Powder movement rules of laser powder bed fusion additive manufacturing aluminum alloy based on discrete element method

Abstract

Laser Powder Bed Fusion (LPBF) is a promising metal additive manufacturing technology based on layer by layer powder spreading, and powder bed uniformity has a great influence on the forming quality. By Discrete Element Method and powder spreading experiment, the interaction and movement between powder were studied during powder spreading, including powder jamming, rebound, splash, eddy, and empty powder area. Additionally, five kinds of powder spreading schemes were explored, and the new process of one-way reciprocating with tri-splint blade was designed to change the motion state of powder spreading from “blade pushing powder” to “blade holding powder.” By increasing the distance between the blade and the working platform form 0 to 20 µm with the distance between the upper surface of the substrate and the working platform 50 µm, defects such as powder splash and empty powder decreased. And the uniform powder bed of aluminum alloy powder was achieved with the new process of one-way reciprocating with tri-splint blade structure.

Keywords

Laser powder bed fusion powder spreading discrete element blade Al-Cu

Introduction

Laser Powder Bed Fusion (LPBF) additive manufacturing is one of the most promising technologies in the metal additive manufacturing, due to outstanding performance in strength, precision, and density.^1,2 LPBF additive manufacturing is based on layer-by-layer stacking, and metal powder melts and solidifies layer by layer after absorbing energy source.^3,4 Therefore, the characteristics of powder bed formed by powder spreading affect the dimensional accuracy, quality stability, and structural integrity of PBF forming parts. Uniform and smooth powder bed is the basic requirement to achieve high-quality PBF manufacturing. Compared with other metals, Powder Bed Fusion aluminum alloy is relatively difficult and requires higher LPBF equipment, which restrict the research progress and application of aluminum alloy.^5,6 LPBF forming aluminum alloy technology is mainly determined by the special physical properties, including low density and low restitution coefficient. In addition, aluminum alloy has higher reflectivity to laser and higher thermal conductivity, which required higher energy to melt powder.⁷ Higher thermal conductivity leads to faster heat transfer and low molten pool temperature, leading to unmelt powder.^8,9 With the improvement of higher-power laser technology, LPBF forming of aluminum alloy materials is gradually maturing: Most research focused on PBF process exploration, defects and mechanism of forming parts,^10,11 composition optimization,¹² and molten pool analysis,^13,14 etc.

Powder size distribution, powder spreading layer thickness, powder spreading speed, and blade type will affect the characteristics of powder bed in LBPF,¹⁵ but it is difficult to obtain the information of powder accumulation, powder spreading, and powder collision motion by experiments. Discrete Element Method (DEM) provides a reliable numerical simulation to reveal the influencing factors and intrinsic mechanisms of powder layer quality from the perspective of particle microscopic dynamic behavior. And DEM was applied to explore the dynamic behavior in the powder spreading process of additive manufacturing, and predict the powdery behavior on patrical-scale to reveal the influence mechanism of powdery spreading quality. Haeri¹⁶ optimized the end shape of the blade, and obtained that the particle volume fraction and surface roughness in the final powder layer can be improved by using the super elliptical scraper by DEM. Riener et al.¹⁷ studied the influence of powder size distribution and morphology on the performance of AlSi10Mg parts formed by LPBF. Nan^18,19 compared the blade and roller powder spreading devices, and studied the powder velocity and trajectory in the powder pile and the powder spreading quality by DEM. Liao²⁰ used discrete element to study the influence of powder spreading thickness on powder flow, defect formation, and powder quality. Chen et al.²¹ established numerical model considering the contact force and cohesion between powder to predict the macroscopic profile during powder spreading. Sun et al.²² studied the influence of surface morphology and process parameters on powder spreading quality by DEM. Nan et al.^23–25 studied the phenomena of powder flow, segregation, and jamming in the powder spreading process of 316L materials by DEM.

In this work, typical powder flow phenomenon and formation mechanism in aluminum alloy powder spreading process, including jamming, rebound, splash, eddy, and empty powder area were studied. To decreased powder spreading defects, the distance between the blade and the working platform was optimized, and the blade structure was redesigned. A new powder spreading process of one-way reciprocating tri-splint blade for aluminum alloy was explored to obtain the uniform powder bed of aluminum alloy powder.

