1. Sensor fault detection and isolation of an industrial gas turbine using partial block-wise adaptive kernel peA
- Author
-
Mania Navi, Nader Meskin, and Mohammadreza Davoodi
- Subjects
Aeroderivative gas turbine ,0209 industrial biotechnology ,Engineering ,Dynamical systems theory ,business.industry ,Industrial gas ,02 engineering and technology ,Turbine ,Fault detection and isolation ,Adaptive kernel PCA ,Isolation ,Nonlinear system ,020901 industrial engineering & automation ,020401 chemical engineering ,Control theory ,Dynamical time-varying systems ,Principal component analysis ,Isolation (database systems) ,0204 chemical engineering ,business ,Fault detection ,Block (data storage) - Abstract
In this paper, sensor fault detection and isolation of nonlinear time-varying dynamical systems is investigated based on a fast partial block-wise adaptive Kernel Principal Component Analysis (KPCA) scheme. Using the proposed partial adaptive KPCA, faults are diagnosed perfectly and it is possible to prevail the shortcomings of the conventional KPCA and PCA methods. It is shown through simulation studies that the occurrence of sensor faults in the nonlinear dynamical model of an aeroderivative gas turbine can be detected and isolated effectively using the proposed approach. 1 2017 IEEE. This publication was made possible by NPRP grant No.5 - 574 - 2 -233 from the Qatar National Research Fund (a member of The Qatar Foundation). The statements made herein are solely the responsibility of the authors. Scopus
- Published
- 2017