AN INTEGRATED APPROACH FOR IMAGE-BASED COMPUTATION BASED ON THE X-FEM: ACCURACY AND HIGH-ORDER APPROXIMATION

International audienceIn this contribution an integrated strategy is proposed for image-based computations. Indeed, it is still difficult to handle the amount of geometrical information contained in the images produced by modern acquisition techniques such as microtomographs. Classically, voxel-based finite element models are used as the model can be automatically constructed by converting each voxel into a finite element. This approach leads to huge numerical models and a poor geometrical description if the resolution is not sufficient. An integrated computational approach was presented in [8] in order to circumvent these limitations. The proposed strategy is based on the X-FEM and levelsets in order to avoid the meshing of complex geometries. More recently [9], the approach was improved in order to uncouple the computational model from the geometrical one. Validations against the voxel-based Finite Elements is 2D and 3D proposed in order to assess the accuracy of the approach