The sheet metal clamping system was validated end-to-end for a real-world process chain through case studies of representative industrial parts. The LPBF process with the sheet metal clamping system was evaluated to verify the thermal deformation results and deviation of the part position. The behavior of the sheet metal clamping system during the milling process was evaluated according to the frequency response function and surface roughness of the milled part. Finally, the residual height after the break-off test and final form deviation of the parts were measured.
4.1 Case studies
Five case studies were carefully selected according to applications, AM part geometries, and dimensions. Figure 16 shows the different parts and their settings in the case studies. Figure 16(a) shows a cuboid with a narrow and large volume, which introduces the highest energy and thermal stress into the sheet metal during the LPBF process and leads to high deformation. The cuboid represents the worst-case scenario for the LPBF process and deformation. The cuboid was not used in the milling process because there are no specific surfaces to treat. Figure 16(b) shows a dental bridge with a superstructure framework representing typical small and freeform parts with filigree milling surfaces. It is retained either by two screws with a conical contact surface or by two prepared teeth abutments; therefore, small tolerances are required. Figure 16(c) shows a dental bar, which is flat but wide. Thus, the wide sheet metal (60 mm) was needed. The dental bar is a superstructure framework for a toothless jaw and is retained by four screws. Two versions of the bar were manufactured. Two narrow sheets were used for the first version. The challenge was to print over the gap between the sheets without interrupting the LPBF process. For the second version, the same dental bar was printed on a wide sheet, as shown in Fig. 16(d), to demonstrate the feasibility of printing and milling on the wide sheet. Both the dental bridge and dental bar require high mechanical strength to transfer the loads induced during chewing. Also, the contact surfaces between the superstructure and screw for the implant must be milled for optimal fitting. Finally, the superstructure must satisfy biomechanical and esthetic functions, so the dental bridge’s superstructure is coated with a polymeric or ceramic material. The surface roughness of the LPBF part, which is covered by the coating, must not be milled because the roughness is beneficial for bonding [5][23][6]. Figure 16(e) shows a bracket, an engineering part used to fasten to another component. The bracket is very tall in relation to the sheet metal with a large volume. This presents a particular challenge for the drilling and milling feed forces as the large lever generates a high moment in the filigree structure at the top of the part.
4.2 Thermal deformation of the sheet metal in the z-direction
A critical consideration of the LPBF process is the deformation of the sheet metal. This can lead to inaccuracy of the LPBF part and disturb the recoating process. Numerical simulations were performed to predict the LPBF part’s deformations and determine whether the sheet metal clamping system can be used without disturbing the LPBF process. Critical regions can be identified to determine whether wide or narrow sheet metal is suitable. Further measures can be applied, such as increasing the thickness of the sheet metal. A Renishaw AM 400 HT was used to validate the design of the sheet metal clamping system. This is a larger machine with chamber dimensions of 248 × 248 × 285 mm3. It has a pulsed Nd:YAG fiber laser with a maximum power of 400 W, a wavelength of 1060 nm, and a laser spot size of 70 µm. The following parameters were set for the hatching: laser power of 120 W, a scan speed of 600 mm/s, point distance of 30 µm, and exposure time of 50 µs. The hatch distance was 0.084 mm, and the border distance was 0.02 mm. For the contours, the laser power was set to 70 W with a scan speed of 300 mm/s, point distance of 15 µm, and exposure of 50 µs. The numerical simulation was performed with the same parameters used in the sheet metal deformation simulation, as given in Table 2 of Sect. 2.2. The laser power was adjusted to 120 W to match that of the Renishaw AM 400 HT, and the scan speed was set to 600 mm/s. The simulation setup is shown in Fig. 17, and the boundary conditions are presented in Table 6. The GOM ATOS Core 200 3D scanner was used for scanning. The 3D measurement system was based on the optical stereo camera principle.
