Numerical investigation of crack propagation in composite materials with manufacturing defects

The thesis is dedicated to the study of the effects of voids on the mechanical properties of a ±45° unidirectional glass fiber reinforced composite. Voids are a widely studied defect on composite materials, as they are very common and have a significant detrimental effect on the mechanical properties. Also, as their total elimination is a difficult and expensive process, they need to be tolerated and their effects need to be accounted for. The reduction in mechanical properties, such as young’s modulus and ultimate stresses, is due to the fact that, at the voids tip, a stress intensification phenomenon arises, and facilitates the creation of a crack. The lack of a quantitative analytical characterization in literature of this effect, is due to the statistical variability of the voids shape, size, location and volume fraction, as they all play a non negligeable role in the stress intensification phenomenon. This opens the possibility of a numerical study, based on finite element method. This thesis proposes a procedure for such numerical study. The procedure is based on a comparison of the real microstructure of a specimen with an ideal one, reconstructed by eliminating all the defects. The microstructure of the real specimen has been reconstructed via a CT scan of the material. The simulations have been done using the LS-Dyna software. The object of the simulation was a representative volume element (RVE), which is the smallest volume of the composite that is still capable to reproduce the mechanical properties of the whole specimen. To account for the statistical nature of the problem, various RVE’s have been subjected to the simulation. The set up of the simulation, which involves boundary conditions, material properties, type of solver, material failure criteria and more, has been chosen starting from theorical assumptions, empirical data, and widely used methods in finite element simulations of RVE’s. By comparing different possibilities and analysing them, a final trade off between accuracy and speed of the solution has been chosen. The real specimen has been also subjected to a tensile tests, an the results were used to check the correctness of the numerical solution. Once the definitive simulation set up was confirmed, the comparison of various RVE’s with the ideal RVE has been done, so that the statistical difference between different types of defects has been taken into consideration.