Optimized forecast components-SVM-based fault diagnosis with applications for wastewater treatment

Hongchao Cheng, Yiqi Liu, Daoping Huang, Bin Liu

Research output: Contribution to journalArticle

2 Citations (Scopus)
3 Downloads (Pure)

Abstract

Process monitoring of wastewater treatment plant (WWTP) is a challenging industrial problem, due to its exposure to the hostile working environment and significant disturbances. This paper proposed a novel fault diagnosis method, termed as optimization forecast components-support vector machine (OFC-SVM). The method firstly improved the forecastable component analysis (ForeCA) for feature extraction. Secondly, in order to further enhance the method, the quadratic Grid Search (GS) algorithm is utilized to optimize the parameters of the proposed method. Thirdly, to properly evaluate the method performance, a new evaluation index is proposed, named Pre Alarm Rate (PAR), aiming to achieve the quantitative trade-off between false alarm rate (FAR) and missed alarm rate(MAR). Then, the new ROC curve can be further derived by PAR. Finally, the performance of OFC-SVM is strictly compared with other five methods as well as validated by a Monte Carlo model and a full-scale WWTP.
Original languageEnglish
Pages (from-to)128534-128543
Number of pages10
JournalIEEE Access
DOIs
Publication statusPublished - 3 Sep 2019

Keywords

  • fault diagnosis
  • grid search (GS)
  • feature extraction
  • forecastable component analysis
  • support vector machine
  • wastewater treatment

Fingerprint Dive into the research topics of 'Optimized forecast components-SVM-based fault diagnosis with applications for wastewater treatment'. Together they form a unique fingerprint.

Cite this