Data-Driven Tailoring Optimization of Thermoset Polymers Using Ultrasonics and Machine Learning

Area of Science:

• Materials Science
• Polymer Science
• Non-destructive Testing

Background:

• Thermoset polymers are crucial for high-performance applications due to their durability.
• Traditional characterization methods are destructive, costly, and time-consuming.
• Optimizing thermoset properties requires efficient methods to understand manufacturing impacts.

Purpose of the Study:

• To develop a novel non-destructive, data-driven method for tailoring thermoset properties.
• To correlate manufacturing parameters (curing temperature, stoichiometry) with material properties.
• To enable efficient and reliable optimization of thermoset manufacturing.

Main Methods:

• Utilized ultrasonics to monitor curing kinetics via real-time sound speed measurements.
• Employed machine learning (k-nearest neighbors) to build predictive models.
• Manufactured and tested thermoset epoxy samples with varied curing conditions.

Main Results:

• Successfully predicted curing kinetics and final elastic properties using manufacturing parameters.
• Demonstrated the ability to predict manufacturing parameters from material properties.
• Established a correlation between curing temperature, stoichiometry, and material performance.

Conclusions:

• The developed non-destructive method efficiently optimizes thermoset tailoring and manufacturing.
• Ultrasonics and machine learning offer a powerful alternative to traditional characterization techniques.
• This approach facilitates precise control over thermoset properties for advanced applications.