Fault Detection and Diagnosis by Integrated Principal Component Analysis

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.