Materials and numerical simulation model

As shown in Figure 1, powder feeding for PBF powder spreading process is mainly divided into two kinds methods: down powder feeding and upper powder feeding. For down powder feeding in Figure 1(a) and (c), powder warehouse send the required powder up to the working platform from the bottom. For upper powder feeding in Figure 1(b) and (d), the required powder dropped from powder warehouse to the working platform, and then the blade pushes the powder to the substrate. Scraper is mainly divided into blade or roller. Blade is often used in industrial machines. In this paper, discrete element is used to study the single-layer powder spreading process of upper powder feeding with blade.

Figure 1.

Two powder feeding methods: (a, c) down powder feeding and (b, d) upper feeding powder.

Discrete element model of powder spreading

Contact models are an important basis for Discrete element method. Particle motion inevitably causes collisions between particles, leading to the generation of force between particles. The dynamics of the particles during milling were calculated using Newton’s equation of motion:

m_{i} \frac{d v_{i}}{d t} = \sum F_{c, j} + m_{i} g

\frac{d (I_{i} \cdot w_{i})}{d t} = \sum M_{c, i}

Where, $m_{i}$ and $I_{i}$ are the particle mass and the moment of inertia, respectively; $v_{i}$ and $w_{i}$ are the translation velocity and the angular velocity of the particles, respectively; $F_{c, j}$ is the interaction force between the particles or the wall; $M_{c, j}$ is the torque caused by the contact force.

The elastic contact force during contact between particles is described by HertzMindlin model, while the adhesion force is calculated by JKR theory:

F_{J K R} = {\sqrt{8 π Γ E^{*}} a}^{3 / 2}

Where, $Γ$ is adhesive surface energy; $E *$ is the equivalent Young’s modulus; $a$ is contact radius.

Figure 2 shows the traditional two-way single layer powder spreading model with empty return stroke, which is composed of powder warehouse, blade, working platform, and substrate (1000 μm × 40 μm). In the process of powder spreading, the substrate first drops down to layer thickness t, and then the powder begin to fall from powder warehouse to the working platform. After the required powder accumulation completed, blade begins to push powder moving from left to right, and the powder bed is formed in substrate. After the laser scanning powder, power in substrate is melted and solidified. For the next power spreading, the substrate drops down to layer thickness t. The blade moving speed is 20 mm/s, and the simulated time step is 5E-08 s.

Figure 2.

Schematic diagram of powder spreading model.

As shown in Figure 2, t is the height of each downward movement of the substrate, namely the layer thickness of the laser process parameter, which is related to the powder size. In this work, t is 30 μm. H is the distance between the upper surface of the substrate/part and the working platform, so H is 30–50 μm between the substrate and the working platform. η is the distance between the blade and the working platform, which is taken as 0 or 20 μm in this work.

Material physical properties and particle interactions

The metal powder used in this work are aerosolized AlCu5MnCdVA aluminum alloy powder, provided by Avic Mette Powder Technology Beijing Co., LTD. The physical parameters of powder were reported by Avic Mette Powder Technology Beijing Co., LTD, including reposed Angle (43°), loose density (1.44 g/cm³), vibration density (1.75 g/cm³), chemical composition, particle size distribution, etc. The particle size distribution ranged from 15 to 60 µm with D₉₀ = 57 µm (Figure 3).

Figure 3.

Particle size distribution.

The discrete element simulation of the powder spreading process was applied by Altair^® EDEM™ and Table 1 shows the physical properties of particles and equipment, including powder, platforms, and blade. The Generic EDEM Material Model Database (GEMM) of EDEM was used to calculate particle interaction parameters, including restitution coefficient and friction coefficient, as shown in Table 2.

Table 1.

Physical and mechanical properties of the particles and equipment.

Property and parameters	Particle	Equipment	Blade
Particle diameter D₉₀ (µm)	57	\	\
Particle density ρ (kg/m³)	2820	7980	1000
Poisson ratio ν	0.3	0.3	0.45
Young modulus E (GPa)	72	210	5.8 MPa

Table 2.

Interaction parameters of particle-particle in simulations.

