Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

Explainable Prediction of the Mechanical Properties of Composites with CNNs

Created by
  • Haebom

Author

Varun Raaghav, Dimitrios Bikos, Antonio Rago, Francesca Toni, Maria Charalambides

Outline

This paper presents a deep learning model for predicting the mechanical properties of composite materials. To address the high computational cost of conventional finite element analysis (FE), we developed a model using a convolutional neural network (CNN) to predict the elastic modulus and yield strength of composite materials. Specifically, to overcome the limitations of previous studies, which typically employ a simple neural network structure (T30772), consider only elastic properties, and lack transparency, we applied explainable AI (XAI) techniques to enhance the interpretation and reliability of the model's prediction results. We trained the CNN using a dataset generated from transverse tensile tests and achieved higher accuracy than the ResNet-34 model. We validated the model's reliability by demonstrating that the CNN utilizes important geometric features that influence composite material behavior using SHAP and Integrated Gradients techniques.

Takeaways, Limitations

•
Takeaways:
◦
A novel method for accurately predicting the mechanical properties of composite materials by combining CNN and XAI techniques is presented.
◦
Contributes to solving the computational cost problem of existing FE analysis.
◦
Improve reliability and enhance engineers' understanding by interpreting the model's prediction results.
◦
Achieved higher prediction accuracy than ResNet-34.
•
Limitations:
◦
The current model is limited to certain types of composite materials and test conditions, and further research is needed to determine its generalizability.
◦
Model performance may be affected by the size and diversity of the dataset used.
◦
The interpretation results of the XAI technique may not always be clear.
◦
Generalization performance verification is needed for other types of composite materials or more complex mechanical behaviors.
👍
Made with Slashpage