This research combines knowledge graphs with retrieval-augmented generation to enhance failure mode and effects analysis (FMEA) capabilities.

• Integrates domain-specific engineering knowledge with large language models
• Significantly improves factual accuracy in failure analysis processes
• Creates more robust FMEA frameworks for manufacturing ramp-up phases
• Addresses traditional LLM limitations in specialized engineering contexts

For engineering teams, this approach enables more reliable risk identification during product development, reducing costly failures and improving safety outcomes in industrial applications.

Knowledge graph enhanced retrieval-augmented generation for failure mode and effects analysis