Property and parameters	Particle-particle
Coefficient of restitution, e	0.15
Coefficient of static friction, e	0.44
Coefficient of rolling friction, e	0.01

Discrete element powder spreading process scheme and innovative design of blade structure

Table 3 and Figure 4 show five process schemes of PBF powder spreading. Scheme 1 is traditional one-way powder spreading process with η = 0 µm and H = 30 µm with empty return stroke. The powder spreading schemes of one-way powder spreading process with η = 0 µm/20 µm and H = 50 µm, bidirectional powder spreading with η = 20 µm and H = 50 µm without empty return stroke were studied in Scheme 2–4.

Table 3.

Powder spreading process scheme.

Scheme no.	Spreading form	Distance between the blade and the working platform η (µm)	Distance between the upper surface of the substrate/part and the working platform H (µm)	Blade structure
1	One-way	0	30	Single
2	One-way	0	50	Single
3	One-way	20	50	Single
4	Two-way	20	50	Single + single
5	One-way	20	50	Tri-splint blade

Figure 4.

Schematic diagram of Scheme 1–5 powder spreading process: (a) Scheme No. 1, (b) Scheme No. 2, (c) Scheme No. 3, (d) Scheme No. 4, and (e) Scheme No. 5.

Additionally, this work designed the powder spreading process of one-way reciprocating powder spreading with tri-splint blade, named as scheme 5. The specific structure is shown in Figure 5. The blade adopts a tri-splint blade structure, which effectively reduces the uneven powder laying phenomenon caused by powder splash, jamming, and rebound.

Figure 5.

Powder spreading process and blade structure of Scheme 5.

Figure 5 shows the one-way reciprocating powder spreading process with tri-splint blade of Scheme 5. Powder spreading device includes tri-splint blade, funnel, and powder warehouse. Funnel and tri-splint blade are the overall assembly structure to avoid empty return stroke. Powder dropped from the powder warehouse into the funnel, and the substrate dropped down a layer thickness t. Then powder dropped from the funnel into the place between splint 2 and splint 3. Next, the tri-splint blade with funnel begins to spread powder from left to right. After this layer is finished, the laser scans the powder to complete melting and solidification, and the substrate drops a layer thickness t. Then, the powder in the funnel again fall to the place between the splint 1 and 2 splint, from right to left and the second-lay powder spreading start, in turn cycle.

Results

Powder spreading experiment

Powder spreading is one of the most important and time-consuming operations before parts printing. Unreasonable powder bed quality will directly lead to failure of printing. For the actual printing process, many factors affect the smoothness and uniformity of powder bed, including the leveling of substrate, substrate surface flatness, powder material and particle size distribution, spreading speed, blade and wind field structure.

In this work, single-layer powder spreading experiment was carried out. The blade was made of rubber, and the blade structure was shown in Figure 6.

Figure 6.

Blade structure: (a) scraper assembly, (b) rubber blade, and (c, d) blade cross-section structure.

Figure 7 shows the typical uneven powders spreading in the experimental powder spreading process. And the uneven powder bed directly affects the LPBF-forming quality and the powder bed quality for the next powder spreading.

Figure 7.

Typical defects in powder spreading experiment: (a) forming bin, (b) powder bed, (c) uneven powder caused by broken blade and (d) uneven powder caused by substrate.

As shown in Figure 7(c), a significant scratch in the uneven powders spreading appeared, which is caused by the structural imbalance of blade itself during the powder movement. Therefore, the powder spreading process parameters and the manufacturing and assembly accuracy of the powder spreading equipment lead to the typical uneven powders spreading, especially the wear of the blade, and regular changing of the blade is an ideal solution.

Figure 7(d) showed the typical uneven powders spreading caused by substrate. Currently, the installation and disassembly of the substrate is done by manual operation. Additionally, the uneven substrate due to the wear of the substrate led to uneven powders spreading.

