Elsevier

Additive Manufacturing

Volume 36, December 2020, 101421
Additive Manufacturing

Influence of particle shape in additive manufacturing: Discrete element simulations of polyamide 11 and polyamide 12

https://doi.org/10.1016/j.addma.2020.101421Get rights and content

Highlights

  • Development of a method to generate optimized multisphere representation while conserving the original form factors.

  • Deposited layers of polyamide 11 and 12 have different packing and roughness properties.

  • Influence of particle shape in recoating process – spheres may not always be the best choice.

Abstract

Understanding the mechanical behavior of powders used in additive manufacturing is a fundamental aspect to improve the quality and reliability of final parts. In this context, the role of particle shape for the powder bed quality is still not completely understood and sometimes overlooked. In this study, we propose a novel method to generate multisphere or clump representation from bi- or tridimensional templates of arbitrarily shaped particles. Particles from powders typically used in additive manufacturing, polyamide 11 (PA11) and polyamide 12 (PA12) are scanned (X-ray tomography) and used as prototypes for multisphere representation. Additionally, templates of modified PA11 particles (rounded through precipitation process) and SEM images of PA12 particles were used in multisphere reconstruction and further investigated. The multisphere representations retains not only most of the original template volume but also form factors associated with flowability characteristics – aspect ratio, flatness ratio, and elongation ratio. Using multisphere representations of the aforementioned powders, realistic discrete element method (DEM) simulations of the recoating step in additive manufacturing process are performed. The speed of the recoating mechanism follows realistic process velocities (100–250 mm/s). Packing density and roughness of powder beds are measured as a function of the recoating speed for different samples. Our results show that low aspect ratio (elongated) particles tend to form more compact layers of powders at lower recoating velocities. For higher recoating velocities, spherical particles perform better than elongated particles, due to better flowability characteristics. There is a clear dependency of recoating velocity and ideal particle shape for the deposition process, in contrast with the common assumption that spherical particles always perform better and should always be preferred.

Introduction

Parts of arbitrary complex shapes can be manufactured by melting sequentially deposited layers of fine particles [1]. This technique is called selective beam melting additive manufacturing and has become more and more important for many industrial sectors in the last years. However, restrictions such as processing speed and product quality in beam-based additive manufacturing processes still limit the technology from a broader range of applications [2].

Improvements in the quality of finished parts and reduction in manufacturing time requires a broader understanding of the mechanical behavior of fine particles composing the layers of powders deposited during the manufacturing process [3]. Particle-based numerical simulations are a valuable tool to predict the mechanical behavior of powders and have been recently employed by several authors to obtain crucial information to understand the role of powder in additive manufacturing process [4], [5], [6], [7]. In [6] it is investigated the influence of the shape of the blade used in the recoating step in the quality of the recoated powder layer. The formation of jamming and empty patches is investigated using numerical simulations of non spherical particles in [5]. In [4] mechanisms like shear induced dilation were identified to have a direct influence in powder bed quality during the investigation of process parameters such as recoating speed and layer thickness. In [7] realistic particle-based simulations were performed for different recoating velocities, and an expression to describe powder bed roughness as a function of the recoating mechanism velocity was derived.

Particle shape have a large influence in the behavior of granular material. Some distinct outcomes such as particle alignment and interlocking are only observable when dealing with non-spherical particles [8], and some granular properties such as angle of repose and packing fraction are highly dependent on particle shape [7], [9], [10], [11]. The role of shapes of different polymeric powders used in the additive manufacturing is still poorly explored. The goal of powder layer deposition is to obtain homogeneous powder layers with high density and low roughness to improve part quality and process speed. Particles of irregular shape are known to have poorer flowability compared to spherical particles [12], however packing fraction increases for shape anisotropic particles when compared to spherical particles [8]. As reported in [13], ellipsoids can randomly pack up to ϕ = 0.68 for aspect ratios near 0.6 (oblate shapes) and ϕ = 0.71 for aspect ratios 1.5 (prolate shape), and even approach ϕ = 0.74 for ellipsoids with other aspect ratios. Randomly packed spheres, as comparison, reach a maximum pack of ϕ = 0.64.

In this work we investigate the role of particle shape in the quality of the powder layer. For this task, we developed a method to generate optimized multisphere representations of real particles, obtained through X-ray tomography. These particles are reconstructed in the proposed multisphere framework, retaining the form factors of the original particles while using a relatively low number of spheres for multisphere representation, allowing us to investigate multiple scenarios involving hundreds of thousands of particles.

