Introduction

Single Point Incremental Forming (SPIF) has been proving to have tremendous potential to bridge the cost problem to manufacture low batches or customized prototypes. However, to conduct experimental tests, the development of this technology has been impaired by the limited machinery available to carry on the process (usually adapted milling machines or robotic arms) which cannot withstand large forming forces, since they were designed for machining and positioning operations respectively, and consequently, they can only be used with soft materials (e.g., aluminum) or harder materials but at the expense of using very thin sheets [1]. Simultaneously, SPIF is a process characterized by large forming times, which limit the batch size. In this work, a new concept of parallel kinematics Incremental Forming machine, by means of a Stewart platform, was employed (see [2]) allowing for larger forming forces and faster feed-rates.

Some effort has been devoted to study the influence of feed rate in the process feasibility. Ambrogio et al. [3, 4] or Hamilton and Jeswiet [5] studied the effect in terms of shape accuracy and surface roughness. They concluded for instance that aluminum material keeps formability properties unchanged with feed rate variation [3] and for Titanium Ti6Al4V, hardness and microstructure also seem to remain practically unaffected [4]. In both works, a CNC lathe was used to promote very high feed rates, but naturally at the expense of reducing the range of geometric possibilities within the SPIF process. Camara [6] studied the influence of heat treatment on formability of 1050 aluminum sheets for SPIF. Hamilton and Jeswiet [5] formed 0.8 mm thick 3003 aluminum at feed rates ranging from (5080 to 8890 mm/min) and analyzed roughness and thickness distribution. Here, conclusions point to the possibility of grain refinement and consequently formability improvement.

In this work, feed rate will range from 1500 to 12,000 mm/min. The contribution to the state-of-art will be given by analyzing not only aluminum material (for the sake of comparison against the cited works) but also high-strength steels, namely dual-phase ones. The authors understand that analyzing this kind of material has importance in the context of possible market applications, like in automotive industry. Furthermore, the apparatus used allow the production of asymmetric parts within the range of speeds proposed. Formability and processing forces will be analyzed in the forthcoming sections.

Experimental tests

To overcome SPIF machine limitations, Alves de Sousa et al. [2] have developed and built a new concept of incremental forming machine based on parallel kinematics (Fig. 1). The so-called SPIF-A machine provides large payload (20kN) and flexibility driven by a 5°-of-freedom hydraulic Stewart platform on the kinematic side. A sixth degree of freedom would be the drilling one, but the tool was chosen to behave passively, freely rotating driven by friction forces only. Actuators are from hydraulic nature and controlling strategy is based on real-time machine [7] using Matlab/Simulink™ which allows performing the process using moderately high speed and 5-axis tool-path strategies.

Fig. 1
figure 1

SPIF-A Machine and forming table adapted to 250X250mm

Full size image

To evaluate the effects of tool tip velocity the frustrum of a cone geometry was considered (see Fig. 2, after Ghulam et al. [8]). This continuously varying-wall-angle geometry allows obtaining a clear comparison of forming forces, being additionally a very popular example to evaluate formability in terms of maximum wall angle. Complementarily, the sheets were grid-marked by electrochemical etching (Universal Marking Systems), which permitted to plot the fracture forming line for different feed rates.

Fig. 2
figure 2

Frustrum of cone geometry, as in Ghulam et al. [8]

Full size image

Two different materials were analyzed. Aluminium, AA1050-H111 series, and Dual-phase steel in three different grades, DP600, DP780 and DP1000. Apart from aluminum, these materials were hardly tested up to date given the lack of SPIF machinery able to withstand large forming forces. The chosen lubricant follows the conclusion by Azevedo et al. [9]. To keep the number of experiments at a reasonable number, several parameters need to be fixed for a clear comparison, as listed in Table 1.

Table 1 Fixed parameters during experiments
Full size table

These set of parameters generally allow for a reasonably good surface finish [10]. Bearing in mind these fixed experimental variables, a set of 4 different velocities (feed rates) was defined as in Table 2. Velocity 1 can be considered extremely slow, and serves to establish a reference result where no effects of feed rate may be observable. Velocity 4 can be considered sufficiently fast to produce SPIF parts in a cadence compatible with industrial needs. Such values are compatible with production of non-axisymmetrical parts by the SPIF-A, although not being studied here. Doing so, forming time can be reduced 5–6 times from velocity 1 to velocity 4, as shown in Fig. 3. Intermediate feed rates 2 and 3 will serve the purpose of establishing a trend for the results.

Table 2 Forming velocities
Full size table
Fig. 3
figure 3

Forming times regarding feed rate. Frustrum geometry depth can vary depending on the material employed

Full size image

Results

Forming forces

The first set of results to be analyzed are the forming forces. The SPIF-A machine is equipped with 3 load cells embed at the tool holder [2]. Vertical (compressive) and lateral (radial) forces were continuously recorded during forming process. As expected for SPIF process, compressive forces are approximately 3-times higher than lateral ones [11]. Recorded peak compressive forces are circa 470 N for aluminum and 2600, 3700 and 4300 N for DP600, 780 and 1000 respectively, which is in accordance the materials mechanical resistance.

