Manufacturing | Product Design & Operational Excellence
Quality Control & Benchmarking of CAD/CAM 3D Designs using Deep Learning
Objective
An Automated Quality Management process for manufacturers to validate any shortcomings in 3D CAD designs, such as Holes (missing edges) or Intersections (overlapping faces), thereby reducing manual intervention in issue identification and resolution.
Challenges
- High manual effort in identifying and resolving design errors.
- Inconsistent quality control due to human error.
- Time-consuming process affecting production timelines.
Solution Proposed
- Leverage Deep Learning-based Graph Convolutional Networks (GCNs) to treat design errors.
- Perform geometric deformation to decide the best fit based on Minimum Intrinsic Dissimilarity.
- Implementing GCNs for their ability to handle complex geometric data and perform accurate mesh repairs.
- The innovation lies in the automated framework to repair and re-mesh 3D designs, significantly reducing manual intervention.
Outcome
- Reduction in manual intervention by 70%.
- Improved accuracy in design validation.
- Faster issue resolution with a notable reduction in time.
Design & Thinking Wins
- Automated framework to repair and re-mesh 3D designs.
- Enhanced quality control with minimal human error.
- Streamlined production process with faster turnaround times.
Disclaimer: The outline showcases the typical challenges, solutions, designs, and outcomes for industries and functions, in general, based on Affine’s prowess in the Industry. The outcomes would be much higher for specific clients as they would be based on their data and specific problems to be solved.
