IPCA在故障检测与分离中的应用

Fault Detection and Diagnosis by Integrated Principal Component Analysis

摘要: 基于ＰＣＡ模型，综合利用Ｔ＾２检验值，Ｑ检验值和故障补偿等集成化主元分析（ＩＰＣＡ）方法对系统的工况故障和仪表故障进行检测和分离，并对仪表故障进行补偿，该方法既能进行故障检测，又能对不同的故障进行分离。仿真实例表明，这种方法是有效的。

Abstract: A new method, integrated principle component analysis (IPCA), of detecting faults of sensors and proecss, seperating them from each other and compensating for sensor faults is presented in this paper. It is based on the PCA model and employes such techniques of IPCA as T 2 check value, Q check value and fault compensation. Simulation results show that it is fairly effective.