The subject of this project is the 3D printing (SLS) of PA12/glass beads composite for applications in aerospace industry. The SLS process uses laser sintering of composite powder with polymer matrix containing glass beads. One of the limiting points of polymers composites for their use in aerospace systems is their durability, and more specifically their resistance to failure due to fatigue cracking. The objective of this project will focus on the study of finished products obtained by SLS of composites powders and their resistance to cracking. The objectives of this work are to understand failure mechanisms in these highly heterogeneous materials at two scales, the scale of the microsctructure and the scale of the workpiece, by combining experimental characterization of cracks networks by mechanical testing, 3D imaging by X-rays laboratory microtomography image analysis, and numerical simulations. The identified microstructural damage models will be used to construct a crack propagation model at the scale of the workpieces, and will account for specificities related to the material and the process: the highly heterogeneous nature of the microstructure and its strong anisotropy due to the layered structure obtained by SLS. Then, it will be used to optimize the process parameters and the shapes of products in the design step. Up to now, the damage mechanisms in compounds obtained by SLS 3D printing are not very well understood, even less for products obtained from composite powders. The objectives imply several challenges related to the numerical simulation of complex crack networks in highly heterogeneous materials, the detection of micro cracks by 3D imagery imaging within combined with in situ mechanical testing, the modelling of damage and its identification at both micro and macro scales. The mechanical parameters, including the damage ones, will be characterized at the micro and macro scales by approaches combining tomography within microstructures (damage at the interfaces, damage related to the layered structure of the material) or at the scale of the workpiece, and numerical simulations through inverse approaches. The studied material is obtained from composite powder made of a polymer matrix of PA12 and containing glass beads. The powder is then sintered by laser to obtain 3D workpieces by PRISMADD. This project will allow optimizing the process parameters of the 3D process and the geometries of the workpieces with respect to failure criteria and lightweight. A numerical simulation code working able to capture damage mechanisms at both microscopic and macroscopic scales will be developed, based on the phase field method. This technique allows modelling initiation, propagation and merging of complex 3D crack networks in heterogeneous media. The method will be extended to the behaviour related to the material, characterized by a strongly nonlinear anisotropic behaviour. The tasks will consist into: (a) developing an efficient modeling numerical framework for simulating complex networks of cracks in highly heterogeneous microstructures from voxel models such as those arising from X-rays computed micro tomography imaging (XRµCT) and at the scale of the workpieces; (b) manufacturing by SLS 3D printing samples for a set of controlled process parameters; (c) characterize the strength properties of the new manufactured materials, with both macroscopic experimental mechanical testing and imaging at microscale, based on in situ mechanical testing in imaging devices and full-field kinematic measurement techniques, in 2D (optical observation) and in full 3D (XRµCT) ; (d) proposing microstructural and macroscopic damage models, identifying them by the mentioned experiments, and developing simplified multiscale damage models for bridging micro and macro damage; (e) optimizing the process parameters and the geometries of the produced workpieces with respect to the strength resistance of the produced products.