Experimental investigation on process parameters of laser-assisted robotic roller bending of a thin-walled structure

Laser-assisted robotic roller forming (LRRF) combines the process capabilities of robot-based manufacturing and laser-assisted forming. In this work, the LRRF process was applied to bending DP1180 steel sheets to thin-walled structures designed for seat trails. Comprehensive experimental investigations were conducted to explore the influences of laser power, forming passes and scanning speed on the forming forces, springback and bending radii of final parts. Experimental results show that the effects of process parameters on the springback and bending radii are similar to those on the forming forces, while forming passes make an insignificant difference to the springback. The optimized process window was subsequently determined out of the balance between geometrical accuracy and experimental efficiency. By applying the optimized process parameters (laser power of 750 W, 6 forming passes, scanning speed of 5 mm/s), the peak force during LRRF was reduced to ∼2.1kN. Meanwhile, a thin-walled profile with higher precision was achieved. Specifically, the springback angle was reduced to ∼4.1° and a compact profile with a radius-to-thickness ratio of ∼1.0 was obtained.


Introduction
Ultra-high strength steels have been widely applied in the manufacturing of thin-walled structures to achieve better performance in various working environments.Due to their difficult-to-form characteristic at room temperature, ultra-high strength steels are usually processed through the traditional multi-pass roll forming process to achieve complex cross-sectional profiles [1].However, increasing requirements for compact geometrical profiles with small springback and sharp bending radius usually bring out additional roll configurations; the tendency towards customized geometry design also demands the redesign of forming dies.Both issues lead to additional process costs, thus limiting the industrial application of this forming process.Therefore, the development of novel sheet metal forming methods for the flexible production of complex thin-walled structures is essential.
Robot-based sheet metal forming has been considered an alternative to traditional metal forming methods, which responds promptly to customer-specific requirements and avoids additional costs for die manufacturing [2].Robot-based incremental forming is one typical process that has been widely applied in the manufacturing of complex thin-walled structures [3].To achieve better geometrical accuracy of the products, various studies have been focusing on parameter optimization for this forming IOP Publishing doi:10.1088/1757-899X/1307/1/012010 2 process.Dai et al. [4] developed a multi-pass incremental forming process and determined the optimal process parameters by orthogonal experimental design.Subsequent tool path compensation for the height of the stepped plane significantly decreased the maximum deviation at the stepped feature by 70%.Störkle et al. [5] improved the geometrical accuracy in double-sided incremental forming of DC04 steel sheets by varying the workpiece orientation and forming sequence.Mohanty et al [6] added two degrees of freedom (DOF) manipulators to the conventional incremental forming process to improve process flexibility.The effects of additional process parameters (part inclination and rotation) on forming force and formability were analytically investigated to improve the formability of aluminum sheets with steep wall angle.
Heat-assisted forming technology has also been widely applied and studied to achieve higher forming accuracy for difficult-to-form materials.Hao et al. [7] adopted the pre-heat treatment and heat-assisted forming strategy in the incremental forming of TA1 titanium sheets, which achieved better maximum forming angle and surface quality.Wei et al. [8] proposed a novel heat-assisted incremental bending process using the induction heating system to fabricate large curvature metal plates.The developed forming process was proved to be effective in improving forming accuracy.Ortiz et al. [9] applied an intelligent process model that compensates for springback deviations to improve the geometrical accuracy in hot single point incremental forming (SPIF) of one typical aerospace Ti-6Al-4V part, achieving an improvement in dimensional accuracy of up to 49%.Specifically, laser-assisted forming has attracted attention due to softening effects in limited areas without mechanical properties transition of entire steel sheets.Mohammadi et al. [10] introduced laser heating to SPIF and achieved a 42% reduction of bulge height by appropriate selection of laser positioning strategy.Gisario et al. [11] developed an external force laser-assisted bending process to bend Titanium Grade-2 sheets.The application of optimal laser parameters achieved sharp bending angles of nearly 140° and small bending radii of ~2 mm.
Based on the process capabilities of laser-assisted forming and robot-based manufacturing, one laserassisted robotic roller forming (LRRF) process has been proposed and applied to bending ultra-high strength steel sheets.With the introduction of synchronized laser heating and incremental roller forming, the generation of cracks in traditional cold forming can be prevented [12] and compact thin-walled structures with smaller springback and sharper bending radii can be achieved [13].Prior investigations were mostly focused on the process capability [12] as well as the microstructure evolution during LRRF [16], while the process window for this process was not well understood.Thus, full-factor experiments were performed in this study to investigate the effects of multiple process parameters on the forming forces, springback and bending radii of DP1180 thin-walled structures manufactured by LRRF.The optimized process window was subsequently determined out of the balance between geometrical accuracy and forming efficiency, which facilitated the manufacturing of compact DP1180 thin-walled structures designed for seat trails.

