Experiences with Process Fault Detection Methods via Parameter Estimation

Fault detection and fault diagnosis of technical processes become more important in the course of progressive automation. Computer based fault supervision methods are developed which allow the early detection and localization of process faults during normal operation. The use of process models enables to estimate process state variables and parameters which are influenced by faults. The contribution concentrates on fault diagnosis based on process parameters. The general procedure is outlined, comprising parameter estimation, feature extraction, fault decision and classification. Experimental results are given for the detection of several faults in an electrical drive/centrifugal pump set and an steam heated heat exchanger.

[1]  G. Geiger,et al.  Fault Identification of a Motor-Pump System Using Parameter Estimation and Pattern Classification , 1984 .

[2]  David Mautner Himmelblau,et al.  Fault detection and diagnosis in chemical and petrochemical processes , 1978 .

[3]  W. Goedecke.  Fault Detection in a Tubular Heat Exchanger based on Modelling and Parameter Estimation , 1985 .

[4]  L. F. Pau.  Failure Diagnosis and Performance Monitoring , 1986, IEEE Transactions on Reliability.

[5]  P. M. Frank,et al.  Entdeckung von Komponentenfehlern in dynamischen Systemen durch hierarchische Zustandsschätzung / Component failure detection in dynamic systems via hierarchical state estimation , 1986 .

[6]  C. Baskiotis,et al.  Parameter identification and discriminant analysis for jet engine machanical state diagnosis , 1979, 1979 18th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[7]  Alan S. Willsky,et al.  A survey of design methods for failure detection in dynamic systems , 1976, Autom..

[8]  Rolf Isermann,et al.  Process fault detection based on modeling and estimation methods - A survey , 1984, Autom..

[9]  David Singer,et al.  Augmented Models for Statistical Fault Isolation in Complex Dynamic Systems , 1985, 1985 American Control Conference.

[10]  Peter Young,et al.  Parameter estimation for continuous-time models - A survey , 1979, Autom..

[11]  R. Isermann.  Fehlerdiagnose mit Prozeßmodellen / Failure diagnosis by process models , 1984 .