Footprint of knowledge acquisition improvement in failure diagnosis analysis

Mohammad Yazdi*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

49 Citations (Scopus)

Abstract

Fault tree analysis (FTA) as an effective and efficient risk assessment tool are widely used to analyze the reliability of a complex system. In this context, FTA can properly improve the safety performance of the system by preventing an event which may lead to occurrence of a catastrophic accident. However, traditional FTA is still suffering from dynamic structure demonstration and importantly epistemic uncertainty processing. In this study, a novel methodology is introduced using Bayesian updating mechanism to deal with dynamic structure and 2-tuple fuzzy set named as intuitionistic fuzzy numbers are employed to cope with subjectivity of uncertainty processing. Accordingly, the most critical system components which affect the system reliability are recognized by using an appropriate sensitivity analysis method. The proposed methodology is then applied on a real case study application (a brake fluid filling system) in order to examine the effectiveness and feasibility of the approach. The results illustrated that the new methodology can have enough benefits for diagnosing the systems' faults compared with listing approaches of safety and reliability analysis. In terms of empirical case study, “electromotor failure” was evaluated as the second most critical basic event in conventional-based approaches, whereas in the novel methodology “high pressure liquefied material” was recognized as the second one.

Original languageEnglish
Pages (from-to)405-422
Number of pages18
JournalQuality and Reliability Engineering International
Volume35
Issue number1
DOIs
Publication statusPublished - Feb 2019
Externally publishedYes

Keywords

  • 2-tuple fuzzy set
  • automotive industry
  • fault tree analysis
  • subjectivity
  • tactic knowledge

Fingerprint

Dive into the research topics of 'Footprint of knowledge acquisition improvement in failure diagnosis analysis'. Together they form a unique fingerprint.

Cite this