Elsevier

Additive Manufacturing

Volume 12, Part B, October 2016, Pages 314-320
Additive Manufacturing

Full length article
Optimized build orientation of additive manufactured parts for improved surface quality and build time

https://doi.org/10.1016/j.addma.2016.06.003Get rights and content

Abstract

The layered structure of Additive Manufacturing processes results in a stair- stepping effect of the surface topographies. In general, the impact of this effect strongly depends on the build angle of a surface, whereas the overall surface roughness is additionally caused by the resolution of the specific AM process. The aim of this work is the prediction of the surface quality in dependence of the building orientation of a part. These results can finally be used to optimize the orientation to get a desired surface quality. As not all parts of the component surface are equally important, a preselection of areas can be used to improve the overall surface quality of relevant areas. The model uses the digital AMF format of a part. Each triangle is assigned with a roughness value and by testing different orientations the best one can be found. This approach needs a database for the surface qualities. This must be done separately for each Additive Manufacturing process and is shown exemplarily with a surface topography simulation for the laser sintering process.

A validation of the model is done with a monitor bracket of EOS GmbH. Measurements of five different orientations of the part, optimized according selected surface areas, show a good accordance between the real surface roughness and the predicted roughness of the simulation.

Introduction

Additive Manufacturing (AM) processes directly produce a real part from a computer-aided design (CAD) file without the need of a tool. The CAD part has only to be saved as a STL (standard triangulation language) file and depending on machine and process sliced into two dimensional layers of about 20 μm to 300 μm. The assignment of the spatial position in the building chamber and support structures, if needed, are done with machine specific software. In general, raw material is treated or deposited layer-by-layer at the cross section of a part and the building platform lowers each cycle by one layer thickness and material is allocated again. The first commercial process stereolithography (SLA) utilizes photopolymers by curing it with ultraviolet laser. Raw materials can also be powders of polymers, metals and ceramics for processes like selective laser sintering (SLS), selective laser melting (SLM) and electron beam melting (EBM), where the powder is melted by a laser or electron beam. Different from these processes the fused deposition modeling (FDM) uses extruded filaments to deposit layers [1], [2].

The layered structure of all AM processes leads involuntarily to a stair-stepping effect on part surfaces which are curved or tilted in respect to the building platform. Additionally, some processes need support structures to attach the part to the building platform, sustain overhangs or dissipate heat, where residues of removed support structures on the surface are present and deteriorate the surface quality. These effects strongly depend on the orientation of a part surface inside the building process. Hence, the orientation of a whole part has a big influence not only on the resulting surface quality. Also the accuracy of part details and the building time and costs due to the build height are strongly influenced by the build orientation, too [3].

Some work was done in the past years to determine optimized build orientations regarding various target values for different AM processes [4], [5], [6], [7]. Reeves and Cobb presented a surface roughness prediction for layered manufacturing (LM) to optimize the building direction in terms of overall surface roughness for given parts [8]. In particular, they focused on the SLA technology, which is based on the photopolymerization of a liquid resin, therefore having smaller surface roughness as in SLS. Hence, they could express the mean roughness index (Ra) as a function of surface angles only resulting from the stair-stepping effect. For SLS parts, no reasonable correlation between the theoretical and the measured roughness for surface angles between 90° and 180° (i.e. surfaces whose surface normal are directed to the bottom) could be found. The independence of the stair-stepping effect in this case is caused by the filleting effect Bacchewar et al. described [9].

Although a few attempts were done in the past on automation tools for an optimized build orientation for AM, the standard procedure is still dominated by manually packing and rotating parts for build jobs. A basic orientation optimization tool is included in the newest versions of the Magics program, but this tool considers primarily the packing density of parts and not individual build restrictions evoked by the used AM process.

This work starts with a surface topography simulation of SLS parts, which is more complex than the other aforementioned processes FDM, SLA and LOM as the raw material is a powder. This prediction of surface roughness is then used to find the optimal build orientation regarding surface quality.

Section snippets

Theoretical description of the models

In the early years of AM, Frank and Fadel set up the first simple tool for selecting an optimal orientation for parts in the SLA process [13]. They already identified a lot of factors which are influenced by changing the build orientation of a part in AM processes. The surface finish is one of the most important properties which can be optimized and is considered in nearly any work. The stair-stepping effect occurs for all AM processes and can be described with the building direction of a

Case study

For the validation of the model a real part was investigated with the algorithm as described in the previous section. The part, which is a monitor bracket, is shown in Fig. 5 as the CAD model and its implementation as a production part of EOS GmbH. This part was chosen because it has a complex structure but only a few simply structured areas are visible in its application. Hence, specific surface areas for surface quality improvement can be determined and also measured experimentally.

In the

Results and discussion

The comparison of the roughness values predicted by the orientation optimizer and measured from the real parts is shown in the diagram of Fig. 9. The error bars shown here are an approximation of the lower bound defined as 7% of the roughness value. Error bars of roughness depths are normally about 10% but naturally depend strongly on the statistical amount of measurements.

But even a plot with the lower bound approximation shows a very good agreement between the simulated roughness values and

Summary and outlook

An orientation optimization tool for AM processes which uses surface roughness and build height as optimization objectives was presented. The previous surface topography simulation of the LS parts [18] sets up the surface roughness database for the optimization tool. Using the example of a production part of EOS GmbH the function of the tool is demonstrated and the resulting optimization potential regarding surface quality and build height can be recognized. A significant improvement of the

Acknowledgements

The authors want to thank all industry partners of the DMRC as well as the federal state of North Rhine-Westphalia and the University of Paderborn for the financial and operational support within the project “STEP: Surface Topography Analysis and Enhancement of Laser Sintered Parts”. Special thanks go to EOS GmbH for contributing the case study part.

References (17)

There are more references available in the full text version of this article.

Cited by (96)

  • Optimization of key quality indicators in material extrusion 3D printing of acrylonitrile butadiene styrene: The impact of critical process control parameters on the surface roughness, dimensional accuracy, and porosity

    2023, Materials Today Communications
    Citation Excerpt :

    It should be noted that the porosity is a parameter that is investigated and adjusted to the required level, for specific 3D printing applications [18], such as in the manufacturing of scaffolds with 3D printing [19,20]. The effect on the dimensional accuracy and surface roughness of 3DP parts of characteristics such as building style, part orientation, printing speed, layer thickness, nozzle temperature, chamber temperature, and platform temperature has been studied [9,21–31]. For the surface roughness evaluation, different techniques have been practiced, with the assessment of a pin profilometer being the most common one [27], while optical scanning [32] and CT scanning have also been employed [21].

View all citing articles on Scopus
View full text