Flow stage of aluminum alloy powder spreading

The formation of powder bed is mainly composed of two processes, powder dropping and powder spreading: (1) Figure 8 shows the volume proportion distribution of different particle sizes after powder dropping and powder stacking condition, in which the color represents the particle size, red represents the maximum particle size of 65 µm, and blue represents the minimum particle size of 15 µm. Figure 9(a) and (b) showed the dropping-powder stages, in which the powder is stored in funnel and falls from the funnel to the working platform under the action of gravity. (2) Figure 9(c), (d), and (f) showed the powder-spreading stage. The powder is pushed by a blade along the working platform into the substrate, and powder bed is formed on the substrate.

Figure 8.

Schematic diagram of the volume proportion of powder with different particle sizes and the state after powder dropping: (a) the volume proportion of powder and cumulative volume proportion and (b) powder stacking condition after powder dropping.

Figure 9.

Typical stage and powder flow velocity of powder spreading process for scheme 1: (a) t = 0.01 s, (b) t = 0.11 s, (c) t = 0.36 s, (d) t = 0.41 s, (e) t = 0.42 s, and (f) t = 0.45 s.

Figure 9 shows six typical stages of powder bed, in which the color represents the speed, red represents the maximum speed 30 mm/s and above, and blue represents the minimum speed 0 mm/s. In addition to the two stages of powder falling, the spreading of powder bed is mainly composed of four steps: (1) as shown in Figure 9(c), particles enter the substrate from the powder pile under the action of blade and particle gravity; (2) as shown in Figure 9(d), the powder moves on the substrate and forms a powder bed under the combined action of friction between substrate and powder, collision force between particles and drag between blade and powder particles. (3) as shown in Figure 9(e), powder spatter and rebound happened during powder spreading; (4) as shown in Figure 9(f), powder spreading is completed and a stable powder bed is formed.

Discussion

Formation mechanism of Powder rebound, splash, and eddy

Figure 10 shows the powder rebound and splash, where location A and B are two connections between the substrate and the working platform. Blade pushed powder from left to right. Due to the existence of dead angle position, blade/substrate and working platform formed extrusion on powder, resulting in the powder jamming at location B. After continuing to move, extrusion disappeared and counter-acting force pushed powder back, resulting in powder rebound and splash.

Figure 10.

Rebound and splash of powder.

As shown in Figure 10, part of the rebounding powder flies into the air, while the others collides with the powder on the substrate, causing the originally stationary powder to move backward to the left, forming a eddy phenomenon, as shown in Figure 11.

Figure 11.

Powder eddy.

Formation mechanism of empty powder area

The powder bed of scheme 1 after powder spreading is shown in Figure 12. The empty powder area in the powder bed is mainly divided into two parts: area A and area B. Area A is formed due to inertia, where metal powder is moved at the speed of 20 mm/s by the blade. Area B area is formed due to powder rebound, which leads to the reverse movement of the powder, resulting in eddy, forming a large area of empty powder near location B.

Figure 12.

One-way powder spreading state of scheme 1 with 30 µm.

Figure 13 shows the powder bed of scheme 2 with H = 50 µm, and the empty powder area gradually decreased. Due to the inertia of metal powder, there is still inevitable empty powder A region, but empty powder A region decreased with the increasing distance H = 50 µm. Additionally, due to the melting and solidification of powder, friction will be increased with rough surface of parts, and empty powder A area will gradually reduce.

Figure 13.

One-way powder spreading state of scheme 2 with 50 µm.

To decrease powder jamming, scheme 3, scheme 4, and scheme 5 were designed to optimize the powder spreading process, including increasing the distance between the blade and the working platform, two-way powder spreading and blade structure design.

Optimization of blade structure and powder spreading process

To eliminate powder extrusion of the blade, substrate, and working platform in position B, the innovative design of powder spreading processes were carried out, as shown in Table 3 and Figures 4 and 5. Figure 14 shows the powder bed of scheme 3, scheme 4, and Scheme 5.

(1) Scheme 3: The distance between the substrate and the working platform is increased from 0 to 20 µm to reduce extrusion pressure: as shown in Figure 14(a), the empty powder B area is significantly reduced, the extrusion area of the powder in the triangle zone is increased, and the internal contact surfaces and pores of the powder are increased, so as to scatter the mutual rebound force between the particles.