Table 6
Boundary conditions for the simulation
LPBF part - sheet metal
|
Bonded
|
Sheet metal groove - pin
|
Frictional, µ = 0.2
|
Sheet metal plane–base block
|
Frictional, µ = 0.2
|
Pin–base block
|
Bonded
|
Figure 18 shows the simulation and measurement of the sheet metal deviation from the computer-aided design (CAD) drawing in the clamped condition after the LPBF process. The simulation and measurement results were compared, and the regions with high and low deformations corresponded. To better differentiate the deviation, Fig. 18 shows different scales for each case study. As expected, the largest deformations were observed at the end of the cuboid, where the maximum simulated and measured values were 0.46 and 0.37 mm, respectively. The difference between the measured and simulated deformations was attributed to the simulation’s complex boundary conditions and the manufacturing tolerances of the sheet metal. Despite the differences between the simulation and measurement, the simulation does show trends that can be used to identify critical areas before the LPBF process is started.
Large deformation of the sheet metal can lead to a large deviation of the LPBF part from the tolerance. To evaluate the clamping system’s effect on the part tolerance, the sheet metal flatness needed to be monitored at different process steps. The flatness represents the minimum and maximum distances between two parallel planes [24]. Figure 19 shows the sheet metal’s flatness when clamped after different process steps based on 3D scans. Clamping smoothed the sheet metal from previous processes by 44% on average for all case studies. The average deformation of the clamped sheet metal increased by 23% before and after the LPBF process. The deformation increased by 500% after the LPBF process when the clamps were released because of the released residual stresses. The deformation was then reduced by 61% on average when the sheet metal was clamped again for the milling process. The maximum deformation of the clamped sheet metal during the milling process was 1.24 mm. Therefore, the clamped sheet metal’s average deformation before milling was 48% greater than that of the clamped sheet metal during LPBF. Overall, the cuboid showed the highest deformation when unclamped after LPBF and clamped before milling. Due to the fact that the cuboid had the highest and most connected printed volume and area without a support structure. These deformations were investigated without considering heat treatment. As a next step, the same investigations should be performed with heat treatment, which would reduce the residual stresses that caused the high deformation after the clamping was released. The deformation in the clamped position before milling could be reduced through heat treatment.
4.3 Deviation of the part position in the x-y direction
The maximum deviation of the part position from the CAD drawing is important for position detection during the milling process. Because of the sheet metal’s thermal expansion from the measurement point, the largest deviations occur farthest from the clamping point. The deviations in the x- and y-directions of the boundary box were measured and compared to the CAD drawing for each case study.
Figure 20 shows the maximum deviations after each process step. After the LPBF process with the sheet metal clamping system, the average maximum deformations were 0.034 mm in the x-direction and 0.165 mm in the y-direction. The overall maximum deviation before the milling process was 0.23 mm. The average maximum deviations between the clamped positions after the LPBF process and before milling were 32% in the x-direction and 27% in the y-direction. The larger distances and, therefore greater thermal deformations in the y-direction led to greater deformations of LPBF parts than in the x-direction. The absolute maximum deviation in the milling process was 0.23 mm. These results demonstrate the need for position detection to facilitate the milling process.
4.4 Dynamic compliance
A stiff clamping system with low dynamic compliance is required for a stable milling process without chattering. A frequency response function was created to investigate the sheet metal’s dynamic stability in different case studies. The LPBF parts were excited with an impact hammer (Type 9722A500, Kistler). A piezoelectric charge accelerometer (Type 4393, Brüel & Kjær) was used to measure the response. A signal acquisition box recorded the excitation and response signals. Finally, the frequency response function was calculated from the acquired data. Figure 21 shows the frequency response functions for the different case studies and the corresponding measurement setup. The measurements were performed before the milling process. The dynamic compliances of the LPBF part, LPBF support structure, and sheet metal were measured.
