| Resumo : |
The present work aims to contribute towards the understanding of the mechanical behavior of 3D-printed composites based on thermoplastic matrices reinforced by continuous fibers. The employed research methodology includes the experimental characterization of mechanical properties and the implementation of a computational homogenization method which predicts elastic mechanical properties and allows building numerical failure envelopes, assessing stresses at the microscopic level. Experimental mechanical properties (stiffness and strength) are evaluated in tension (longitudinal and transverse to fibers), compression (longitudinal), in-plane shear and interlaminar shear. A methodology is proposed and implemented using the commercial software MATLAB to computationally identify the fiber arrangement from cross-section micrographs of specimens and determine a statistically equivalent fiber distribution. From these results, a periodic equivalent volume is then adopted in the computational homogenization and also in the numerical failure analysis. The homogenized elastic properties, as well as stresses at microscopic level, are computed using the asymptotic homogenization technique. This technique is implemented according to a novel methodology with support of the commercial finite element software ABAQUS, using Python subroutines encoded into an Application Programming Interface (API). The micromechanical model is also employed to build mechanism based failure envelopes, which accounts for different failure criteria for matrix, fiber and interface. Additionally, an expansion of the Puck and Schürmann Inter-Fiber Fracture criterion (originally formulated to evaluate intrinsically brittle materials) is proposed, in order to provide more suitable analytical failure envelopes for 3D-printed fiber reinforced composite materials. The obtained results with both approaches are considered to be consistent. While the expansion of Puck and Schürmann Inter-Fiber Fracture criterion fits well the expected results on preliminary studies, the mechanism based failure envelopes are more suitable for advanced studies, where the effects of fiber arrangement must be assessed. Furthermore, the micromechanical analysis is useful whenever experimental bi-axial testing results are not available. |