(2) Scheme 4: The two-way powder spreading process is adopted to reduce the area of empty powder A: as shown in Figure 14(b), two-way powder spreading does not contribute to the evenness of layering. On the contrary, since the second spreading from the right to left is carried out on the powder after the first layering rather than on the solid, the effect is worse. On the other hand, the reverse second spreading will produce powder jamming again in location A, resulting in powder rebound in location A to produce empty powder A area.

(3) Scheme 5: A rear blade on the basis of the original blade was set to prevent the powder from rebound back; And a front blade on the basis of the original blade was set to prevent powder splash: as shown in Figure 14(c), the empty powder area disappeared significantly and the powder filled the upper part of the substrate.

Figure 14.

Comparison of powder spreading uniformity of three improved schemes: (a) scheme 3, (b) scheme 4, and (c) scheme 5.

Compared with the traditional one-way powder spreading process, the funnel and blade are assembled as a whole structure, which solves the problem of empty return and achieves reciprocating powder spreading. To reduce the defects of powder bed, such as splash, rebound, and empty powder area, the motion state of powder spreading is changed from traditional “blade pushing powder” to “blade holding powder” by using tri-splint blade structure.

Figure 15 showed powder flow velocity of powder spreading process for scheme 5. With the help of tri-splint blade structure, powder were both hold and pushed instead of only pushed: rear blade prevented the powder from rebound back and front blade prevented powder from splashing away.

Figure 15.

Typical stage and powder flow velocity of powder spreading process for scheme 5: (a) powder falling, (b) begin powder spreading, (c) powder spreading and (d) complete powder spreading.

Along the powder-spreading direction, particle size distribution gradually increases, mainly because the powder with small particle size is mostly located at the bottom of the powder pile due to gravity reasons, resulting in the powder with small particle size first spread on the substrate. Therefore, it is suggested to add a micro-vibration device at the bottom of the substrate in order to eliminate the empty powder area while homogenizing the distribution of powder size.6

Conclusions

In this work, a simplified DEM model of powder spreading process for PBF additive manufacturing was established. The typical flow states and interactions of particles in the powder spreading process were studied. Six typical states of powder movement were obtained, and the formation mechanism of powder rebound, spatter, and empty powder area defects were studied. Additionally, by optimizing the powder spreading process parameters and blade structure, a new powder spreading process for PBF aluminum alloy by one-way reciprocating tri-splint blade was designed, which changed the traditional powder spreading motion state of “blade pushing powder” to “blade holding and pushing powder,” and a uniform powder bed without rebound and empty powder area was obtained.

Considering the triangular position of the substrate, the working platform and the blade is full of powder, the general forming area is preferentially placed in the center of the substrate to avoid the uneven spread of powder at the edge, such as empty powder area, uneven distribution of powder and particle size difference.

Based on the previous topology optimization^26,27 and mesoscopic molten pool,²⁸ micro-vibration and wind field optimization are introduced to further optimize the particle distribution of the powder bed in the next. On this basis, the kinetic and thermodynamic process simulation of “mesoscopic powder spreading – mesoscopic molten pool – microscopic phase field – macroscopic process simulation” will be integrated and combined with structural topology optimization to form a systematic process simulation scheme for PBF additive manufacturing.

Footnotes

Handling Editor: Sharmili Pandian

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by key Research Project of Natural Science of Anhui Province (KJ2019A1157), key Research Project of Natural Science of Anhui Province (KJ2020A1117). Author Pan Lu has received research support from Anhui Top Manufacturing Technology Co., Ltd and Anhui HIT-3D Technology Co., Ltd provide desktop metal 3D printers and experiment.

Availability of data and material

The test data used to support the findings of this study are included within the article. Readers can obtain data supporting the research results from the test data table in the paper.

ORCID iD

Pan Lu

References

Wang

Zhang

Wang

XM.

Progress and challenges of laser direct manufacturing of large titanium alloy structural parts. China Laser 2009; 36: 3204–3209.

Lin

Huang

WD.

Laser additive manufacturing for high-performance metal components. Chin Sci Inf Sci 2015; 45: 1111–1126.

Xia

, et al. A multiscale understanding of the thermodynamic and kinetic mechanisms of laser additive manufacturing. Engineering 2017; 3: 675–684.

