Influence of recoil pressure, mushy zone flow resistance and reflectivity on melt pool shape in laser powder bed fusion simulation

Jun Katagiri, Masahiro Kusano, Sukeharu Nomoto, Makoto Watanabe

TL;DR

This study investigates how key simulation parameters—recoil pressure, mushy zone flow resistance, and laser reflectivity—influence the melt pool shape in laser powder bed fusion (L-PBF) simulations. While simulations showed that recoil pressure and reflectivity affect melt pool depth, none of the tested parameters could accurately reproduce experimental melt pool dimensions, highlighting the need for more complex models.

What Problem Does This Paper Address?

Laser powder bed fusion (L-PBF) is a promising additive manufacturing technique, but defects often arise due to variations in the melt pool shape. Accurately simulating this melt pool is crucial for predicting and preventing defects. However, L-PBF simulations are complex, involving numerous parameters that are difficult to determine empirically. This paper addresses the lack of systematic evaluation of how specific physical parameters—recoil pressure (due to vapor ejection), mushy zone flow resistance (in the semi-solid region), and laser reflectivity—impact melt pool dimensions (depth and width). Understanding this sensitivity is vital for optimizing simulation accuracy and reducing the time and cost associated with experimental parameter tuning.

Methods

The researchers employed a multi-physics Computational Fluid Dynamics (CFD) simulation using the OpenFOAM software, specifically the TBinterFoam solver. This solver incorporates solid-liquid-gas phase change and laser multiple reflections.

Key aspects of their methodology include:

  • Powder Bed Simulation: A Discrete Element Method (DEM) using LIGGGHTS was used to model the initial powder bed formation.
  • Governing Equations: Standard conservation equations for energy, mass, and momentum were solved using the finite volume method (FVM).
  • Physical Models:
    • Recoil Pressure: Modeled using an evaporation-driven force (Eq. 6).
  • Mushy Zone Flow Resistance: Modeled using the Voller-Prakash model (Eq. 7), which accounts for flow resistance in the solid-liquid coexistence region.
  • Laser Reflection: A ray-tracing method was used to simulate multiple laser reflections, with reflectivity calculated via the Fresnel reflection model (Eq. 15).
  • Parametric Study: The study systematically varied three parameters:
  • Recoil pressure parameter ($\alpha$)
    • Mushy zone flow resistance parameter ($C$)
    • Fresnel laser reflection parameter ($\epsilon$)
  • Experimental Validation: Single-track experiments were conducted using Inconel738LC to obtain experimental melt pool dimensions for comparison.

Datasets

The study did not utilize pre-existing datasets for training or evaluation. Instead, it generated simulation data by systematically varying parameters and compared these results against experimental data obtained from single-track L-PBF experiments performed by the authors. The material properties for Inconel738LC were based on calculations from CALPHAD software (JMatPro).

Key Insights and Ideas

  • Recoil Pressure's Depth Impact: Increasing the recoil pressure parameter ($\alpha$) significantly increased the melt pool depth (D), likely due to enhanced surface excavation by vapor ejection. However, it had little effect on the melt pool width (W).
  • Reflectivity's Depth Impact: Similarly, increasing the laser reflectivity parameter ($\epsilon$) also led to a larger melt pool depth (D), as more laser energy was absorbed by the material. Melt pool width (W) remained largely unaffected.
  • Mushy Zone Resistance's Limited Effect: The mushy zone flow resistance parameter ($C$) primarily influenced the velocity fields within the mushy zone itself, but it had a negligible impact on the overall melt pool dimensions (D and W).
  • Discrepancy with Experiment: A critical finding is that none of the tested parameter variations could accurately reproduce the melt pool dimensions observed in the single-track experiments. This suggests that the current simulation models, even with multi-physics considerations, are insufficient to capture the full complexity of the L-PBF process.
  • Need for Advanced Models: The authors highlight that to accurately simulate the melt pool shape, especially the semicircular profile often seen in experiments, coupling CFD with solid particle dynamics (like DEM or SPH) is likely necessary. Additionally, more sophisticated laser reflection models that account for temperature dependency might be required.

Key Results

  • Recoil Pressure ($\alpha$): Melt pool depth (D) increased with increasing $\alpha$. Melt pool width (W) remained relatively constant.
  • Mushy Zone Flow Resistance ($C$): Melt pool depth (D) and width (W) showed minimal variation across a wide range of $C$ values (10⁻¹ to 10⁹).
  • Laser Reflectivity ($\epsilon$): Melt pool depth (D) increased with increasing $\epsilon$. Melt pool width (W) was largely unaffected.
  • Quantitative Comparison: While trends were observed, the absolute values of melt pool depth and width obtained from simulations consistently differed from experimental measurements. For instance, in Figure 6, the simulated depth varied significantly from experimental data across different $\alpha$ values.

Limitations and Future Work

  • Limitations:
    • The simulation did not include solid particle dynamics, which the authors believe is crucial for reproducing the semicircular melt pool shape observed in experiments.
  • The current laser reflection model (Fresnel) might not fully capture the complex interactions, especially at different angles and temperatures.
  • The Kozeny-Carman equation used for mushy zone resistance might not be applicable to the inertial flow regime that can occur in L-PBF.
  • The standard deviation of the Gaussian laser distribution was not extensively evaluated.
  • Future Work:
    • Incorporate solid particle dynamics (e.g., DEM-CFD or SPH coupling) to better simulate melt pool shape and phenomena like spattering.
  • Investigate the effect of microstructural parameters on the Kozeny-Carman equation and explore alternative flow resistance models like the Ergun equation for inertial flow.
  • Evaluate more advanced laser reflection models, including temperature-dependent reflectivity (e.g., Hagan-Ruben model), and assess the impact of the Gaussian distribution's standard deviation.
  • Conduct simulations with varying laser power (P) and scan speed (V) to further validate the model.
  • Develop CFD models coupled with microstructure formation analysis for microstructure control.

Summary updated

Abstract

Melt pool dimensions (depth: D and width: W) are strongly correlated with defects generated in laser powder bed fusion. The melt pool dimensions have been evaluated by CFD simulation including solid–liquid-gas phase change and laser multiple reflection. Such a multi-physics simulation requires a large number of parameters. Preliminary simulations for parameter identification require a lot of time and effort. The parameter sensitivity to the melt pool dimensions is important; however, the systematic evaluations have hardly been found in the literature. In this study, we performed systematic parametric evaluation for three parameters related to recoil pressure, flow resistance in solid–liquid mushy zone and Fresnel laser reflection. As a result, the D values increased with increasing recoil pressure, but the W values did not. The flow resistance force influenced the velocities in the mushy zone but not the melt pool dimensions. The D increased with increasing reflectivity, but the W did not. The melt pool dimensions varying recoil pressure, flow resistance force and reflectivity were all inconsistent with the experimental melt pool dimensions. In order to agree with the experimental melt pool dimensions, the problems to be solved were discussed by comparing the previous studies.
BibTeX
@inproceedings{katagiri2023influence,
  author = {Jun Katagiri and Masahiro Kusano and Sukeharu Nomoto and Makoto Watanabe},
  title = {{Influence of recoil pressure, mushy zone flow resistance and reflectivity on melt pool shape in laser powder bed fusion simulation}},
  booktitle = {Case Studies in Thermal Engineering},
  year = {2023},
  doi = {10.1016/j.csite.2023.103477},
}
Figure