Design automation for polymer containers using efficient global optimization
A machine-learning-based solution for optimizing polymer container design using efficient global optimization, enhancing production yield and design efficiency.
Overview
This solution leverages machine learning to automate the design process for polymer containers by developing a surrogate model that integrates prior design data. Through efficient global optimization (EGO), feasible design solutions are efficiently identified, retaining design knowledge and applying optimization rules to meet specific requirements. The method has been successfully applied in injection molding gate design and polymer lens assembly, significantly improving production yields.
Technical specifications
- Machine Learning Integration: Utilizes prior design data to create a surrogate model of physical systems.
- Efficient Global Optimization (EGO): Searches for optimal design solutions by applying various optimization rules.
- Successful Applications: Proven in injection molding gate design with Moldex3D and polymer lens assembly, achieving production yields over 90%.
- Process Steps: Includes a feasibility study, model training, and testing, with plans for large-scale model application and transfer learning.
Technology readiness level
This solution has reached Technology Readiness Level 8, indicating it has been validated in an operational environment and is ready for full-scale deployment.
About University of Cincinnati
The University of Cincinnati is a comprehensive public research university with an applied, urban-serving character and a significant clinical enterprise. Industry engages through one of the nation's largest cooperative education programs, placing students year-round with corporate R&D and operations teams and creating an on-ramp to sponsored research. An innovation district near campus hosts co-located corporate labs, startup space, and shared prototyping facilities, while the university's integration with a major hospital system enables clinical studies and translation. Research is supported by competitive federal funding from agencies such as NIH and NSF, along with state and industry partnerships. A dedicated technology transfer office manages IP, licensing, corporate agreements, and startup formation, providing flexible models for collaboration.
Inquire about this partnering listing
Industry applications
delivered to your inbox.