146. Integrated Data-Supported Smart Root Cause and Process Failure Analysis
Invited abstract in session MD-28: Human-AI Collaboration and Ethics, stream Decision Support Systems.
Monday, 14:30-16:00Room: Maurice Keyworth 1.03
Authors (first author is the speaker)
| 1. | Mahdi Mokhtarzadeh
|
| Industrial Systems Engineering and Product Design, Ghent University | |
| 2. | Sidharta Gautama
|
| Industrial Systems Engineering and Product Design, Ghent University | |
| 3. | Jorge RodrÃguez-EcheverrÃa
|
| Industrial Systems Engineering and Product Design, Ghent University |
Abstract
In production systems, a frequent disconnection exists between Process Failure Mode and Effects Analysis (pFMEA) and Root Cause Analysis (RCA) during failure analysis, leading to the inefficient utilization of the causal knowledge documented in pFMEA. Furthermore, despite advancements in technology, pFMEA risk analysis has remained subjective and expert-dependent.
This study introduces a data-supported integrated pFMEA-RCA framework that combines quantitative data from the production process with pFMEA qualitative expert knowledge to transform traditional static pFMEA into a connected, continuously updated system with autonomous reasoning capabilities. The framework constructs a failure mode tree based on pFMEA information, links quality control data to this tree, and utilizes a Beta-binomial Bayesian approach for updating and inferring from the tree. This improves failure analysis causality and objectivity and effectively bridges the gap between pFMEA and RCA.
The framework is validated through a case study, complemented by sensitivity analysis to demonstrate its robustness. Its key benefits include the ability to objectively prioritize critical failure modes and causes, identify overlooked issues and inconsistencies in pFMEA documents, support data-driven decision-making and automated analysis, produce computer-readable and structured pFMEA documents, and evaluate normal and abnormal operating conditions.
Keywords
- Analytics and Data Science
- Decision Support Systems
- Quality Management
Status: accepted
Back to the list of papers