主元空间中故障可重构性、可分离性研究  

Fault reconstruction and separability research in principal component subspace

作　　者：肖应旺[1] 陈呈国 黄业安[2] 刘冬杰[3] 杨军[1] 姚美银[1] 

机构地区：[1]华南师范大学南海校区软件学院,广东佛山528225 [2]华南师范大学计算机学院,广东广州510631 [3]华南师范大学教育信息技术中心,广东广州510631

出　　处：《计算机与应用化学》2015年第6期733-738,共6页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(61174123);广东省自然科学基金资助项目(9151063101000043)

摘　　要：基于主元分析(Principal Component Analysis,PCA)的统计过程性能监测尽管不依赖于精确的数学模型,但也限制了它的故障诊断能力。本文在故障子空间和PCA监测模型及故障重构技术的基础上,研究了基于T2统计量的故障诊断问题,获得了主元空间中故障可重构性、可分离性的必要充分理论条件。通过对双效蒸发过程的仿真监测,证实了所获理论结果的有效性;表明通过故障重构不仅为故障识别提供了基础,而且重构故障幅值波形还为判断传感器故障类型提供了依据。Nevertheless, the most significant advantage of principal component analysis（PCA） is that no precise process model is needed, the development of capacity for fault diagnosis based on PCA is restricted to some extent. Based on fault subspace, PCA monitoring model and fault reconstruction technology, the fault diagnosis issue is explored by using T2 statistical index and the necessary and sufficient theoretical conditions of fault reconstruction and separability in principal component subspace are obtained. Through the simulation monitoring of double-effect evaporator process, these results show that the acquired theoretical results are effective. These simulation results also show that fault can be effectively identified by using the fault reconstruction technology and sensor fault type can be decided by using reconstruction fault amplitude.

关 键 词：故障子空间和PCA监测模型 故障重构 主元空间 故障可重构性和可分离性 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]