论文部分内容阅读
基于主元分析(Principal Component Analysis,PCA)统计过程监控方法,由于其不需要数学模型,因此目前在过程监控领域获得了广泛应用,但这也限制了其在故障诊断方面的能力。针对此问题,本文从故障子空间与PCA监控模型的角度,利用故障重构技术,对基于PCA的T~2统计量进行重构,获得了主元子空间中T~2统计量的故障可重构性理论条件,提出了具体的故障识别指标和诊断算法,解决了基于主元子空间故障重构技术的故障诊断问题,弥补了Dunia等人的方法只在残差子空间中讨论故障重构与识别问题。通过对双效蒸发过程的仿真监控,表明了所获得的理论条件、故障识别指标和诊断算法能对传感器故障和过程故障进行有效地识别,证实了所获理论、识别指标和诊断算法的有效性。
Based on Principal Component Analysis (PCA) statistical process monitoring method, it has been widely used in the field of process monitoring because it does not require mathematical models, but this also limits its capability in fault diagnosis. To solve this problem, this paper reconstructs the T ~ 2 statistics based on PCA from fault subspace and PCA monitoring model using fault reconstruction technique and obtains the T ~ 2 statistics of fault in principal component subspace Reconfigurable theory conditions, a specific fault identification index and diagnosis algorithm are proposed to solve the problem of fault diagnosis based on principal component subspace fault reconstruction technology, to make up for the method of Dunia et al. Structure and identification problems. Through the simulation monitoring of the double-effect evaporation process, it shows that the theoretical conditions obtained, the fault identification index and the diagnosis algorithm can effectively identify the sensor fault and the process fault, and confirm the validity of the obtained theory, the identification index and the diagnosis algorithm .