Quality-Relevant Monitoring and Diagnosis with Dynamic Concurrent Projection to Latent Structures

Qiang LIU, S. Joe QIN, Tianyou CHAI

Research output: Journal PublicationsJournal Article (refereed)peer-review

16 Citations (Scopus)

Abstract

In this paper, a data-driven dynamic concurrent projection to latent structures (DCPLS) approach is proposed for quality-relevant fault diagnosis of dynamic processes. First, a novel DCPLS algorithm is proposed for dynamic modeling which captures the dynamic correlations between quality variables and process variables. Quality-specific variations, process-specific variations, and quality-process covariations of dynamic processes are monitored respectively. Secondly, a multi-block extension of DCPLS is designed to compute the contributions according to block partition of the lagged variables, in order to help localize faults. Finally, the application results on strip-thickness relevant fault diagnosis for a practical cold rolling continuous annealing process (CAP) demonstrate the effectiveness of the proposed methods.
Original languageEnglish
Pages (from-to)2740-2745
Number of pages6
JournalIFAC Proceedings Volumes
Volume47
Issue number3
DOIs
Publication statusPublished - Aug 2014
Externally publishedYes
Event19th IFAC World Congress on International Federation of Automatic Control, IFAC 2014 - , South Africa
Duration: 24 Aug 201429 Aug 2014

Bibliographical note

ISBN: 9783902823625 <br/>This work was supported in part by the Natural Science Foundation of China (61304107, 61020106003, 61290323, 61333007, 61203102, 61104084), the China Postdoctoral Science Foundation funded project (2013M541242), the International Postdoctoral Exchange Fellowship Program, and the IAPI Fundamental Research Funds (2013ZCX04, 2013ZCX05).

Keywords

  • Dynamic concurrent projection to latent structures
  • Dynamic process modeling
  • Fault diagnosis

Fingerprint

Dive into the research topics of 'Quality-Relevant Monitoring and Diagnosis with Dynamic Concurrent Projection to Latent Structures'. Together they form a unique fingerprint.

Cite this