Page 109 - MSAM-4-1
P. 109

Materials Science in Additive Manufacturing                  Topology optimization of an aluminum bicycle pedal
                                                                                    crank using laser powder bed fusion



            A                   B















            Figure 14. 3D surface texture after micro sandblasting of Alsi10Mg part:
            (A) transversal, (B) longitudinal

            and with the same laser parameters. The largest difference
            in Rsk occurs in the printed component’s transversal   Figure 15. The 3D surface texture of the printed component after micro
                                                               sandblasting Alsi10Mg part
            direction. The positive Rsk in this section indicates a surface
            with predominantly peaks and asperities characteristic of
            a downskin surface, differing from the valley-dominated   valleys than its average height. As expected, the printed
            surfaces of the other parts (upskin surface). The down   component displayed higher roughness than the cube,
            surface  cloud’s  quality  is  only  improved  with  more   which is more complex geometry can explain.
            aggressive post-processing, such as a cycle on the vibratory   Khan  et al.  noted that roughness in LPBF parts is
                                                                           45
            tumbler.                                           highly dependent on geometry, build position, and support
              Additional topological analysis using a 3D optical   volume. Horizontal surfaces typically have smoother
            profilometer (Figure  14A and  B for transversal and   finishes than inclined or vertical surfaces, which are more
            longitudinal, respectively) provided a more comprehensive   affected by the staircase effect. Larger volumes of support
            representation of surface roughness than the 2D study.   structures help reduce surface roughness by improving
            Here, roughness was expressed using areal parameters   heat dissipation and reducing energy accumulation in
            rather than directional profile measurements, with the   processed layers. However, areas  where the support
            arithmetical mean height (Sa) and maximum height (Sz )   structures are connected to the part often show localized
            used instead of the profile-based parameters Ra  and Rz .   roughness. Post-processing methods, such as sandblasting,
                                                                                                            46
            The Sa values were 30.1 µm and 25.2 µm for surfaces 1 and   significantly impact roughness on the surface. Mehta et al.
            2, respectively, and Sz values were 229 µm and 198 µm,   reported a reduced Sa of up to 43.2% after sandblasting.
            respectively. The 15% reduction in roughness between   Parameters such as laser power and energy density can
            the two surfaces is attributed to the micro sandblasting   be adjusted to improve surface roughness. Maamoun
                                                                   47
            treatment and the natural fluctuations in the deposition   et al.  achieved roughness as low as 4.5 µm by increasing
            process. Face 2 exhibited greater homogeneity than face   energy density to 65 J/mm³, though laser scanning speed
            1, though neither showed symmetrical height distribution   must be carefully controlled due to its inverse relationship
            around the mean plane. Face 1 had a skewed height   with roughness. Alternatively, surface treatments such
                                                               as polishing shot blasting, and shot peening can further
            distribution (Ssk) below the mean plane with a value   reduce roughness, with the latter reducing Sa by up to
            of 0.161, while face 2 had an Ssk of −0.604, indicating   79%.  However, these processes also increase production
                                                                   48
            remaining surface peaks on face 2. This observation is   time and costs. Finally, Chu et al.  observed that satellite
                                                                                          49
            consistent with the 2D data, where face 2 showed higher   particles in the LPBF microstructure lead to defects such
            peak values (Sp) of 133 µm compared to 72.3 µm for face   as lack of fusion, contributing to surface roughness. Visual
            1, while the deepest valleys (Sv) were 95.5 µm and 126 µm,   analysis using ImageJ software and comparison to Chu
            respectively.
                                                               et al.’s study suggests that the metallic powders supplied
              For the printed component (Figure 15), the arithmetical   by Hypermetal had sufficient flowability and sphericity
            mean height (Sa) was 23.4 µm, while the maximum height   to produce a uniform topographic profile. The powder’s
            (Sz) was 203 µm. The height distribution skewness (Ssk)   narrow size distribution (d10, d50, and d90) did not appear
            was −0.812 – a higher value than that found in the cube   to hinder the compaction process, though further analysis
            – indicating that the printed component had greater   using laser diffraction could provide additional insights


            Volume 4 Issue 1 (2025)                         10                        doi: 10.36922/MSAM025040003
   104   105   106   107   108   109   110   111   112   113   114