Abstract

Stereolithography (SLA) is one of the most important techniques used in rapid prototyping processes. It has a great industrial interest because it allows for dramatic time savings with respect to traditional manufacturing processes. One of the main sources of error in the final dimensions of the prototype is the curl distortion effect owing to the shrinkage of the resin during the SLA process. Presents a study of the influence of different constructive and numerical parameters in the curl distortion, an analysis which was made using the computer code stereolithography analysis program, developed to model SLA processes using the finite element method. Also briefly presents this code.

Abstract

Residual stresses and distortion in Additive Manufactured (AM) parts are two key obstacles which seriously hinder the wide application of this technology. Nowadays, understanding the thermomechanical behavior induced by the AM process is still a complex task which must take into account the effects of both the process and the material parameters, the microstructure evolution as well as the pre-heating strategy. One of the challenges of this work is to increase the complexity of the geometries used to study the thermomechanical behavior induced by the AM process. The reference geometries are a rectangular and a S-shaped structures made of 44-layers each. The samples have been fabricated by Directed Energy Deposition (DED). In-situ thermal and distortion histories of the substrate are measured in order to calibrate the 3D coupled thermo-mechanical model. Once the numerical results showed a good agreement with the temperature measurements, the validated model has been used to predict the residual stresses and distortions. Different process parameters have been analyzed to study their sensitivity to the process assessment. Different preheating strategies have been also analyzed to check their effectiveness on the mitigation of both distortions and residual stresses. Finally, some simplifications of the actual scanning sequence are proposed to reduce the computational cost without loss of the accuracy of the simulation framework.