Answers are grounded in this paper and cite their sources.
Purpose This study aims to propose and experimentally validate a voxel-based numerical procedure for predicting dimensional deviations in acrylonitrile butadiene styrene (ABS) parts manufactured by material extrusion. Design/methodology/approach The workflow combines finite-element simulation of the printing process with virtual inspection of the resulting geometry. A benchmark part, including external surfaces, circular holes, rectangular pockets and a square pocket, was manufactured, measured with a coordinate-measuring machine and compared with numerical predictions. A voxel-size sensitivity analysis was performed by varying the voxel size from 0.4 to 2.0 mm. Findings Experimental deviations from nominal dimensions ranged from −2.08% to −0.02%, confirming the tendency of ABS parts to shrink. For the finest voxel size, numerical–experimental differences ranged from −1.15% to 1.99%, and 8 out of 10 measured dimensions showed absolute errors below 0.25 mm. The 0.4 mm voxel size yielded the lowest mean absolute error (0.22 mm), but required approximately four times the computational time of the 2.0 mm configuration. Intermediate voxel sizes, especially 1.2 and 1.6 mm, showed slightly higher errors but lower computational cost. Research limitations/implications This voxel-based approach was validated only for ABS parts via material extrusion. An isotropic material formulation and simplified thermal boundary conditions were adopted. Therefore, the process-induced anisotropy of the printed ABS component was not explicitly calibrated. Future work should include direction-dependent material characterisation and validation across different geometries and printing conditions. Future work will test complex geometries and temperature-dependent properties. The study provides evidence that voxel-based simulations can support dimensional prediction in MEX, though further validation across different materials, geometries, printers and process conditions is required. Practical implications This numerical tool can help industrial practitioners predict dimensional deviations in material extrusion parts prior to printing. The workflow provides practical guidance for selecting voxel resolution by balancing prediction accuracy, feature complexity and computational cost. Originality/value The individual finite element, voxel-discretisation and virtual-inspection methods are established. The originality of this study lies in its integration within a metrology-oriented workflow that compares simulated and CMM-measured geometries using equivalent feature-reconstruction criteria. Rather than merely confirming that smaller voxels improve geometrical representation, this study demonstrates that the resulting improvement in dimensional prediction is feature-dependent and may be non-monotonic. Within the investigated material, benchmark geometry, printer and process conditions, the results provide preliminary guidance for voxel-resolution selection by illustrating the trade-off between feature-level accuracy and computational cost.
No takes yet. Share an insight, caveat, or question.
Vendittoli et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: