Materials

In this research, DAIKO SF 625 welding wire, which has a broad operating temperature range from -269°C to 1000°C, was used. Its diameter is 1.2 mm, and the chemical composition of the wire is presented in Table 1 (according to the AWS A5.14 standard):

The material used in the experimental process was S45C (JIS G4051) steel. The chemical composition of the substrate included 0.48 C, 0.35 Si, 0.9 Mn, 0.03 P, 0.035 S, 0.2 Ni, 0.2 Cr, and 0.3 Cu (in wt.%). Its dimensions are 200 mm × 200 mm × 10 mm in length, width, and height, respectively. On the other hand, the plate was cleaned before the experiment.

CMT-WAAM system

All the samples were fabricated on the CMT-WAAM system, as shown in Figure 1. It consists of a CMT welding unit (TPS 320i) and a 6-axis robot (FD-V8 OTC Daihen). During welding, the motion of the welding wire is regulated with an average frequency of approximately 70 Hz24, 25, 26, 27, and the power source is supplied intermittently during each short-circuit period to create the arc and melt the metal wire. As a result, molten metal droplets were steadily controlled into the welding pool.

With the CMT welding principle, the welding wire is first moved closer to the substrate to start the welding process. An arc is formed, and the welding head is controlled to process the molten metal into the welding pool by moving the welding wire closer and farther away to minimize the thermal effect in the welding area. At the end of a welding line, the process is repeated. The weld beads are built on each other until the part is completed.

Building thin-walled samples and data collection methods

In this study, three rectangular thin-walled samples were subjected to three different welding speeds ( = 65, 50, and 35 cm/min), and the wire feed speed () was held constant at 6.5 m/min, as shown in Table 2. The linear heat input () is calculated according to Eq. (1).

where η is the energy conversion efficiency (normally, η = 85% for CMT), and U and I are the voltage and current, respectively. With the CMT power source, U and I were automatically set according to the value of the wire feed speed, as shown in Table 2.

The thin-walled samples were made of 20 layers, and the programmed dimension of each layer was the same, with dimensions of 130 mm in length and 60 mm in width. The starting point, the ending point, and the transfer point between the layers are the same and are in the middle of the length dimension. During the deposition process, the arc was emitted continuously according to the CMT principle. Moreover, commercially pure argon was used (99.99% Ar) with a flow rate of 16 L.min to protect the welding pool. All the samples were fabricated at room temperature. The distance between the torch tip and the workpiece surface was fixed at 10 mm. The angle between the axis of the welding torch and the substrate was also maintained at 90 degrees.

After fabrication, all the samples were scanned by a Kreon Zypher II scanner to collect their actual shape. The scanned data were then processed with Geomagic Design X software, as shown in Figure 2.

In this study, the geometric characteristics of the thin walls investigated are illustrated in Figure 3, including the total width (), the effective width (), the total height (), and the effective wall height (). The average layer height () is determined by the ratio between and the number of printed layers.

The method used to analyze the geometrical characteristics is presented in Figure 4. Each thin-walled sample was evaluated at four cross-sections (cs1, cs2, cs3, and cs4). The distance between two adjacent cross-sections is 25 mm. cs1 and cs2 are symmetrical with respect to cs4 and cs3, respectively, through the center plane along the length of the scanned sample. This ensures the survey along the entire length of the deposition path and monitoring the difference in profile at the beginning and end of the deposition line. These intersection profiles were processed in AutoCAD software to measure , , , and .

The material deposition efficiency () is a value that represents the efficiency of the entire WAAM process in general. It is determined by the ratio of the effective cross-sectional area to the total cross-sectional area at the local location under consideration, as described in Eq. (2), where the effective cross-sectional area is calculated via the values of effective width () and effective height ().

The process for evaluating the surface roughness parameters is shown in Figure 5. For each sample, two portions of the opposite surface on each wall along the length of the sample were cut. Each portion has the same dimensions of 70 mm × 30 mm. These four surface samples were subsequently used to analyze the surface roughness parameters via Omnisurf 3D software. The studied roughness parameters are the maximum roughness , average roughness , and mean square roughness , which are calculated via Eqs. (3), (4), and (5), respectively.

where represents the coordinates of the grid points and represents the surface area.

Geometric characteristics

The results regarding the geometric characteristics (including , , , , and ) of the thin-wall samples fabricated via CMT-WAAM are displayed in Table 3, Table 4, respectively. For each type of output result, the average values from the measurements were calculated to construct graphs.

The graphs that depict the influence of the welding speed on , , and are presented in Figure 6. The decreases proportionally with the welding speed . The decreases from 2.63 mm to 1.83 mm when increases from 35 cm/min to 65 cm/min. tends to decrease rapidly at lower speeds (from 35 to 50 cm/min), whereas it slowly decreases at high values of , from 50 to 65 cm/min (Figure 6a).

Figure 6b shows that both the values of and gradually decrease with increasing . Specifically, the overall width of the wall () is 9.39 mm, 8.00 mm, and 6.83 mm when is set at 35, 50, and 65 cm/min, respectively. The effective width of each sample () is also proportional to the overall thickness, and its values are 5.9 mm 5.04 mm and 4.3 mm at = (35, 50, and 65) cm/min, respectively.

Figure 7 illustrates the four surface portions of the thin-walled sample “No. 1” that were extracted from the 3D scanned data through the procedure in Figure 5. These representative surfaces were used to measure surface roughness parameters (, , and ) via Omnisurf 3D software. Figure 8 shows the morphology characterization of the surface portion in Figure 7.

The surface roughness parameters of all the samples are shown in Table 5 and Figure 9. The average measurement values are represented to evaluate and compare all the samples.

The maximum roughness value decreases from 168.968 µm to 108.464 µm when increases from 35 to 50 cm/min, respectively. When increases from 50 to 65 cm/min, tends to increase from 108.464 µm to 128.811 µm. Similarly, the average roughness and the mean square roughness have similar trends. For example, decreases from 5.902 µm to 4.326 µm when increases from 35 cm/min to 50 cm/min, whereas it increases from 4.326 µm to 5.394 µm from 50 cm/min to 65 cm/min. decreases from 9.005 µm to 6.544 µm when increases from 35 cm/min to 50 cm/min, and it increases from 6.544 µm to 8.115 µm from 50 cm/min to 65 cm/min.

Effects of on , , and

As mentioned previously, the linear heat input () plays a crucial role in determining the dimensions and properties of the deposited layers. When the increases, a greater amount of energy is applied to melt the filler wire and the base metal, resulting in larger melting pool dimensions and deeper penetration into the base material. Consequently, increasing tends to result in thicker layers and increased (Figure 11a). Similarly, and increase with increasing (Figure 11b).

The effect trend of on , , and is opposite to that of . Figure 6 shows that all the , and values decrease with increasing v. This can be explained by the nature of the process parameters. indicates the rate at which the amount of molten material is deposited into a welding line, whereas refers to the amount of heat applied per unit length of deposition. As indicated in Eq. (1), the is high at a low value of 28, 29. At a fixed value of , when the increases (or decreases), the ability to melt the metal wire is greater, more material is melted and added to the weld pool, and the solidification process also takes longer 19. Therefore, the melting pool and weld bead become increasingly larger. This is why all the characteristics , , and increase (Figure 11).

Effects of on surface roughness

The influence of on surface roughness parameters is similar to the influence of , meaning that they decrease when the increases from 191 to 245 J/mm and increase again when the increases from 245 to 352 J/mm (Figure 12). The reason is that when , increasing reduces the / ratio, and the staircase effect between layers decreases, leading to a decrease in surface roughness, and the surface quality gradually stabilizes. When is too low, the amount of heat in the weld pool is large, and the cooling rate is low. After deposition, the surface unevenness and surface roughness are high. As increases, this heat buildup decreases, the material melts and solidifies more stably, the surface roughness becomes more uniform, and the values decrease19. However, if continues to increase, the stability of the arc gradually deteriorates. This adversely affects the surface roughness. As a result, the surface roughness increases, and the surface quality deteriorates.