Concerning AA1050-H111, it is interesting to note that compressive forming forces increase very clearly from the beginning of the manufacturing process until 50 % of depth (Fig. 4), then decrease until necking takes places at the cone wall with subsequent fracture. The effect of velocity on compressive forces appear to be not very significant (about 10 % variation) leading to a quicker increase of forces until the force peak and then to an also quicker force value decrease until failure. Regarding radial forces (Fig. 5), the evolution is positive practically until the end of forming (85–90 % depth), then decreasing due to necking. Furthermore, the effect of feed rate is only moderately pronounced at the last forming stages (70–100 % depth), with the lower feed rate yielding lower forces.

Fig. 4
figure 4

Vertical Force—Aluminium

Full size image
Fig. 5
figure 5

Radial Force—Aluminium

Full size image

Analyzing the results of dual-phase steels (Figs. 6, 7, 8, 9, 10 and 11), it is possible to check the positive evolution of compressive forces up to 35–40 % of forming depth, followed by a sudden decrease at the final forming stages due to necking. Radial forces have a more smooth increase during the whole manufacturing process. If one analyses the evolution of forces from 0 to 85 % forming depth, before the onset of material instability, it is possible to check that the influence of feed rate on force values is kept always limited to circa 5–10 %, which suggests for the range of velocities studied (up to 12 m/min) the effect on forming forces is perfectly acceptable for part production.

Fig. 6
figure 6

Vertical Force—DP 600

Full size image
Fig. 7
figure 7

Radial Force—DP 600

Full size image
Fig. 8
figure 8

Vertical Force—DP 780

Full size image
Fig. 9
figure 9

Radial Force—DP 780

Full size image
Fig. 10
figure 10

Vertical Force—DP1000

Full size image
Fig. 11
figure 11

Radial Force—DP 1000

Full size image

Formability

Incremental forming process is well known for its peculiar deformation mechanism and improved formability when compared to conventional press-forming operations [1, 12]. Traditional forming limit diagrams are of not much use for incremental forming purposes. One of the most used measures to evaluate a given material formability under SPIF is the maximum wall angle [13, 14].

Following this methodology, the formed frustrums of cone were 3D measured (Fig. 12) and the maximum wall angle evaluated using a CAD software. Several measurements were taken to ensure repeatability and the presented values represent the mean value with a 95 % confidence interval.

Fig. 12
figure 12

3D profile measurement

Full size image

Results are summarized in Fig. 13 and illustrated on Figs. 14 and 15. Oppositely to forming forces, results now give a clear idea of the effects deriving from feed rate.

Fig. 13
figure 13

Wall angle at maximum formed depth VS. feed rate

Full size image
Fig. 14
figure 14

AA1050-H111 Conical Frustums—Vel.1, Vel.2, Vel.3 and Vel.4 in sequence

Full size image
Fig. 15
figure 15

DP600 Conical Frustums—Vel.1, Vel.2, Vel.3 and Vel.4 in sequence

Full size image

Aluminum alloy shows insensitivity regarding the feed rate, which is in agreement with the conclusions from Ambrogio et al. [3]. The case of DP600 steel, on the other hand, denotes clearly a decrease in formability properties with a reduction of 68° to 55°. DP780 and DP1000 show even less formability as having a wall angle starting from 45° to 42° degrees respectively and decreasing to 39° and 37°, but not linearly. In fact, after several repetitions, the trend for the maximum wall angle was kept for DP780 and DP1000, with minimum wall angles occurring for the feed rate of 3500 mm/min. Summing up, DP600 material was the one showing greatest sensitivity to this forming parameter.

The beneficial effect of incremental forming on increasing material formability is well-known and well-studied by many authors, especially concerning aluminum alloys [12]. Typical forming limit curves (FLC) were found to be straight lines with negative slope. A better designation to these lines are Fracture Forming Lines [15]. In fact, FLC’s are determined mostly by necking, while in SPIF operations the continuous pressure from the tool during forming prevents and delay this phenomenon. Furthermore, due to the diffuse necking present in incrementally formed parts, the fracture forming line (FFL) seems more appropriate to characterize formability during incremental forming. In this sense, FFL’s were plotted for each material and for each feed rate studied, Figs. 16, 17 and 18.

Fig. 16
figure 16

Fracture Forming Line (FFL): AA1050-H111

Full size image
Fig. 17
figure 17

Fracture Forming Line (FFL): DP600

Full size image
Fig. 18
figure 18

Fracture Forming Line (FFL): DP780

Full size image

The scale in both axis were kept unaltered to provide a more easy comparison between the different graphics. For AA1050-H111 (Fig. 16) there is a high scattered set of points. Furthermore, it can be concluded that the effect of feed rate for this material is not significant which agrees with the results from Fig. 13, where the maximum wall angle for this material was kept constant for the range of parameters studied. Finally, the magnitude of strain observed is in accordance with values observed by Camara [6].

