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

1276

3

In this work, the authors explore the application of a thermal process termed ’ironing’ and evaluate its applicability to reducing the warping of ABS parts produced by FFF. Ironing can be understood as a layer wise surface treatment that allows the heated nozzle to go over printed layers with reduced or no extrusion, partially remelting the surface, as illustrated in the scheme of Figure 1(b). Basic iron ing functionalities have recently been made available in various slicing software. Preliminary results from previous research [17] suggest that the ironing process has the potential to reduce warping in ABS parts. Furthermore, it has been made clear that the thermal process will inevitably increase printing time, material usage, and energy. Even so, it has several advantages over other post-processing methods due to its simplicity and the fact that it is applied during the build process. The methods used by the authors share the focus of producing specimens that consistently display warping under constant printing conditions, to later enable the addition of ironing layers across the part. The CAD software used to model the specimen was SolidWorks 2019, and the specimens were manufactured on an Ultimaker S5. The following criteria were defined to choose the shape of the specimens: su ffi cient surface area to stay attached to the build platform surface (glass) even when the corners are warped, a maximum height of 50 layers [10], a small volume to reduce printing time, and several 90-degree corners, which are stress concentration locations. Combining these conditions with those used in the literature [7, 17, 21], a cuboid with 30x10x5 mm was used for the tests. The models were exported in stereolithography (STL) file to the slicer software Ultimaker Cura 4.6 to generate the g-code. Due to the limitations of the ironing process in the slicer, this g-code file served as input to a mid-processing software responsible for the ironing allocation parameters. For this purpose and based on previous works, a Python script was developed [17] to allow ironing application in any desired layer of the samples. The script finds ironing commands in the g-code, usually only existent in the top layers of each part, and replicates them along each chosen layer. In addition, it allows the definition of an o ff set between the height of the ironed layers and extrusion head, facilitates the definition of the ironing path, and controls the material and cooling flows. After the process, a new g-code file serves as input for the FFF machine. Three approaches to evaluate the ironing process influence on the warping of the specimens were planned: applying ironing on the first layers, applying ironing evenly across the specimen, and applying ironing consecutively on the first layers. In all adopted approaches, the ironing process is performed at the same layer height and with the same nozzle temperature as the previously deposited layer currently being ironed. Similarly, the ironing scanning path in the experiments is always orthogonal to the infill deposition path of the ironed layer, and both are performed with a Zig Zag pattern. Finally, it should be noted that applying the ironing process on the 30 × 15 × 5mm cuboids increases the printing time by 2 min and 42 seconds for each layer. Further details regarding FFF process parameters and the experimental procedure can be found in the work of Jose´ Lopes [22]. For each parameter under study, a group of four specimens was produced. Since the time between printing di ff erent layers of a part grows with the number of parts being simultaneously produced, the number of parts per print might influence warping-related results. In this research, due to the scarcity of data on the e ff ect of ironing on such distor tions, all samples were produced one at a time. Distortions also vary from print to print due to external variables such as bed coating conditions. For this reason, specimens without ironing were printed between all group samples. They were printed using optimised printing parameters that impose a consistent level of warping on the part and used as a benchmark. All the specimens were printed using the same filament material, red ABS with 2.85 mm from Ultimaker, a Print Core AA and a 0.4mm nozzle, a layer height of 0.1 mm, and an infill density of 100%. Achieve consistent Warping : All the ironed and non-ironed specimens were printed using parameters that would promote a constant level of warping (see Figure 2). Furthermore, from preliminary tests, it was clear that the bed tem perature greatly impacted warping. Therefore, tests were conducted to reach the temperature with the most consistent warping results. The bed temperatures tested were: no heat, 60ºC, 65ºC, 69ºC, 74ºC and 80ºC. As shown in Figure 2. Methodology 2.1. Manufacturing