Section snippets

Multisphere method

The discrete element method (DEM) is a particle-based numerical technique that describes the translational and rotational motion of N individual particles by means of integration of Newton's equation of motion [14], [15], [16]. Collisions between particles are described by a physical model which relies on inter-particle parameters, particle size and shape to simulate realistic granular behavior [17], [18], [19], [20], [21].

The complex shape of the particles are modeled here by means of the

Effect of particle shape on the roughness of the powder bed – experimental validations

To validate the numerical simulations of the recoating process, we measured the surface roughness of a recoated layer of virgin PA12 powder. The evaluation of the powder roughness has been proved by a surface measurement of several single layers of PA12 powder. The recoating blade dispenses the powder with a thickness of 120 μm and a recoating speed of 100 mm/s. As shown in Fig. 14, the surface measurement is done by a projection of a planar, periodical and equidistant structured pattern on the

Conclusions

We developed a multisphere reconstruction approach followed by an optimization step to reduce the number of spheres used in the multisphere representation. The optimized representation preserves the original form factors with errors not larger than 5% and also the original volume with errors smaller than 6%. Additionally, the optimization step reduced the average numbers of spheres used for a multisphere representation from the range of 102–103 to 101, which allowed us to investigate different

Authorship contribution statement

Daniel Schiochet Nasato: DEM simulations, data processing, discussion of the results, draft of the manuscript.

Thorsten Pöschel: literature review, model development, mathematical description, discussion of the results, manuscript preparation.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

We thank the German Research Foundation (DFG) for funding the Collaborative Research Center 814 (CRC 814) – Additive Manufacturing, sub-project B1. We thank Martin Heinl, Lehrstuhl für Fertigungsmesstechnik, Friedrich-Alexander-University of Erlangen-Nuremberg for providing the experimental measurements.

References (40)

  • C.-Q. Li et al.

    Multi-sphere approximation of real particles for DEM simulation based on a modified greedy heuristic algorithm

    Powder Technol.

    (2015)
  • N. Amenta et al.

    A new Voronoi-based surface reconstruction algorithm

    Comput. Geometry

    (2001)
  • M.A. Dechet et al.

    Production of polyamide 11 microparticles for additive manufacturing by liquid–liquid phase separation and precipitation

    Chem. Eng. Sci.

    (2019)
  • S. Berretta et al.

    Morphology of polymeric powders in laser sintering (LS): from polyamide to new peek powders

    Eur. Polym. J.

    (2014)
  • T.D. Ngo et al.

    Additive manufacturing (3d printing): a review of materials, methods, applications and challenges

    Compos. Part B: Eng.

    (2018)
  • R. Campbell et al.

    Additive manufacturing: rapid prototyping comes of age

    Rapid Prototyp. J.

    (2012)
  • E. Parteli

    Dem simulation of particles of complex shapes using the multisphere method: application for additive manufacturing

    AIP Conf. Proc.

    (2013)
  • Y.M. Fouda et al.

    A DEM study of powder spreading in additive layer manufacturing

    Granular Matter

    (2020)
  • S. Haeri et al.

    Discrete element simulation and experimental study of powder spreading process in additive manufacturing

    Powder Technol.

    (2016)
  • T. Börzsönyi et al.

    Granular materials composed of shape-anisotropic grains

    Soft Matter

    (2013)
  • Cited by (27)

    • Simulation of the thermomechanical behavior of discrete particles in the laser directed energy deposition process

      2022, Powder Technology
      Citation Excerpt :

      Dao and Lou [14] developed a SPH model for detailed 3D simulations of L-PBF and L-DED processes, including the powder flow, laser-powder interaction, heat transfer, solid-liquid phase change, solid-solid-liquid interactions, interfacial forcing at air-liquidand, and melt-pool dynamics. Nasato and Poschel [15] investigated the role of particle shape in the quality of the powder layer. The authors developed a method to generate optimized multi sphere representations of real particles.

    • A review of powder deposition in additive manufacturing by powder bed fusion

      2022, Journal of Manufacturing Processes
      Citation Excerpt :

      However, their choice to use 2D particle traces obtained via SEM in order to create a 3D model of particles is controversial, since the reconstruction of a 3D object based on 2D data induces unwanted uncertainty into the simulations and might cause non-realistic results [40]. There are also studies that optimize this process by managing to achieve acceptable resemblance of powder particle models with the real particles by using a minimum of spheres [48], which do not necessarily only have contact points with each other, but might overlap to create a complex shape with the minimum amount of spheres [49]. This is incredibly important, since, even though it is easy to use a large number of spheres to combine them together and develop the model of the real particle with high dimensional accuracy, the more the spheres, the more the computational cost, hence the more time necessary for the simulation to run to completion.

    View all citing articles on Scopus
    View full text