Geometric deviation modeling with Statistical Shape Analysis in Design for Additive Manufacturing

Effective modeling of geometric deviations is an important issue in Design for Additive Manufacturing (DfAM), since it enables the evaluation of geometric consistency and the optimization of geometric design. Motivated by the awareness that process-related factors have non-trivial effects on geometric deviations, a new method is proposed in this paper which combines Statistical Shape Analysis with Gaussian Process to enable the modeling of deviations with consideration of process parameters. By learning from a number of simulated samples, the method could achieve effective prediction of deviations for new parts. Its applications in surface deformation evaluation and geometric compensation are also discussed, which will bring substantial benefits to DfAM.