Multiscale Fracture Simulations for Composite Materials

2016 
For failure in composite materials the fracture mechanisms at the microscopic scale, such as matrix cracking, fiber cracking, fiber-matrix debonding or combinations thereof determine the growth of crack at the macroscopic scale. The goal is to develop a simple failure criterion in terms of a bilinear Traction-Separation Law (TSL) for unidirectional Carbon-Fiber Reinforced Plastic (CFRP) that represents the crack nucleation and evolution at the macroscopic scale. The microscopic fracture mechanisms are incorporated by calibrating this bilinear TSL for arbitrary loading conditions and material properties. The calibration is performed by an extensive parametric study on 2D a Representative Volume Element (RVE) in ABAQUS. For the parametric study five load cases are studied: uni-axial extension (pure mode I), simple shear (pure mode II) and three mixed loading cases. Fracture is simulated inserting cohesive elements along the boundaries of all bulk elements. The effect of five variables was studied: interface cohesive strength, fiber cohesive strength, interface fracture energy, fiber fracture energy and fiber volume fraction. An average of four realizations, i.e. fiber distributions, resulted in an effective TSL that represents the fracture behavior of the RVE. The resulting effective TSL is approximated by a bilinear TSL that has three parameters: initial slope, cohesive strength (peak) and fracture energy (area). Using the results of the parametric study a correlation is performed for these three parameters. First, the mode I and II parameters are correlated using the results of the uni-axial extension and simple shear load case. Secondly, the three bilinear TSL parameters dependence on the mode mixity is correlated using one of the mixed loading cases. As a result of this procedure simple closed-form expressions are established that can determine the macroscopic fracture mechanism, in terms of a bilinear TSL, as a function of the microscopic properties.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    1
    Citations
    NaN
    KQI
    []