Ai-driven predictive model to optimize injection molding cycle times for PCR PET preforms
Developing an AI-driven predictive model to optimize injection molding cycle times for PET preforms with varying post-consumer recycled (PCR) content. The approach uses machine learning to learn relationships between PCR content, material properties, and molding parameters, integrating process data such as temperature, pressure, injection speed, and cooling rates. A hybrid data-driven and expert-guided optimization strategy supports parameter selection using a simplex experimental setup and reinforcement learning decision model.
Overview
An AI-driven predictive model is being developed to optimize injection molding cycle times for PET preforms with varying levels of post-consumer recycled (PCR) content. The goal is to improve manufacturing efficiency, reduce costs, and support more sustainable production practices. By leveraging machine learning, the model captures complex relationships between PCR content, material properties, and injection molding parameters that are difficult to represent with traditional methods.
Technical specifications
The development combines machine learning with process and materials data to support cycle-time optimization. Key features include:
- Utilization of machine learning to analyze complex datasets involving PCR content and injection molding parameters
- Integration of process data such as temperature, pressure, injection speed, and cooling rates
- Hybrid modeling approach combining data-driven predictions with expert-guided optimization
- Use of a simplex experimental setup and a reinforcement learning decision model for process parameter optimization
The model is validated through comprehensive data collection and experimentation in injection molding facilities, with the intent to ensure robustness and accuracy.
Technology readiness level
The AI-driven model is currently at Technology Readiness Level 6, indicating it has been demonstrated in a relevant environment. Further validation and refinement will be conducted through targeted experiments, leveraging historical data and additional domain insights.
About Fraunhofer USA
Fraunhofer USA is an independent, non-profit applied R&D organization with a network of centers across the United States, backed by the global Fraunhofer network. Many facilities are co-located on research university campuses and in manufacturing regions, enabling access to shared equipment, specialized labs, and talent close to supply chains. Its contract research model emphasizes rapid scoping, prototyping, pilot-scale demonstration, and validation in industry-relevant settings, giving partners a low-risk path from concept to implementation. Research is supported by industry sponsorships and competitive federal and state funding from agencies such as the U.S. Department of Energy and the Department of Defense, with industry-friendly IP terms and licensing to streamline collaboration.
delivered to your inbox.