Infill Topology Conditions the Process Response of FDM-Printed Lightweight Polylactic Acid: A Multi-Response Factorial Variance Decomposition

Process-parameter recommendations in material extrusion are often derived at one fixed infill geometry and then transferred to others, although such studies cannot estimate topology × process interactions. This study tests the transferability of process-parameter recommendations for lightweight polylactic acid (LW-PLA) using the phase 2 experimental campaign reported previously. A 3 × 2 × 2 factorial experiment (Cubic, Gyroid, Sub-Cubic; 230/240 °C; 70/80% flow) was reanalyzed by Type III sum-coded analysis of variance (ANOVA) with effect-size decomposition for ultimate tensile strength (UTS; N = 68), apparent density ( N = 59), and Charpy impact energy ( N = 71). For UTS, the Topology × Flow is the largest term [ η 2 = 33.6%; partial η 2 = 71.3%; F (2,56) = 69.53, p < 0.001], exceeding the flow main effect ( η 2 = 26.5%). Raising flow from 70% to 80% changes UTS by +1.020 MPa for Cubic, +0.703 MPa for Sub-Cubic, and −0.284 MPa for Gyroid; all pairwise interaction contrasts survive Holm correction. Density increases for every topology and most for Gyroid (+14.4%), despite its strength loss. Accordingly, marginal mass efficiency of the flow step is +68.0 MPa cm ³ /g for Cubic, +29.3 for Sub-Cubic, and −9.6 for Gyroid, inconsistent with a simple gravimetric explanation. In Charpy impact, topology is the only detectable systematic term [ η 2 = 41.7%, F (2,59) = 25.36, p < 0.001], with no detectable process main effect or interaction. Within the investigated 70–80% step, the tensile response therefore depends on topology; process recommendations should not be assumed to be transferable across architectures without interaction testing.