A Study of the Method for Supporting an Analysis of the Latent PSFs in the Potential Incidents – Development of the Prototype System for Auto-extraction/Assessment of Latent PSFs that are Included in the Report Documents

Kazuki Kiyota, Yusaku Okada

Research output: Contribution to journalArticle

Abstract

In enterprises, industrial accidents, malfunction of the quality, and abnormality of the system should be solved rapidly by analyzing mechanisms of generating them. However, some problems exist, so that information of potential incidents is not utilized for preventing accidents in the future effectively. There are some problems. Then, this study made a latent PSFs list and devised a prototype system to assist people to analyze potential incidents and provide some advice about safety activities. These were done in the following order: 1) Practicing the factor analysis to 1003 incident reports in a petrochemical plant from the perspective of human factors 2) Setting 127 PSFs based on the analysis 3) Classifying them into three categories related nature of the operation and into nine categories related the operator or environment of the operation 4) Setting keywords which correspond to each PSF 5) Writing “countermeasure”, “method to diffuse a countermeasure”, “method to determine the effect of countermeasure” to each PSF According to the above process, we drew up PSFs table which was the key of subdividing and analyzing potential incident reports. And then, based on this table, we developed a prototype system for auto-extraction and assessment of latent PFSs that are included in the report documents. Now, we make trial run of this system to measure the effectiveness in a petrochemical plant. Here after we will consider the result and develop this system to more practical system.

Original languageEnglish
Pages (from-to)4302-4308
Number of pages7
JournalProcedia Manufacturing
Volume3
DOIs
Publication statusPublished - 2015

    Fingerprint

Keywords

  • Auto-extraction
  • Incident report
  • PSF

ASJC Scopus subject areas

  • Artificial Intelligence
  • Industrial and Manufacturing Engineering

Cite this