Experimental details
The aforementioned LRRF process is applied to bending steel strips in this study, which is schematically illustrated in figure 1.The corresponding experimental setup is presented in figure 2. The as-received DP1180 steel sheets with a thickness of 1.6 mm were cut into 100 mm ⅹ 50 mm strips, and then fixed by the clamp onto the fixture with a fillet radius of 1mm.The forming roller with a diameter of 50 mm and a width of 25 mm is installed at the end flange of the industrial robot through the connecting shaft to bend the steel strips by multiple passes.A high-power laser source is installed beside the end flange to heat steel strips synchronously during the forming process.Full-factor experimental design was devised to investigate the effects of laser power, forming passes and scanning speed on the forming forces and dimensional accuracy.Due to the significant temperature gradient of laser heating, lower scanning speeds were selected for thicker steel strips compared with the previous study [13] to achieve more pronounced softening effects through the thickness direction of irradiated areas.The highest laser power was also determined by preliminary experiments to avoid melting on the irradiated surface.The detailed process parameters are summarized in table 1, with a total number of 32 tests performed.The forming forces during LRRF were measured by the load sensor installed at the end flange of the industrial robot.The springback angles were measured by a digital dipmeter, and the bending radii were obtained from cross-sectional profiles of final parts through tangent circle fitting.

Effects of process parameters on forming forces
For robot-based manufacturing, higher demands for allowable load and robot stiffness of industrial robots usually stand for additional facility costs, which emphasizes the significance of controlling the magnitude of forming forces.The forming forces during 6-pass LRRF with different scanning speeds are presented in figure 3. The peak force of each pass fluctuates with the increase of pass numbers due to the dual influences from the accumulation of work hardening and softening effects by laser heating.
With the increase of laser power, forming forces are significantly decreased because of the lower yield strength of steel sheets at elevated temperatures.The scanning speed of 5 mm/s further decreases the forming forces by increased laser power density, and the forming forces come to a stable level with the laser power of over 700 W. The relationship between forming forces and laser process parameters of 3, 4 and 9 passes is similar to that of 6 passes, and those similar figures are not repeated here.To clarify the effect of forming passes on forming forces, the peak forces during LRRF at different passes are extracted and presented in figure 4. With the increase of forming passes, the peak force is reduced due to minor plastic deformation in each pass.The peak force can be significantly reduced by 6 passes with the scanning speed of 10 mm/s; for the scanning speed of 5 mm/s, 4-pass forming is sufficient because of more significant softening effects by higher laser power density.Also, a plot of average experimental responses for each experimental factor, i.e., the main effect plot (MEP) of peak force is presented in figure 5 to reveal the significance of multiple process parameters on peak force.According to MEP, all three parameters exert significant changes to the peak force.Besides, the degree of variation decreases with the increase of laser power and forming passes, which points out the possibility of parameter optimization for the balance between resource consumption and facility costs.

