PSI - Issue 42

Manuel Sardinha et al. / Procedia Structural Integrity 42 (2022) 1274–1281 M. Sardinha et al. / Structural Integrity Procedia 00 (2022) 000–000

1279

6

The average standard deviation σ 5&6 = 0.220 mm for the ironed B tests is slightly lower than the one obtained for the A series. These conclusions are aligned with the study by Sardinha et al. [17] which indicated that applying ironing on the first layers of specimens could reduce warping. C, D and E Series: To evaluate the impact of the ironing process, tests were performed in which the heat treatment was applied across multiple layers along the specimen. The approach used included the application every two, four and six layers, corresponding to D, C and E series respectively. Results of show that applying this heat treatment across the specimen not only drastically increases the printing time of specimens, but worsens warping, since all the performed tests exhibited on average more warping with than without ironing. F and G Series: In the case of F tests, the heat treatment took place from the 4 th to the 9 th layer, and in the case of G tests, from the 1 st to the 3 rd layer. Results show a clear di ff erence in warping results when applying consecutive ironing up to the 3 rd layer (G series) and after the third layer (F series). F series specimens registered an increase of warping of 49%, whereas G specimens registered a decrease of 79%, suggesting there is no advantage in ironing beyond a layer in the proximity of the 3 rd layer of the specimen. Figure 5 compares a representative example of each series with the non-ironed specimens of series A. Naturally, the overall appearance of each batch of specimens changes with the number and density of ironed layers. Among tested samples, B and E specimens presented almost no signs of the ironing process application. Some scraps can be seen in C specimens but are not very pronounced. D specimens display visible stratified warping and separation between layers. The representative example of F specimens in Figure 5 shows clear signs of melted ironed layers. In the specimens of the G series, even if the warping is attenuated, a pronounced visual defect from the cumulative thermal processing can be identified at the bottom of the specimen.

Fig. 5: Representation of overall appearance of each specimen series.

Ironing specimens every two layers (D series) proved to be the worst solution among the tested hypotheses, increas ing the distortions by an average of 121%. The application of the process every 6 and 4 layers followed, reporting average increases of 79% and 42%, respectively. Series C to F indicate that a more significant amount of ironed layers tends to display the least interesting reduction results. A possible justification for this is that ironing consecutive layers across the specimen significantly increases the number of thermal cycles during production, resulting in non-uniform thermal gradients that can cause internal stresses, leading to geometrical distortions such as warping. The perception that spacing the ironing process across the specimen does not come as a viable solution for warping comes not only from the quantitative results but also from the final visual aspect of the specimens. Specimens with an interval of 2 and 4 layers revealed excessive material accumulation (drained from the nozzle ), which can partially destroy the specimen and even compromise the accuracy of measurements. Furthermore, results indicate that the application of ironing after the first layer of material is deposited, such as the tests from the G series (and B1), can lead to an excess of build platform adhesion, as seen in Figure 6. This e ff ect was also more pronounced when a cumulative number of ironed layers was tested (G series).

Fig. 6: Example of excess of build platform adhesion seen mostly in samples of G series. The bottom view of a detached specimen (left), and residue of ABS polymer on the build platform (right).