November 2023 Volume 5

FORGING RESEARCH

Hz (larger fork) and the second is at 260 Hz (smaller fork). Using frequency simulations, the fork and dimensions were homed in to be in tune with the associated tuning frequency. For the 440 Hz tuning fork the final simulated frequency was 441.79 Hz (Fig 1), and the 260 Hz tuning fork had a final simulated frequency was 261.63 Hz. The expected bead geometry is based on previous testing using the same Wire-Feed-Speed (WFS) and Robot speed for the Robotic DED. These tests resulted in consistent bead heights and widths over

a length of approximately 200 mm. The preforms for this project were designed to be produced out of AWS A5.28; ER90S-D2; a low alloy welding wire. The chemical composition of the wire feedstock is given in Table 1. This wire was chosen for its high strength, durability, and the limitations on forging other alloys. The low carbon content of the alloy also provides good weldability and low amounts of impurities. High-carbon steels are prone to cracking during and/or after the welding process, so they were ruled out while choosing the welding wire.

Table 1: The chemical composition of ER90S-D2

A Wire-Feed-Speed (WFS) of 8.25 m/min and a Robot Speed of 12.7 mm/sec was chosen. These parameters were chosen due to previous testing of different parameter sets and the impact on weld quality. A pulsed-spray waveform was chosen for this deposition. The welding parameters were controlled using a combination of the KUKA teach pendant and the Lincoln Electric welder. A slightly bigger tuning fork was deposited with the Robotic DED to compensate for the dimensional changes post-forging. The shorter tuning fork has a bead offset of 7 mm and the larger tuning fork has a bead offset of 5 mm. These offsets were chosen based on previous multi-bead wall testing. The 7 mm offset led to the formation of a cavity within the preform. A shielding gas blend of Ar (75 wt. %) and CO2 (25 wt. %) at a flow rate of 14 lpm was used. A total of five layers were deposited and the preforms were allowed to cool to room temperature and then removed from the base plate using a horizontal band saw. The forks were then hand-forged at temperatures ranging from 870-980 °C. After the forks were forged and presented, one of them had a small portion removed for metallography analysis. The forged steel was tested against two other samples of as-deposited and heat-treated AWS A5.28; ER90S-D2 steel. Each of these comparison samples came from the same welded wall that was deposited with the same WFS and robot speed as the forged sample. One of these samples was heat-treated using a normalization and tempering process and the other sample remained 'as-deposited'. The three samples’ hardness was measured, and their microstructures were compared.

Results and Discussion

Figure 2: The Robotic DED printed tuning fork (on left) and the subsequent forged tuning fork (on right) The rod width (as shown in Fig 1) is 6.35 mm for the larger fork, the deposited part has the rod width of 10 mm ± 0.5 mm (Fig 2). After manual forging the rod width becomes nearly 8 mm, a total reduction of 20 %. The total length of the deposited part for the larger fork and the smaller fork are 217 mm and 160 mm respectively. These dimensions are slightly longer than the required lengths as shown in Fig 1. A slightly longer length was printed in order for the preform to be manually held during forging. The weld bead offset distance causes the cavity in between as seen in Fig 2. The cavity is more prevalent in the shorter tuning fork. The smaller bead offset distance overcomes the cavitation issue for a larger fork. After forging these cavities are closed, however, there are still some gaps in between the beads as depicted in Figure 2 (indicated by red arrows). Thus, more specific design rules are required to produce WAAM leading to the more dimensionally accurate forging. The Rockwell hardness measurements of the as-deposited, heat treated, and forged samples are shown in Fig 3. There was neither

FIA MAGAZINE | NOVEMBER 2023 83