The influence of defects on the elastic response of lattice structures resulting from additive manufacturing

Stiffness prediction for additively manufactured (AM) lattices is necessary for lightweight components design. For a given lattice structure, the Gibson and Ashby (GA) model can predict relative stiffness as a function of its relative density. However, volumetric porosity and surface roughness defects that are commonplace in AM lattices depreciate the quality of prediction made by the GA model. This is because such defects complicate the elastic behavior of lattice structures. In this work, a modified GA model is proposed that accounts for these defects. A Bayesian inferencing framework is constructed to delineate the influence of spatial distributions of these defects on resulting mechanical response. Principal component analysis (PCA) is used to identify differences between elastic strain fields (∊₁₁, ∊₂₂, and ∊₁₂) resulting from perfect and defective lattices. The insights obtained can provide a viable approach to predict the mechanical response of as-received AM lattices that are often defective, and thereby enable systematic approaches for their design.