The dominant natural frequency peak for the bracket was at 1600 Hz with an amplitude of 0.11 µm/N. The dental bridge’s three dominant frequency peaks were smaller but occurred at 500, 1600, and 2200 Hz with a maximum amplitude of 0.035 µm/N. The different amplitudes for the dental bridge and bracket were probably due to their different dimensions. The bracket had a large lever, which increased the amplitude, while the dental bridge was a small part. The dominant natural frequency peak for the dental bar on a wide sheet was at 2800 Hz with an amplitude of 0.75 µm/N. The dominant relevant frequency peak of the dental bar bridging two narrow sheets was at 1900 Hz with an amplitude of 0.5 µm/N. The dental bar on the two narrow sheets was more stable along the length axis (y-direction) because it had a higher inertia moment than the single wide sheet. However, the narrow sheets were less stable along the transversal axis (x-direction) because of the lower bending stiffness caused by the sheets’ gap.
4.5 Milling roughness
The surface roughness after milling was measured on a face and side for each case study, except the cuboid. The final surface roughness of each LPBF part with the sheet metal clamping system was compared to the surface roughness of a rigid block clamped with a standard parallel clamping jaw. The milling experiment was conducted with a five-axis CNC milling machine (DMU 60 monoBlock from DECKEL MAHO). Cooling lubricant (B-Cool 755 from Blaser SwisslubeAG) was used. Because of the freeform surfaces, different end mills, ball mills, and roughing mills were used. The roughness was measured with a confocal 3D laser scanning microscope (Keyence VK-X200K) having a z-direction resolution of 0.5 nm and a Gaussian filter. The surface roughness measurements are plotted in Fig. 22. The surface roughness of the LPBF parts with the sheet metal clamping system was slightly greater than that of the rigid block. The maximum measured roughness of LPBF parts with the sheet metal clamping system was Ra = 2.8 µm. Despite the rougher surface, Ra = 2.8 µm is sufficient for a large range of end-user applications. Despite the dominant peaks observed with the dynamic stability measurements, the milling roughness experiment showed minor chatter marks. The results showed that the milling parameters should be adjusted to avoid the eigenmode and chatter marks caused by the sheet metal clamping system.
4.6 Milling support break-off test
The LPBF part is connected to the sheet metal by solid milling supports and block supports commonly used for LPBF. The LPBF supports are removed during the milling process. The LPBF parts are removed in two steps from the sheet metal after milling, as shown in Fig. 23. First, the milling supports are cut from the sheet metal with a side cutter. Second, the remaining milling supports are removed with pliers. The residual height of the remaining milling supports after the LPBF part was removed is plotted in Fig. 23. The maximum residual height was 0.297 mm, and the average residual height for all parts was 0.14 mm. The remaining residual height was deemed acceptable for parts with a nonfunctional surface at this position. The milling supports should be positioned on nonfunctional surfaces. If the placement on a functional surface is unavoidable, then a simple manual grinding can remove the remaining material.
4.7 Deviations of the final form
The local form deviations of the case studies were measured after post-processing, as shown in Fig. 24. The dental bridge had a maximum deviation of − 0.32 mm on the as-built surface and 0.07 mm on the milled surface. The dental bar on two narrow sheets had a maximum deviation of 0.08 mm on the as-built surface and 0.04 mm on the milled surface. The dental bar on a single wide sheet showed less deformation than the bar on narrow sheets but also a local maximum deformation of 0.08 mm. The bracket had a maximum deviation of − 0.6 mm on the as-built LPBF surface, which was attributed to thermal stresses and strong forces acting on the LPBF part when two holes were drilled. For all case studies, the maximum deviation of the milled surface was 0.08 mm. The form deviations were mainly caused by deviations of the part position, as discussed in Sect. 4.3. An offset of 0.2 mm was used for the milled surfaces. A larger offset with a value of 0.5 mm is recommended to achieve tolerances for the final part in the future. However, the final form deviations resulting from the sheet metal clamping system are acceptable for many end-user applications.