A RippleNet‐Based Recommended Diagnostics Method Based on Knowledge Graph Integrated Human‐Machine‐Environment
Chuanchao Su, Fuhong KUANG, Peng Hou, Xiaojian Yi, Feng Liu
Quality and Reliability Engineering International
Problems Identified (4)
Limited-factor fault knowledge graphs: Traditional fault diagnosis knowledge graphs focus mainly on machine factors and omit broader human and environmental factors.
Simple KG recommendation techniques: Existing fault diagnosis recommendation methods rely on simple embedding or path-matching techniques that struggle with complex diagnostic needs.
Limited-factor fault knowledge graphs: Traditional fault diagnosis knowledge graphs focus mainly on machine factors and omit broader human and environmental factors.
Simple KG recommendation techniques: Existing fault diagnosis recommendation methods rely on simple embedding or path-matching techniques that struggle with complex diagnostic needs.
Proposed Solutions (5)
Human-machine-environment fault knowledge graph: The paper constructs a multilevel equipment fault knowledge graph that integrates human, machine, and environmental factors.
RippleNet recommended diagnostics algorithm: The paper develops a RippleNet-based recommended diagnostics algorithm that combines historical diagnostic information with fault knowledge graph structure using iterative propagation.
Hybrid embedding-path recommendation: The proposed model integrates embedding-based and path-based recommendation approaches for fault diagnosis and recommendation.
Human-machine-environment fault knowledge graph: The paper constructs a multilevel equipment fault knowledge graph that integrates human, machine, and environmental factors.
RippleNet recommended diagnostics algorithm: The paper develops a RippleNet-based recommended diagnostics algorithm that combines historical diagnostic information with fault knowledge graph structure using iterative propagation.
Results (3)
Improved fault diagnosis performance:
Broad application potential:
Improved fault diagnosis performance:
Research Domain
Industrial equipment fault diagnosis using knowledge graphs and recommendation algorithms