CAD-feature enhanced machine learning for manufacturing effort estimation on sheet metal bending parts
TLDR
Hybrid approach enriches B-rep graphs with manufacturing features for sheet metal bending effort estimation, validated on synthetic and real industrial data.
Reasoning
The paper's strength lies in its hybrid method combining rule-based feature recognition with graph ML, and its real-world validation on industrial bending time data. Weakness is the narrow focus on sheet metal bending, limiting generalizability to other manufacturing domains.
Read-first score
Read-first score 58.9, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 54.
Field roles
Rank sensitivity
Stability: volatile; rank range: 41.
Keyword Scores
Deep Analysis
Innovations
- Hybrid approach enriching B-rep attributed adjacency graphs with manufacturing features recognized by a rule-based module
- Integration of process-specific features (bend characteristics, flange lengths, surface roles) as node attributes to focus learning on relevant geometric patterns
- Validation on a real-world industrial dataset with measured bending times, one of the first such validations on genuine production data
Methodology
The method enriches B-rep attributed adjacency graphs with manufacturing features (bend characteristics, flange lengths, surface roles) recognized by a rule-based module, then applies graph-based machine learning for manufacturability prediction and effort estimation. It is evaluated on a large-scale synthetic benchmark and a real-world industrial sheet metal bending dataset.
Key Results
Combining domain knowledge with graph-based learning improves prediction accuracy on both a synthetic manufacturability benchmark and a real-world industrial dataset with measured bending times.