Multiscale Progressive Failure Analysis of 3D Woven Composites

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Modeling

Background:

• Three-dimensional (3D) woven composites offer advantages over traditional ply-based materials but present significant modeling challenges due to complex multiscale geometries.
• Accurate prediction of mechanical behavior is crucial for the expanding application of these advanced materials.
• Existing modeling techniques often struggle with the intricate nature of 3D woven structures.

Purpose of the Study:

• To develop and apply an efficient, semi-analytical multiscale modeling procedure for 3D woven carbon-epoxy composites.
• To incorporate realistic microstructural features, including binder-tow disbonds and weft-tow waviness, into the model.
• To validate the model's predictions against experimental data for in-plane tensile and shear behavior.

Main Methods:

• Utilized efficient, semi-analytical micromechanical theories, avoiding traditional finite element methods.
• Employed a crack-band progressive damage model for the matrix to capture nonlinear responses.
• Integrated microstructural data from X-ray computed tomography (CT) and scanning electron microscopy (SEM) to define realistic dimensions and volume fractions.
• Incorporated observed microstructural defects like binder-tow disbonds and weft-tow waviness.

Main Results:

• The multiscale model demonstrated good correlation with experimental data for the in-plane tensile behavior of the 3D woven composite.
• The model successfully captured the less brittle nature observed in the in-plane shear response.
• Quantitative predictions for shear behavior were somewhat underpredicted compared to experimental results.

Conclusions:

• The presented semi-analytical multiscale modeling procedure is a viable and efficient alternative for analyzing 3D woven composites.
• The model's ability to incorporate microstructural details enhances its predictive capability for complex composite behavior.
• Further refinement may be needed to improve quantitative accuracy in shear response predictions.