【作者】 张新荣； 徐保国；

【Author】 ZHANG Xin-rong XU Bao-guo College of Communication and Control Engineering,Southern Yangtze University,Wuxi 214122,China.

【机构】 江南大学 通信与控制工程学院；

【摘要】 基于 PCA 理论对故障进行重构,研究了基于 T^?统计量的故障诊断问题．首先将过程数据分别投影于主元子空间和残差子空间;然后依据各过程变量间的强相关性,采用最优化方法沿故障各自的方向对故障变量进行重构．在理论上分别获得了故障可检测性以及主元空间中故障可重构性、可分离性、可识别的条件．仿真研究结果表明,所获得的结果能对工业过程故障进行有效地识别．更多还原

【Abstract】 Faulty reconstruction based on PCA is presented,the fault isolation and identification issues are explored by using T~2 index.Firstly,data is projected onto principal component sub-space and residual sub-space respectively,then fault direction vector is defined via optimization.The theoretical conditions of isolatability and identifiability are obtained.The simulation research results show that the acquired theoretical results can effectively identify fault.更多还原