DocumentCode
3674124
Title
Fault detection in wastewater treatment plants using distributed PCA methods
Author
A. Sanchez-Fernández;M.J. Fuente;G.I. Sainz-Palmero
Author_Institution
Department of System Engineering and Automatic Control, EII, University of Valladolid, Spain
fYear
2015
Firstpage
1
Lastpage
7
Abstract
This paper proposes a distributed fault detection and diagnosis method based on Principal Component Analysis (PCA) in a whole plant monitoring scheme. The method is based on the decomposition of the plant into multiple blocks using plant topology. A local PCA based fault detection method is applied in each block and the results are sent to the central node to fuse the information and to detect and diagnose faults in the global plant. This method is compared with the centralized PCA method and some distributed principal component analysis (DPCA) methods in a wastewater treatment plant (WWTP). The objective is to check which of the distributed methods implemented is the best one in terms of detecting faults and minimizing the communication cost between the blocks. Empirical results on the WWTP show that the DPCA method based on local models has very good results.
Keywords
"Principal component analysis","Covariance matrices","Biological system modeling","Matrix decomposition","Fault detection","Wastewater treatment","Monitoring"
Publisher
ieee
Conference_Titel
Emerging Technologies & Factory Automation (ETFA), 2015 IEEE 20th Conference on
Type
conf
DOI
10.1109/ETFA.2015.7301504
Filename
7301504
Link To Document