Title :
Research on fault detection of FPSO production process based on the kernel function
Author :
Xiaopeng Hao ; Dahua Li
Author_Institution :
Tianjin Key Lab. for Control Theor. & Applic. in Complicated Syst., Tianjin Univ. of Technol., Tianjin, China
Abstract :
FPSO(Floating Production Shortage and Offloading) production process fault detection method based on a combination of kernel functions with multivariate statistical is introduced. Kernel function mapping the nonlinear problem is converted into a high-dimensional linear feature space analysis and mapping data in high-dimensional space to extract the nonlinear characteristics to identify the fault. The experimental results show that the kernel function combined with multivariate statistical methods for the FPSO production process fault detection are very effective.
Keywords :
crude oil; fault diagnosis; independent component analysis; learning (artificial intelligence); oil technology; principal component analysis; process monitoring; production engineering computing; support vector machines; FPSO production process; crude oil; data mapping; fault detection method; fault identification; floating production shortage and offloading production process; high-dimensional linear feature space analysis; independent component analysis method; kernel function; multivariate industry process monitoring; multivariate statistical method; nonlinear characteristic extraction; nonlinear problem mapping; offshore oil engineering; principal component analysis method; support vector machine learning algorithm; FPSO; KICA; KPCA; fault detection;
Conference_Titel :
Information Science and Control Engineering 2012 (ICISCE 2012), IET International Conference on
Conference_Location :
Shenzhen
Electronic_ISBN :
978-1-84919-641-3
DOI :
10.1049/cp.2012.2450