Fault Detection and Diagnosis in Industrial Systems Advanced Textbooks in Control and Signal Processing

Citation

Rigelsford, J. (2001), "Fault Detection and Diagnosis in Industrial Systems Advanced Textbooks in Control and Signal Processing", Industrial Robot, Vol. 28 No. 5. https://doi.org/10.1108/ir.2001.04928eae.003

Fault Detection and Diagnosis in Industrial Systems Advanced Textbooks in Control and Signal Processing

Fault Detection and Diagnosis in Industrial Systems Advanced Textbooks in Control and Signal Processing

L.H. Chiang, E.L. Russell and R.D. Braatz Springer2001279 pp.ISBN 1-85233-327-8£24.50

Keywords: Publication, Fault analysis, Process control

This textbook presents the background theory and practical techniques required for successful process monitoring. Early and accurate fault detection and diagnosis can minimise downtime, reduce manufacturing costs, and increase the safety of modern manufacturing processes. It provides the reader with knowledge required to implement the right techniques for a particular application.

The book comprises of 12 chapters divided into five parts. Part I provides an introduction to process monitoring procedures, measures and methods, while the background theory is presented in Part II. This includes chapters discussing multivariate statistics and pattern classification, and addresses topics including data pre-treatment; univariate statistical modelling; the T2 statistic; discriminant analysis: and feature extraction. Part III, data-driven methods, comprises of four chapters discussing principal component analysis (PCA), Fisher discriminant analysis (FDA), partial least squares (PLS), and canonical variate analysis (CVA). Topics covered include: reduction order; fault detection, identification and diagnosis; dynamic PCA and FDA; a comparison of PCA and FDA; a comparison of PCA and RLS; the CVA algorithm; and subspace algorithm interpretations. Part IV, Application, addresses the Tennessee Eastman Process, application description, and results and discussion. These chapters discuss process flow sheets, variables and faults; data sets; sample sizes; and four case studies. The final part of the book presents analytical and knowledge-based methods, including observer-based methods; parity relations; causal analysis; expert systems; pattern recognition; artificial neural networks; and fuzzy logic.

Fault Detection and Diagnosis in Industrial Systems is a well written and informative text. It provides students and practitioners with the understanding of various process monitoring techniques to ensure the right method is used for a particular application. A plant simulator and homework problems are included to reinforce the concepts presented.

Jon Rigelsford