Title

Modeling of Fiber Pull-Out in Continuous Fiber Reinforced Ceramic Composites Using Finite Element Method and Artificial Neural Networks

Abstract

Finite element models for the debonding of a silicon carbide fiber (SiCf) embedded in silicon carbide matrix (SiC) are developed and analyzed. An axi-symmetric finite element model is developed to simulate the single fiber pull-out process and predict the load-displacement behavior in terms of fiber/matrix interface properties. A two-parameter cohesive damage modeling approach is coupled with a finite element model to simulate crack propagation during a fiber pull-out event. Effects of residual compressive stress acting across the fiber/matrix interface, and residual axial strain in the fiber, on fiber pull-out behavior are investigated. Poisson contraction of the fiber, which reduces resultant radial compressive stresses at the interface and interfacial frictional stress, is taken into consideration. Parametric studies are conducted to evaluate the effects of thickness of specimen, friction coefficient, interface toughness, and residual stresses, on load-displacement behavior. An artificial neural network model using a backpropagation algorithm is proposed to mimic the fiber pull-out and also approximate load-displacement behavior. A multilayer perceptron utilizing a nonlinear activation function is implemented in the neural network model. Analytical modeling and finite element models are used to train and test the proposed neural network model. The developed finite element and neural network models are validated using existing analytical models from the technical literature.

Department(s)

Mechanical and Aerospace Engineering

Second Department

Materials Science and Engineering

Keywords and Phrases

Single Fiber Pull-Out; Fiber/Matrix Interface; Cohesive Damage Model; Finite Element Model; Artificial Neural Networks

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2013 Elsevier, All rights reserved.

Share

 
COinS