Image-Based Peridynamic Modeling-Based Micro-CT for Failure Simulation of Composites.

Area of Science:

• Materials Science
• Computational Mechanics
• Image Analysis

Background:

• Computed tomography (CT) provides detailed material insights.
• Computational mechanics enables cost-effective structural response prediction via numerical simulation.
• Carbon-silicon carbide (C/SiC) composites require accurate failure analysis.

Purpose of the Study:

• To numerically simulate the tensile cracking behavior of C/SiC composites.
• To develop an image-based peridynamics (IB-PD) model for failure simulation.
• To integrate deep learning for efficient material identification and modeling.

Main Methods:

• Utilized 3D CT data to create geometric models.
• Employed a deep learning-based image recognition model for material identification.
• Applied the bond-based peridynamics (BB-PD) model for numerical simulation of cracking.

Main Results:

• The IB-PD model accurately reconstructed the composite microstructure from CT data.
• Simulations effectively predicted interfacial debonding and crack propagation influenced by defects.
• The model demonstrated capability in simulating matrix damage.

Conclusions:

• The proposed IB-PD approach offers an accurate and efficient method for simulating composite failure.
• This technique reduces the need for expensive experimental testing.
• The study highlights the potential of integrating CT, deep learning, and peridynamics for material analysis.