ECCOMAS 2024

Part-Scale Simulation of Ti-6Al-4V Microstructure Evolution during Laser Powder Bed Fusion (LPBF) Process

  • Yang, Yang (TU Delft)
  • van Keulen, Fred (TU Delft)
  • Ayas, Can (TU Delft)

Please login to view abstract download link

Microstructure evolution is dependent on the thermal history in the LPBF process, and the distribution of different solid-state phases is an important aspect of the microstructure dictating the mechanical properties of printed parts. Therefore, specific thermal histories can result in customized phase distributions, tailored to meet specific requirements or demands. However, in additive-manufactured Ti-6Al-4V parts, the brittle α′ (martensite) phase accounts for the largest proportion that needs to be eliminated with a post heat treatment. Thus, it is crucial to perform part-scale thermal and solid state phase transformation simulations to assess the distribution of different phases and to find the critical process window that controls the volume fraction of solid state phases. A framework consisting of a semi-analytical method for thermal analysis [1] one way coupled to a Ti-6Al-4V phase transformation model [2] is presented to predict the phase distributions for Ti-6Al-4V components. Effects of volumetric energy density and the scanning strategy with various numbers of scanning lasers are studied. The semi-analytical method relies on the analytical solution for a line heat sources in a semi-infinite space to capture steep temperature gradients, while a complementary field is used to impose realistic boundary conditions. The latter can be resolved numerically with a coarse spatial resolution leading to high computational efficiency. The Ti-6Al-4V phase transformation model is based on four solid transformation processes including β to α and β to α′ transformations in cooling; α′ to β + α and α to β transformations in heating. Our simulation results show that average α′ content decreases with the increasing volumetric energy density while scanning with multiple lasers simultaneously leads to a more uniform phase distribution. After the current layer is deposited, α′ phase in this layer always accounts for the largest volume fraction around 90% due to high cooling rates. However, annealing from the subsequent deposition processes triggers the decomposition of α′ phase into α and β phases, which then leads to the decrease of α′ phase in the layer.