A Numerical Study of Topography and Roughness of Sloped Surfaces Using Process Simulation Data for Laser Powder Bed Fusion.

Abstract

The simulation of additive manufacturing has become a prominent research area in the past decade. Process physics simulations are employed to replicate laser powder bed fusion (L-PBF) manufacturing processes, aiming to predict potential issues through simulated data. This study focuses on calculating surface roughness by utilizing 3D surface topology extracted from simulated data, as surface roughness significantly influences part quality. Accurately predicting surface roughness using a simulation remains a persistent challenge. To address this challenge, the L-PBF technique with two different cases (pre- and post-contouring) was simulated using two-step process physics simulations. The discrete element method was utilized to simulate powder spreading, followed by the Flow-3D melting simulation. Ten layers were simulated at three different linear energy density (LED) combinations for both cases, with samples positioned at a 30-degree angle to accommodate upskin and downskin effects. Furthermore, a three-dimensional representation of the melted region for each layer was generated using the thermal gradient output from the simulated data. All generated 3D layers were stacked and merged to consolidate a 3D representation of the overall sample. The surfaces (upskin, downskin, and side skins) were extracted from this merged sample. Subsequently, these surfaces were analyzed, and surface roughness (Sa values) was calculated using MATLAB. The obtained values were then compared with experimental results. The downskin surface roughness results from the simulation were found to be within the range of the experimental results. This alignment is attributed to the fact that the physics simulation primarily focuses on melt pool depth and width. These promising findings indicate the potential for accurately predicting surface roughness through simulation.