To decrease the weight of new cars, aluminium alloys are progressively replacing steels in parts of the "body-in-white". However, the use of aluminium alloys requires sometimes new joining techniques to replace classical welding-points. Self-piercing riveting (SPR) or clinching are two of these relatively new joining techniques in which joining comes from the materials plastic deformation. In this study, we focus on the numerical modelling of self-piercing riveting and its experimental validation.Various studies are dealing with the numerical simulation of riveting, Abe et al. [1], Porcaro et al. [2]. The originality of our approach, initially presented in Bouchard et al. [3], is the use of damage in the modelling of material behaviours (sheets and rivet). Throughout this paper, we will show the importance of taking damage into account.In order to achieve the numerical analysis of SPR process, we use the finite element software To determine the parameters of the isotropic hardening and of the damage law, we use an inverse analysis. The methodology is based on an evolutionary algorithm. Due to the number of parameters, we perform for each part two different mechanical tests to valid the identified parameters. Details are given in the article.In this paper, we use a 2D axisymmetric configuration to model the f f SPR process. When the computation is done, we validate the riveting process simulation thanks to the experimental loaddisplacement curves and the geometrical cuts of the samples.The second aim of this paper is to present the ability of our model to simulate the mechanical strength of the joined specimenunder static loading (tensile, shear and mixed solicitations). With FORGE 2005®, it is possible to export mechanical fields from a 2D axisymmetric mesh to a 3D mesh. The creation of the 3D sample is achieved in many steps using the final geometries of the SPR process. First, a 3D geometry is extrapolated from the 2D geometry. Then a new mesh is generated and the mechanical fields (residual stresses, damage, …) are transported using an interpolation technique. To finish, the 3D circular meshes are cut to have a rectangular numerical specimen closest to geometry of the experimental.As in Langrand et al. [5], the experimental data are obtained thanks to an ARCAN test illustrated in Fig. 1a. This specific device enables to mix and control tensile and shear loadings on a riveted cross-shaped specimen (Fig. 1b). The experimental campaign and the ARCAN test are detailed in this paper.