Skip to content

WP-02 Feature-Based Manufacturability Inference

Field Value
Priority High
Status Framing
Problem Space DFM and manufacturing reasoning
Linked Capability manufacturing physics, feature recognition, explainable review logic
Best Collaboration Shape applied thesis + pilot + rules study
Owner unassigned

What This Work Package Is

Map CAD and drawing features to manufacturability-relevant reasoning that can later support review comments, rule suggestions, and DFM guidance.

Why It Matters To RapidDraft

  • Directly supports the DFM product story
  • Connects geometry understanding to production-aware output
  • Can produce explainable logic rather than only black-box predictions

Expected Deliverables

  • candidate feature schema
  • manufacturability risk taxonomy
  • first rule or inference prototype
  • validation on a narrow manufacturing family

Candidate Lanes

Lane People or Group Why They Matter Best First Ask Priority
FAU KBE and design-methods lane Sandro Wartzack clearest public signal for KBE, engineering drawings, tolerancing, and decision reuse advisor call on feature schema and explainable DFM framing High
TUM CCBE AI4CADCAM Dr. Stavros Nousias, André Borrmann strong CAD feature recognition and downstream CAD/CAM reasoning signal thesis or mini-project around feature recognition for packaging machinery parts High
Hochschule Kempten IDF Frank Schirmeier and Ulrich Göhner direct public signal on graph neural networks on CAD structures and industrial GenAI applied mini-project tied to SOMIC or SME pilot framing High
TU Darmstadt PLCM Benjamin Schleich strong engineering automation and PLM-informed product reasoning lane later-stage thesis or grant-framing discussion Medium

Ideal Partner Profile

  • applied product-development or manufacturing-methods strength
  • comfort with explainable engineering logic
  • ability to work on a narrow domain rather than generic "AI for manufacturing"

Candidate Evidence Types

  • feature-based DFM work
  • engineering knowledge capture
  • process-aware CAD or product development research
  1. Pick one first manufacturing family before opening broad outreach.
  2. Use FAU first for schema and framing, then Kempten or TUM CCBE for applied implementation depending on the chosen vertical.
  3. Keep the first prototype focused on explainable risk logic, not full autonomous DFM.
  4. Treat Kempten as the strongest packaging-aware applied path once the SOMIC-specific angle is ready.

Open Questions

  • Which process family gives the best first collaboration wedge?
  • How should this package connect to standards-as-data work later?

Sources

  • TextCAD/04_Marketing and Outreach/13_Universities/deep-research-report Balanced.md
  • TextCAD/04_Marketing and Outreach/13_Universities/deep-research-report Monster.md
  • TextCAD/04_Marketing and Outreach/13_Universities/deep-research-report.md