Acta： Revealing the mechanism of defect formation in additive manufacturing

Main author: Yanming Zhang, Wentao Yana*
The first unit: National University of Singapore
Published Journal: Acta Materialia

Research background
Industry pain point: Although laser powder bed melting (LPBF) technology can manufacture complex components, the lack of consistent product quality is still the core bottleneck restricting its industrial application. Research has shown that up to 35% of process defects are directly related to powder splashing and entrainment.
Scientific challenge: Traditional experimental methods are difficult to capture microsecond level dynamic processes, and existing numerical models lack accurate descriptions of the gas liquid solid three-phase coupling effect, resulting in unclear mechanisms for defect formation.
Innovation breakthrough point: This study establishes for the first time a CFD-DEM-CALPHAD multi physics field coupling model, breaking through the limitations of traditional simulation methods in modeling phase transition kinetics and metallurgical reactions.

research contents 
Modeling method:
Coupling Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM) to achieve bidirectional coupling between molten pool flow and powder motion
Integrate CALPHAD thermodynamic database to accurately describe metallurgical reactions in multi material systems
Develop a steam jet dynamic model to reproduce the microstructure evolution of Knudsen layer

Figure 1: Schematic diagram of computational domain and mesh.

Experimental verification:
Adopting multiple material systems such as 316L stainless steel and NiTi alloy
Combining high-speed schlieren imaging with ultrafast X-ray observation technology
Build a 4 million grid computing domain, with a single case computation time of 7 days (i9-12900K)

Figure 2: Multiphase flow in the melting process.

Research results
Thermal Splash Dynamics:
70% of the splashing comes from the molten powder in the steam jet zone (Type I)
20% is generated by sudden fragmentation of the molten pool (Type II)
10% from melt pool fluctuations (Type III)

Figure 3: Formation of hot spatters.

Defect formation mechanism:
150 μ m aggregates entering the laser action zone can lead to an 18% increase in porosity
The defect size of Ti particle inclusions in multi material LPBF reaches 45-80 μ m
Splashing momentum changes the flow field at the tail of the molten pool, causing element segregation (Ni segregation degree reaches 62%)

Figure 4: Large agglomeration formed by the coalescence of hot spatters.

Defect criteria:
τ<τc
The critical time τ _c decreases from 157 μ s to 67 μ s as the scanning speed increases

Deep insight
▶  Technological innovation value:
Establish a fully coupled dynamic model of gas melt pool powder with a resolution of 6 μ m
Revealing the chain reaction mechanism of thermal splashing agglomeration ("defect avalanche" effect)
Propose a prediction criterion for particle inclusion defects based on metallurgical reaction kinetics

▶  Engineering application inspiration:
Developing online monitoring algorithm: implementing defect warning through real-time ratio of τ/τ _c
Optimizing inert gas flow field: controlling the spatial distribution of splashing and redeposition
Multi material process design: Avoiding the combination of liquid-solid phase inversion materials

▶  Current challenges:
The high fidelity model has a high computational cost (single orbit simulation takes 7 days)
The impact of cross airflow on actual working conditions has not been modeled yet
Ultra fine powder (<20 μ m) motion trajectory prediction deviation>12%

▶  Future direction:
Developing GPU accelerated heterogeneous computing framework
Study on the metallurgical behavior of splash matrix interface
Exploring new technologies for controllable utilization of splashes (such as in-situ alloying)

Source: Yangtze River Delta Laser Alliance