Get access

Risk-based fault diagnosis and safety management for process systems

Authors

  • Huizhi Bao,

    1. Process Engineering, Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada A1B 3X5
    Search for more papers by this author
  • Faisal Khan,

    Corresponding author
    1. Process Engineering, Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada A1B 3X5
    • Process Engineering, Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada A1B 3X5
    Search for more papers by this author
  • Tariq Iqbal,

    1. Process Engineering, Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada A1B 3X5
    Search for more papers by this author
  • Yanjun Chang

    1. Process Engineering, Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada A1B 3X5
    Search for more papers by this author

Abstract

An innovative methodology of risk-based fault diagnosis and its integration with safety instrumented system (SIS) is proposed in this article. The proposed methodology uses control chart technique to distinguish abnormal situation from normal operation based on three-sigma rule and linear trend forecast. Time series moving average techniques are used to perform real-time monitoring and noise filtering in fault diagnosis processes. Furthermore, risk indicators are used to identify and determine potential fault(s) to minimize the number of false alarms. The proposed methodology is implemented in G2 development environment. Two case studies of a tank filling system and a steam power plant system with SIS1s and SIS2s are conducted in G2 environment. A technique breakthrough from univariate monitoring to multivariate monitoring for fault diagnosis has been achieved during the verification in the steam power plant system. © 2010 American Institute of Chemical Engineers Process Saf Prog, 2011

Ancillary