nTop Training: Automotive on Thursday, October 7 at 11am ET / 8am PT. Register here.

Design of surface roughness on CAD of AM lattices using areal surface parameters for better design validation

AM lattices often have a down skin surface roughness different from the up skin one. The workflow shown below enables the design of a virtual “CAD with Designed Surface”. The latter has a surface roughness closer to the actual AM lattice, measured using X-ray Computed Tomography.

Younes Chahid
December 16, 2020

Structural or fluid simulations in the Additive Manufacturing (AM) field are performed a utilizing Computer-aided design (CAD) models of lattice structures that do not contain any information related to the surface roughness. During the Laser Powder Bed Fusion (LPBF) significant surface roughness is produced especially in low overhangs and down skin of lattice structures as can be seen in Fig. 1.

Example of a Boolean subtracted one-unit cell of an AM lattice (a) obtained using X-ray Computed Tomography (XCT) showing different up (b) and down skin (c) surface roughness.

Fig. 1: Example of a Boolean subtracted one-unit cell of an AM lattice (a) obtained using X-ray Computed Tomography (XCT) showing different up (b) and down skin (c) surface roughness.

To account for these differences, the expected surface roughness can be designed before the manufacturing process on the CAD of the AM part (Fig. 2), creating a virtual “CAD with Designed Surface” (CADwDS).

In this study, a full factorial Design of Experiment (DOE) has been performed to correlate nTop Platform surface roughness block parameters (frequency, amplitude, seed) with the conventional engineering ISO 25178 areal surface parameters (Sa, Sq, Sp, Sv, Spd). A total of 54 DOE surfaces were generated on both a planar and cylindrical geometry.

Example of surface roughness design with increasing amplitude, frequency and fixed seed on both planar and cylindrical surfaces.

Fig. 2: Example of surface roughness design with increasing amplitude, frequency and fixed seed on both planar and cylindrical surfaces.

The results of the study show a successful correlation between the frequency parameter and the conventional areal surface parameter “Spd” and a separate correlation between amplitude and a developed “max(Sp,Sv)” parameter (also to Sa and Sq). These results demonstrate that it is possible to design surface roughness on CAD models using areal surface roughness parameters in such a way as to replicate the surface that is produced during Powder Bed Fusion.

In order to design surface roughness on CAD models of lattice structures two workflows have been developed: one to extract surface roughness from an AM lattice using X-ray Computed Tomography (XCT) as seen in Fig. 1 and another one to design a different upper and down skin surface roughness on a strut-based lattice as seen below.

Workflow of adding up skin and general surface roughness (a) combined (c) with down skin one (b) using areal surface parameters as inputs (Spd, Sa, Sq…) and a custom nTop workflow.

Fig. 3: Workflow of adding up skin and general surface roughness (a) combined (c) with down skin one (b) using areal surface parameters as inputs (Spd, Sa, Sq…) and a custom nTop workflow.

The result was a CADwDS that had a closer surface roughness and smaller mean deviation (Fig.4) to the XCT of the AM lattice compared to the original perfect CAD versus the XCT. More information can be found in my published research paper, Parametrically designed surface topography on CAD models of additive manufactured lattice structures for improved design validation. 

Colour map of the deviation analysis of the XCT vs. original CAD (a) and of the XCT vs. CADwDS (b) showing less deviation.

Fig. 4: Colour map of the deviation analysis of the XCT vs. original CAD (a) and of the XCT vs. CADwDS (b) showing less deviation.

In conclusion, the research in this field will enable the design of virtual CAD models that contain a realistic and better expectation of the anticipated defects or deviations prior to the AM process, assisting in the design validation process and cost-effectively assessing different printing parameters, strategies, surface treatments and behavior of AM parts.

If this interests you, read my first nTop guest blog, How to Design and Optimize a Patient Specific Additively Manufactured Hip Implant Stem.

You can also check out my presentation from nTop’s Metal DfAM series, Metrology and DfAM for Lattice Structures.

Written by
Younes Chahid has a first class Mechanical Engineering degree and is currently a final year PhD student on the topic of Design and Metrology of AM Lattice and Trabecular Structures at the University of Huddersfield. Younes has been selected in iMeche 2019 Rising Stars list and is also the winner of the international Additive World Design for AM 2020 challenge. Younes is also the founder and mentor of the award-winning University of Huddersfield 3D Printing Society. His experiences are in generative design, design for AM using topology optimisation, lattice structures and metrology using X-ray CT for dimensional, surface roughness and porosity analysis.

Up Next

Looking back at 2020: nTopology’s Year in Review

Over the past 12 months, nTop Platform has evolved into an advanced engineering product development platform. Engineering teams are using our software to tackle the most challenging problems. With 2020 soon coming to a close, now is the perfect time for an end-of-the-year retrospective. Here are the nTop highlights of the year.

Alkaios Bournias Varotsis
Topology Optimization Additive Manufacturing Constraints in nTopology

In the recent nTopology 2.24 update, we introduced the topology optimization overhang constraint for Additive Manufacturing. In this blog post, we take a deeper look at the unique capabilities of this new feature and how it allows you to create optimization workflows that are different from every other solution currently in the market.

Alkaios Bournias Varotsis