STRENGTH AND RUPTURE MODELING OF UNIDIRECTIONAL POLYMER COMPOSITES

2001 
With the increasing use of composite materials a detailed understanding of the micromechanical events that lead to material failure is necessary. In this work, numerical techniques for prediction of the static strength and rupture lifetime of carbon fiber/polymer matrix composites are developed. Statistical static strength and rupture lifetimes of Grafil 700 carbon fiber/Polyphenylene Sulfide (PPS) thermoplastic matrix are presented for purposes of comparison. The key ideas in developing micromechanical models for failure are the interaction between the fiber and matrix, its response to time and temperature, and the fiber strength distribution. Fiber/matrix interaction is studied through the use of model composites. Three-dimensional model composites, representative of unidirectiona l polymer matrix composites, are fabricated with hexagonal fiber packing. Strain gages mounted onto the fibers provide strain concentration measurements due to a break in one of the fibers. Three-dimensional finite element analysis is performed to model the load-sharing. There is very good agreement between the finite element analysis and experimental results. A comparison is also made between the load-sharing calculated by finite element analysis and that obtained by a modified Hedgepeth and Van Dyke (HVD) 'shear-lag' analysis. The HVD analysis yields greater stress concentrations on the unbroken neighboring fibers. The finite element loadsharing is incorporated into a Monte Carlo simulation for the static strength of carbon fiber/polymer matrix composite. The composite material Weibull distribution computed by the simulation has a location and shape parameter of 1.71 GPa and 47.3, respectively, at a length of 76 mm. The experimentally obtained Weibull location and shape parameters are 1.57 GPa and 29.4, respectively, at a length of 76 mm. The Monte Carlo approach for rupture modeling is briefly discussed.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    9
    References
    0
    Citations
    NaN
    KQI
    []