A BP Wavelet Neural Network Structure for Process Monitoring and Fault Detection

Author

Shi, Hongbo ; Huang, Chuang

Author_Institution

Res. Inst. of Autom., East China Univ. of Sci. & Technol., Shanghai

Volume

2

fYear

0

fDate

0-0 0

Firstpage

5675

Lastpage

5681

Abstract

Wavelet theory cooperated with neural networks was applied to chemical process monitoring in this contribution. A newly developed BP wavelet neural network (WNN) structure is presented. It was used to summarize the corrupted process information into a nonlinear dynamic mathematical model. The monitoring charts are based on the multivariable residuals derived from the difference between the process measurements and the wavelet network prediction. The effectiveness of the proposed method was tested through the study on Tennessee Eastman (TE) problem, and simulation was developed. The results show that the current performance of the process can be evaluated

Keywords

backpropagation; chemical engineering computing; neural nets; process monitoring; wavelet transforms; BP wavelet neural network; Tennessee Eastman problem; chemical process monitoring; corrupted process information; fault detection; multivariable residuals; nonlinear dynamic mathematical model; wavelet network prediction; Artificial neural networks; Chemical processes; Computerized monitoring; Control charts; Fault detection; Low-frequency noise; Neural networks; Principal component analysis; Process control; Wavelet transforms; BP wavelet network; Fault detection; Process monitoring; chemical process;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on

Conference_Location

Dalian

Print_ISBN

1-4244-0332-4

Type

conf

DOI

10.1109/WCICA.2006.1714162

Filename

1714162