Modeling strength and failure variability due to porosity in additively manufactured metals

Sensitivity of void mediated failure to geometric design features of porous metals

“…We varied the notch radius R in the range 0.05-0. to reproduce the experimentally observed average porosity φ = 0.008 and spatial correlation length ≈ 0.05 mm as described in detail in Ref. [9]. The largest pore in the ensemble was 81 elements (mean radius of 0.13 mm), and the average pore size was 4.1 elements (mean radius of 0.05 mm) across 10,000 realizations.…”

“…Thus it is difficult to represent the effects of this porosity with a bulk quantity such as overall material density. As in our previous work [9], we employ a model of distributed porosity which accurately reproduces the sizes and spatial correlations of defects observed in the material. This two-level model represents pores above a given size threshold explicitly in the finite element meshes, and sub-threshold pores are modeled with an initial damage field.…”

“…In the previous study [9] we calibrated an isothermal variant of the Bammann-Chiesa-Johnson (BCJ) viscoplastic damage model [25,26,27,28] to experimental data on AM 17-4 PH stainless steel tension specimens loaded to failure. We employ that calibrated model to represent the deformation and failure behavior of the specimens in this study.…”

“…As discussed in Ref. [9], fixing the element size across all the mesh realizations ameliorates the impact of mesh dependence due to softening on the quantities of interest.…”

“…In our previous work [9] we developed a model of porous metal deformation and failure tuned to experimental data from AM 17-4H steel [10]. That work employed Bayesian techniques to represent the observed variable mechanical response due to porosity and other microstructural effects.…”