Fast Fatigue Life Prediction of Polymers Through Combined Constitutive Mathematical and AI-Based Modeling

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Science

Background:

• Fatigue life prediction is crucial for material design, but experimental testing is costly and time-consuming.
• Existing constitutive fatigue models integrate low-cycle and high-cycle behavior using damage evolution and a moving endurance surface.
• Polymers often exhibit S-N curve characteristics similar to metals, allowing parameterization via Coffin-Manson-Basquin models.

Purpose of the Study:

• To develop a computationally efficient method for predicting high-cycle fatigue life in polymers and ductile materials.
• To leverage machine learning (ML) for accelerating fatigue life predictions.
• To reduce the dependency on extensive experimental fatigue testing.

Main Methods:

• Developed a constitutive fatigue model integrating low-cycle and high-cycle behavior with a moving endurance surface.
• Expressed model parameters using Coffin-Manson-Basquin parameters, applicable to polymers.
• Utilized the constitutive model to generate high-quality training data for machine learning models.

Main Results:

• The combined constitutive modeling and ML approach significantly improves computational efficiency for high-cycle fatigue prediction.
• The model successfully generates data for training ML models, offering a viable alternative to traditional methods.
• The framework expedites the high-cycle fatigue design process for polymers and other ductile materials.

Conclusions:

• An advanced modeling framework combining constitutive mathematical modeling and AI accelerates high-cycle fatigue design.
• This approach offers a computationally efficient and data-driven alternative to experimental fatigue testing.
• The study demonstrates the potential of integrating physics-based models with ML for material design optimization.