Sensor fault detection and isolation of an industrial gas turbine using partial kernel PCA
Abstract
In this paper, partial kernel principal component analysis (PKPCA) is studied for sensor fault detection and isolation of an aeroderivative industrial gas turbine. Principal component analysis (PCA) is an effective tool for process monitoring task, however it can achieve acceptable results only for linear processes. In the case of nonlinear processes such as gas turbines, kernel PCA approach can be used which leads to more accurate health monitoring. In order to achieve fault isolation, partial KPCA is proposed where the parity relation concept is used to generate a set of residual signals. The simulation studies demonstrate that using the proposed methodology, the occurrence of sensor faults in an industrial gas turbine can be effectively detected and isolated.
Publication Title
IFAC-PapersOnLine
Recommended Citation
Navi, M., Davoodi, M., & Meskin, N. (2015). Sensor fault detection and isolation of an industrial gas turbine using partial kernel PCA. IFAC-PapersOnLine, 28 (21), 1389-1396. https://doi.org/10.1016/j.ifacol.2015.09.719
