Linguistic assessment approach for hierarchical safety analysis and synthesis

Jun Liu, Luis Martínez López, Jian Bo Yang, Jin Wang

    Research output: Chapter in Book/Report/Conference proceedingChapter

    Abstract

    Engineering systems in industry are most often concerned with safety issues. Many of these systems are intended to work properly even in contexts where information is missing, incomplete or unreliable. This chapter introduces a safety model based on the concept of approximate reasoning for safety analysis. Parameters of the safety level, including failure rate, failure consequence severity and failure consequence probability, are all described by fuzzy linguistic variables. A fuzzy rulebase is used to capture the uncertainty and the non-linear relationships among these parameters. A safety estimate for possible causes of a technical failure can be obtained by the approximate reasoning approach. A safety synthesis is then applied to integrate all possible causes for a specific technical failure, or applied at the safety estimate made by a panel of experts. The synthesis is based on an ordinal fuzzy linguistic approach by means of a direct computation on linguistic values instead of the approximation approach by their associated membership functions. The use of the ordinal fuzzy linguistic approach makes the safety analysis more effective. Subsequently, the ranking and interpretation of the final safety synthesis of a concerned system are also described. Application of this proposed approach is demonstrated by a real-world case study in the offshore engineering. © 2008 Springer-Verlag Berlin Heidelberg.
    Original languageEnglish
    Title of host publicationStudies in Computational Intelligence|Stud. Comput. Intell.
    PublisherSpringer Nature
    Pages211-230
    Number of pages19
    Volume117
    DOIs
    Publication statusPublished - 2008

    Fingerprint

    Dive into the research topics of 'Linguistic assessment approach for hierarchical safety analysis and synthesis'. Together they form a unique fingerprint.

    Cite this