Bead-resolved prediction of warpage, residual stress, and high-stress locations in metal-filled PLA additive manufacturing

Green metal-filled PLA parts are fragile in the as-printed state, and part-scale simulation tools cannot resolve where bead-level stresses concentrate. We present a bead-resolved thermo-viscoelastic eigenstrain framework that ingests a pre-computed thermal history and predicts the as-printed geometry, residual stress field, and highest-stress locations of green metal-PLA parts. The solver chain is validated against the ABS planar-spring benchmark of Cattenone et al., reproducing the eleven-probe experimental mean to −5.2% with a lower normalised RMSE (20.7% vs 23.5% of the experimental mean) than the benchmark’s Abaqus simulation, with two quantities calibrated to the benchmark and no rescaling. Predicted geometry matches 3D scans of IN718-PLA and Cu-PLA specimens to 0.07–0.10 mm typical lateral and 0.04–0.09 mm vertical deviation, measured against six separately fabricated specimens per material. A three-case sequential-isolation design recovers the closed-form cross-material stress ratio σ_Cu/σ_IN = (E_Cu/E_IN)·(α_Cu/α_IN), with α the along-raster contraction coefficient, to within 6% and verifies numerically that displacement is decoupled from modulus: a 2.2× modulus change alters the predicted geometry by under 0.002 mm. The bead-level map of elastic stress demand is robust to joint constitutive perturbation (ρ ≥ 0.98), and 195 of the top 200 highest-stress beads coincide between materials, which motivates single-simulation feedstock screening.