Ananda Varun, Saravana Kumar Gurunathan, Jayaganthan Rengaswamy, Srinivasan Balamurugan
Aero-Component and Systems Design Department, Honeywell Technology Solutions Lab Pvt. Ltd., Bangalore 560103, India.
Department of Engineering Design, Additive Manufacturing Group & Centre of Excellence in Materials and Manufacturing for Futuristic Mobility, Indian Institute of Technology Madras, Chennai 600036, India.
Materials (Basel). 2022 Aug 26;15(17):5909. doi: 10.3390/ma15175909.
The Laser Powder-Bed Fusion (LPBF) process produces complex part geometry by selectively sintering powder metal layer upon layer. During the LPBF process, parts experience the challenge of residual stress, distortions, and print failures. Lattice-based structures are used to support overhang parts and reduce distortion; this lattice support has complex geometry and demands high computational effort to predict distortion using simulation. This study proposes a computational efforts reduction strategy by replacing complex lattice support geometry with homogenization using experimentally determined mechanical properties. Many homogenization models have been established to relate the lattice topology and material properties to the observed mechanical properties, like the Gibson-Ashby model. However, these predicted properties vary from as printed lattice geometry. In this work, the power-law relationship of mechanical properties for additively manufactured Inconel 718 part is obtained using tensile tests of various lattice support topologies and the model is used for homogenization in simulation. The model's accuracy in predicting distortion in printed parts is demonstrated using simulation results for a cantilever model. Simulation studies show that computational speed is significantly increased (6-7 times) using the homogenization technique without compromising the accuracy of distortion prediction.
激光粉末床熔融（LPBF）工艺通过逐层选择性烧结粉末金属来制造复杂的零件几何形状。在LPBF工艺过程中，零件面临残余应力、变形和打印失败等挑战。基于晶格的结构用于支撑悬垂部分并减少变形；这种晶格支撑具有复杂的几何形状，使用模拟来预测变形需要很高的计算量。本研究提出了一种计算量减少策略，即使用实验确定的力学性能通过均匀化来替代复杂的晶格支撑几何形状。已经建立了许多均匀化模型来将晶格拓扑结构和材料性能与观察到的力学性能联系起来，比如吉布森-阿什比模型。然而，这些预测的性能与打印后的晶格几何形状不同。在这项工作中，通过对各种晶格支撑拓扑结构进行拉伸试验，获得了增材制造的因科镍合金718零件力学性能的幂律关系，并将该模型用于模拟中的均匀化。使用悬臂模型的模拟结果证明了该模型在预测打印零件变形方面的准确性。模拟研究表明，使用均匀化技术可显著提高计算速度（6至7倍），同时不影响变形预测的准确性。