Focusing in dual-phase steels, there is no data available in literature for comparison. The softer DP600 (Fig. 17) shows more formability than the harder DP780 and DP1000 (Figs. 18 and 19). This is a straightforward conclusion. On the other hand, there is an evident sensitivity of the 3 grades concerning the feed rate. For DP1000, the decrease of formability is drastic, with limiting strains being reduced up to 3 times its initial magnitude. It becomes evident that feed rate must be carefully set when forming these types of steel to keep a good compromise between forming time and preventing necking and ultimately cracking.

Fig. 19
figure 19

Fracture Forming Line (FFL): DP1000

Full size image

Surface finish

A final set of measurements were performed to assess the roughness, and consequently, the surface finish of formed parts. Being an ultimate goal the dissemination of incremental forming processes into industrial environment, it’s important to assess the effect of quickening the process in the final quality. Finally, it is important to check the link between forming forces (Figs. 4, 5, 6, 7, 8, 9, 10 and 11) and final surface quality.

The roughness tests were conducted using an electromechanical Hommelwerke T1000 rugosimeter (Fig. 20). A support was created to ensure the parallelism of the wall and the fixed rugosimeter’s needle. Each piece went through 8 roughness tests at intervals of 45° around its center line. Roughness parameters were analyzed according to DIN4768 standard, namely the arithmetic mean roughness (Ra), the total height of the roughness profile (Rt) and the mean roughness depth (RzD). Results are summarized from Figs. 21, 22 and 23.

Fig. 20
figure 20

Measuring surface roughness

Full size image
Fig. 21
figure 21

Arithmetic mean roughness (Ra) for initial and formed surfaces (forming parameters: tool diameter—16 mm, step-down 0.05 mm)

Full size image
Fig. 22
figure 22

Total height of roughness profile (Rt) for initial and formed surfaces (forming parameters: tool diameter—16 mm, step-down 0.05 mm)

Full size image
Fig. 23
figure 23

Mean roughness depth (RzD) for initial and formed surfaces (forming parameters: tool diameter—16 mm, step-down 0.05 mm)

Full size image

Initial roughness coming from rolling conditions was measured as well for the sake of comparison. For dual-phase steel grades, roughness parameters are higher than aluminum. Bearing that in mind, it is possible to check that forming operation will increase roughness in case of AA1050-H111 whereas it will decrease it for DP steels. Then, concerning the influence of feed rate, it is possible to notice the expected rise in roughness values as the punch travel velocity increases. However, it is not an accentuated variation. For DP600 steel, the maximum value of Ra, for a forming operation with 12,000 mm/min, is even lower than the initial value from rolling. For Rt and RzD measurements, they are not much higher than the initial ones. Nevertheless, the trend is clear: increasing feed rate will increase roughness and consequently deteriorate surface finish. For the harder grades DP780 and DP1000, values for all parameters surpasses the initial ones. For DP1000, roughness parameters Rt and RzD achieve high values and a not smooth surface finish. In this sense, one can conclude about the sensitiveness of surface finish depending on the feed rate used. Finally, and concerning aluminium AA1050-H111, all parameters are kept approximately constant even for the highest feed rate. Once again, confirming the previous results from formability tests, this material appears to be insensitive regards high punch travel velocities.

Conclusions

High production times are one of the factors impairing wide-spread use of incremental forming technologies for industrial applications, by limiting batch sizes for a fixed timespan. In this work, the effects of increasing the feed rate were studied in terms of forming forces, formability and surface finish. To minimize the number of experiments, tool diameter, lubricants and tool-path strategy were kept constant. However, the reader should keep in mind that a larger step down can also reduce forming times. Such effect has been studied for instance by Li et al. [16].

Two types of material were assessed: one grade of aluminium AA1050-H111 and three grades of Dual-phase steel: DP600, DP780 and DP1000. Few results for these latter materials are present in literature given machine limitations, usually adapted milling machines, which is not the case of the SPIFA machine.

Results coming from forming forces, maximum wall angle of a frustrum geometry, fracture forming lines and roughness were clear and in agreement: increasing feed-rate reduces formability and deteriorates surface finish for dual-phase steels. Aluminum 1050-H111 appears to be insensitive to feed rate variations.

Summing up, it is naturally feasible to reduce forming times of simple parts, using a dedicated SPIF machine. For simple geometries, parts can be formed in a matter of minutes, allowing the production of medium sized batches in an acceptable timespan. However, reducing forming times for steel material will bring some pay-off which in this case would be less formability and worst surface finish. It will be up to the designer to choose an acceptable feed rate to provide reliable and well finished parts to the final consumer.