Effects of process parameters on springback and bending radii
As mentioned before, advanced industrial design calls for the manufacturing of thin-walled structures with small springback and sharp bending radii.Thus, the effects of process parameters on springback and bending radii in LRRF are investigated here to assist in the determination of a proper process window for subsequent application.
The springback angles of final parts with different scanning speeds are presented in figure 6.According to the previous study [13], the springback in LRRF is usually related to stress at loading and Young's modulus of materials at unloading.The springback based on the release of elastic deformation is positively correlated to forming forces, and thus can be significantly reduced by the increase of laser power and decrease of scanning speed; while the increase of forming passes exerts insignificant changes to the springback.Although LRRF with more forming passes can significantly decrease the forming forces, the lower Young's modulus of steel strips at unloading due to the heat accumulation still triggers higher elastic recovery, which alleviates the decrease of springback.Similarly, the MEP of the springback angle is presented in figure 7, which demonstrates obvious influences of laser power and scanning speed, while insignificant influence of forming passes on springback.Radius-to-thickness (R/t) ratio was adopted here to evaluate the effects of process parameters on the bending radii.The R/t ratios of final parts with different scanning speeds are presented in figure 8.With the increase of laser power and forming passes, the R/t ratio is decreased significantly because of lower resistance to deformation of steel strips at elevated temperatures.The decrease in scanning speed from 10 mm/s to 5 mm/s also strengthens the softening effects of materials, which leads to smaller R/t ratios.With the low springback angle of ~2°, the smallest bending radius of ~1.6 mm still has a gap from the fillet radius of 1 mm possessed by the fixture, which can be attributed to the prominent impact of robot stiffness deformation on bending radii in LRRF [17].The MEP presented in figure 9 also reveals the effects of process parameters on the R/t ratio.The laser power obviously influences the R/t ratio, followed by the forming passes; the scanning speed exerts the least apparent changes to the bending radii.

Determination and application of optimized process window
According to the above experimental results and analyses, the optimized process window with laser power of 700~800 W, 6 forming passes and scanning speed of 5 mm/s was determined to achieve better geometrical accuracy with controlled heat input and process time.One thin-walled structure designed for seat trails was subsequently manufactured by optimized process parameters, aiming at the validation of process availability in the manufacturing of complex geometrical profiles, with the detailed procedure illustrated in figure 10.The applied laser power was chosen as 750 W to achieve more distinct softening effects of materials while avoiding melting on the irradiated surface.The final complex thin-walled structure fabricated via LRRF is also presented in Figure 11.Table 2 presents the peak forces, springback angles and R/t ratios of final components obtained via robotic roller forming (RRF) and LRRF.Compared with RRF at room temperature, the optimized process parameters reduced the peak force to ~2.1kN, which facilitates the reduction of facility costs in LRRF.The manufacturing of a thin-walled structure with higher dimensional accuracy was also validated: the springback angle was reduced from ~11.7° to ~4.1°, and a compact profile with an R/t ratio of ~1.0 was achieved.

Conclusions
In this study, LRRF experiments with multiple process parameters including laser power, forming passes and scanning speed were primarily performed on DP1180 steel sheets to obtain an optimized process window, which balances the dimensional accuracy and forming efficiency.One complex thin-walled structure designed for seat trails was subsequently manufactured by optimized process parameters to validate the availability of this forming process.The main conclusions are summarized below: (1) Laser power and scanning speed are closely correlated to the softening effects of materials during the LRRF process, and forming passes affect the magnitude of plastic deformation in each pass.All three parameters exert obvious influences on forming forces.(2) Springback angles and bending radii of final parts in LRRF are distinctly affected by laser power and scanning speed.The increase of forming passes can remarkably decrease the bending radii while exerting insignificant influence on springback.(3) Optimized process parameters with laser power of 750 W, 6 forming passes and scanning speed of 5 mm/s were determined to fabricate DP1180 thin-walled structures, which significantly reduced the peak force to ~2.1kN compared with RRF.A cross-sectional profile with higher precision was also achieved, with the springback angle reduced from ~11.7° (RRF) to ~4.1° (LRRF) and radiusto-thickness ratio reduced from ~2.9 (RRF) to ~1.0 (LRRF).

Figure 8 .Figure 9 .
Figure 8. R/t ratios of final parts by LRRF with scanning speeds of (a) 10 mm/s and (b) 5 mm/s.

Figure 10 .
Figure 10.Fabrication procedure for a thin-walled structure designed for seat trails.

Table 2 .
Peak forces and geometrical accuracy of thinwalled structures formed by RRF and LRRF.