Ahmadi

Tabary

Rahmatabadi

, et al. Review of selective laser melting of magnesium alloys: advantages, microstructure and mechanical characterizations, defects, challenges, and applications. J Mater Res Technol 2022; 19: 1537–1562.

Rong

Huang

, et al. Study on room temperature and high temperature mechanical properties of aluminum alloy manufactured by laser powder bed. China Laser 2022; 49: 0802001.

Wang

Gao

, et al. Research status and development of laser additive manufacturing. Appl Laser 2016; 36: 478–483.

Azami

Siahsarani

Hadian

, et al. Laser powder bed fusion of Alumina/Fe–Ni ceramic matrix particulate composites impregnated with a polymeric resin. J Mater Res Technol 2023; 24: 3133–3144.

Zhang

Wei

QS.

Study on optimization of aluminum alloy powder. Appl Laser 2013; 33: 391–397.

Yao

Shi

Zhang

, et al. Temperature field simulation of laser melting forming process in metal powder constituency. Appl Laser 2007; 6: 456–460.

10.

Wang

Dou

, et al. Research progress of splash behavior during laser melting of powder beds. China Laser 2020; 47: 0900001.

11.

Ding

Yang

Bai

, et al. Microstructure and mechanical properties of laser stereoscopic forming AlSi10Mg alloy. China Surf Eng 2018; 31: 46–54.

12.

Yuan

Wei

Wen

, et al. Study on the selective laser melting of AlSi10Mg alloy powder. Therm Process 2014; 43: 4–6.

13.

Wei

Lin

Wang

, et al. Numerical simulation of the metal laser additive manufacturing process. Aviat Manuf Technol 2017; 13: 16–25.

14.

Xia

, et al. Multiscale understanding of the thermodynamics and dynamics of laser additive fabrication. Engineering 2017; 3: 220–239.

15.

Liu

Xia

HX.

Review of mediation-microscopic modeling and simulation methods for selective laser melting process. J Mech Eng 2022; 58: 239–257.

16.

Haeri

Optimisation of blade type spreaders for powder bed preparation in additive manufacturing using DEM simulations. J Powder Technol 2017; 321: 94104.

17.

Riener

Albrecht

Ziegelmeier

, et al. Influence of particle size distribution and morphology on the properties of the powder feedstock as well as of AlSi10Mg parts produced by laser powder bed fusion (LPBF). Addit Manuf 2020; 34: 101286.

18.

Nan

YQ.

Study on dynamics of metal powder based on Discrete Method. J Process Eng 2020; 20: 1313–1320.

19.

Nan

Ghadiri

Discrete element simulation study of metal powder card plug dynamics in additive manufacturing. J Eng Thermophys 2022; 43: 1260–1266.

20.

Liao

FY.

Research on powder bunk exhibition process based on constituency laser melting technology. Chongqing University, 2020.

21.

Chen

Wei

Wen

, et al. Flow behavior of powder particles in layering process of selective laser melting: numerical modeling and experimental verification based on discrete element method. Int J Mach Tools Manuf 2017; 123: 146–159.

22.

Sun

Jiang

GG.

Influence of the rough surface of the constituency laser melting forming area on the powder quality: discrete element simulation. J Mech 2021; 53: 3217–3227.

23.

Nan

Ghadiri

Numerical simulation of powder flow during spreading in additive manufacturing. Powder Technol 2019; 342: 801–807.

24.

Nan

Pasha

Ghadiri

Numerical simulation of particle flow and segregation during roller spreading process in additive manufacturing. Powder Technol 2020; 364: 811–821.

25.

Nan

Mehrdad

Tina

, et al. Jamming during particle spreading in additive manufacturing. Powder Technol 2018; 338: 253–262.

26.

Cao

Wang

Measurement of Young’s modulus by interferometry. Univ Phys 2017; 36: 40–43.

27.

Pan

Zhang

HH.

Research on the key technology of humanoid robot calf based on topological optimization and laser powder bed melting. Appl Laser 2023; 2: 47–55.

28.

Pan

Zhang

Wang

, et al. Mesoscopic numerical simulation and experimental investigation of laser powder bed fusion AlCu5MnCdVA alloys. Mater Res Express 2021; 8: 126525.