ORS3-1016
Engineering Processable Matter: Accelerating Thermoplastic Materials Development with AI and Robotics
When and Where
Nov 30, -0001
12:00am - 12:00am
Presenter(s)
Maciej Haranczyk (IMDEA Materials Institute)
Co-Author(s)
Abstract
The development of advanced thermoplastic formulations is often hindered by costly trial-and-error experimentation and the complex interplay between composition, processing conditions, and resulting performance. While robotics and artificial intelligence are increasingly being adopted in materials development, efficient approaches for screening and optimizing formulations across multiple manufacturing processes remain limited.
In this work, we present a closed-loop Materials Acceleration Platform that integrates automated compounding, processing, characterization, computer vision, and machine learning to accelerate thermoplastic formulation development. Formulations are produced using a robot-assisted twin-screw extrusion workflow and subsequently evaluated through multiple downstream manufacturing routes, including injection molding and pellet-based extrusion additive manufacturing. Experimental data are combined with rheological and thermal characterization results to train machine learning models capable of predicting processability and product quality, enabling the early identification of unsuitable formulations and reducing costly experimental failures.
The platform further incorporates computer-vision-based quality assessment and active learning strategies to guide material selection and experimental planning. By continuously learning from experimental outcomes and focusing resources on the most promising candidates, the approach improves workflow efficiency and accelerates the exploration of composition–processing–performance relationships. The presented methodology demonstrates how the integration of AI, automation, and advanced characterization can support the rapid development of next-generation thermoplastic materials across diverse manufacturing technologies.
In this work, we present a closed-loop Materials Acceleration Platform that integrates automated compounding, processing, characterization, computer vision, and machine learning to accelerate thermoplastic formulation development. Formulations are produced using a robot-assisted twin-screw extrusion workflow and subsequently evaluated through multiple downstream manufacturing routes, including injection molding and pellet-based extrusion additive manufacturing. Experimental data are combined with rheological and thermal characterization results to train machine learning models capable of predicting processability and product quality, enabling the early identification of unsuitable formulations and reducing costly experimental failures.
The platform further incorporates computer-vision-based quality assessment and active learning strategies to guide material selection and experimental planning. By continuously learning from experimental outcomes and focusing resources on the most promising candidates, the approach improves workflow efficiency and accelerates the exploration of composition–processing–performance relationships. The presented methodology demonstrates how the integration of AI, automation, and advanced characterization can support the rapid development of next-generation thermoplastic materials across diverse manufacturing technologies.
