传统的MPCA不能处理非线性过程和实时性问题,并避免了MPCA在线应用时预报未来测量值带来的误差, 提高了批过程性能监测和故障诊断的准确性。

Batch processes fault detection based on characteristic subspace moving window PCA

A characteristic subspace moving window principal component analysis for on-line batch process monitoring and fault detection was proposed. Using proper moving window to update current data subspace and calculating matching degree between the current data subspace and each fault belonged to fundus warehouse step by step, this approach recognizes the current data subspace fault and emphasizes particularly on-line process performance monitoring and exactly fault detecting which results in extraordinary behavior of batch processes.