Modified Micro-Mechanics Based Multiscale Model for Damage Analysis of Open-Hole Composite Laminates under...

Area of Science:

• Composite Materials Science
• Mechanical Engineering
• Computational Modeling

Background:

• Existing multiscale models often simplify complex internal structures in composite materials.
• Accurate prediction of damage initiation and evolution is crucial for structural integrity.
• Fiber arrangement and interfacial properties significantly influence laminate behavior under load.

Purpose of the Study:

• To modify a multiscale model to incorporate complex internal structures like fiber random arrangement and interfaces.
• To develop a feed-forward neural network (FFNN)-based method for evaluating macroscale property degradation.
• To validate the enhanced model through progressive damage analysis of open-hole laminates under compression.

Main Methods:

• Modification of a micro-mechanics failure theory-based multiscale model.
• Inclusion of fiber random arrangement patterns and interfaces using a clustering method.
• Development of a feed-forward neural network (FFNN) for damage evolution and property degradation analysis.

Main Results:

• The interface inclusion leads to premature damage initiation in laminates.
• Fiber random arrangement patterns decrease predicted compression responses.
• The FFNN-based degradation method increases predicted compression strength, particularly for random fiber patterns (6.0% increase).

Conclusions:

• The modified multiscale model effectively captures the influence of complex internal structures on laminate behavior.
• Interfacial effects and fiber arrangement are critical factors in predicting composite damage.
• The FFNN-based approach provides a robust method for assessing property degradation and predicting enhanced compression strength.