Skip to main content
The ARC Training Centre for
Transforming Maintenance through Data Science
Presentations

The Failure Project - Semantic modelling of how processes and things fail

melinda-hodkiewicz

Virtual - Researchers Catch-up host online from Curtin University

Friday 15 March 2024

Engineered Systems failures happen often, however a rare few have significant consequences. Most failures have hidden costs, increased time to complete tasks, extra costs, more time a service is off-line, diversion of management attention.

Every organisation identifies failure events, classifies them, and uses this data for continuous improvement and capital allocation in their own way. But there is no agreement on how to describe failures across organisations and even within organisations the agreed classification codes are often problematic and poorly used. Hence there is no pathway to machine readable approaches to describing failures.

Melinda will describe a new project to develop an ontology for capturing failure related information held in FMEA, failure investigation, maintenance work order and notification texts. Her talk will describe the motivation for the project, the NLP-KG tools (many developed by CTMTDS researchers), our collaborators and some preliminary work.