dc.description.abstract | Laser powder-bed fusion (LPBF) is one of the most common types of additive manufacturing (AM) processes that has gained a lot of attraction by industries. This process induces a high magnitude of a temperature gradient within the fabricated part due to fast thermal and cooling cycles. Therefore, the existence of residual stresses and deformation of produced parts is inevitable. There are a tremendous number of process parameters involved in LPBF that affect the quality of final products, such as laser power, scanning speed, layer thickness, hatching distance, etc. Modeling and simulation of LPBF provide an opportunity for predicting residual stresses and deformation of LBPF-made parts. Therefore, optimizing process parameters for minimizing residual stresses and deformation is required.
Extensive computational time and implementing a proper heat source model are some of the existing challenges in the modeling and simulation of LPBF. Due to micro-scale features of the LPBF melt pool zone, as well as the high-speed process (up to 5 m/s), the computational cost of the simulation process of making macro-scale large parts is highly expensive. On the other hand, extremely fine mesh is required for capturing heat transfer in the laser interaction zone accurately. Consequently, a large number of elements need to be analyzed for solving the problem, which requires a strong resource for computation.
This work presents the multi-scale modeling approach based on two groups of micro/mesoscale and macroscale simulations. Firstly, melt pool dimensions of Hastelloy X material single tracks were measured experimentally. Afterward, the micro/mesoscale simulation of LPBF single track was conducted, while implementing a volumetric heat source model (conical-Gaussian) to extract the transient temperature profile and melt pool dimensions. The percentage difference of melt pool depth and width dimensions derived from simulation results and experimental ones are 13% and 6%, respectively. The validated model was then used for multi-track multi-layer simulation. The effect of thin-wall thicknesses on the melt pool dimensions has been studied as an application of the multi-track simulation process. In the macroscale simulation, the thermo-mechanical model was developed for obtaining residual stresses and deformation of the fabricated part. As a major contribution, novel effective heat flux is proposed and applied for accelerating the simulation. Thermo-mechanical modeling of the cube building process is carried out using an effective heat flux. The residual stress is experimentally measured using an X-Ray analyzer machine. The simulation results show a good agreement with experimental ones while a significant reduction in computational costs is achieved. The average percentage difference in predicting residual stress in longitudinal and transverse directions was 11% and the total computational time was 90 minutes. | en |