1 Introduction

Roller hemming with the help of industrial robots is an established process especially in automotive industry for joining outer body panels with higher ductility by bending the flange of an outer panel over an inner blank [1]. In comparison to conventional hemming, roller hemming does not require individual form tools and therefore has a higher flexibility. Due to the specific designs and the needs of the customers, these parts can often contain sharp design edges and curvatures. The joining along the resulting complex contours represents a challenge for joining by hemming since defects like cracks or wrinkles can occur [2]. This is primarily due to changes in the flange surface area. (cf. Fig. 7). If the surface area is increasing or decreasing, a material flow in the flange area is necessary to suppress wrinkles or cracks. In common hemming processes, the deformation is concentrated to bending of the hemline. But, if there is no material flow in the flange area, these defects arise. So, to successfully form complex-shaped flanges, a material flow in the flange area additionally to bending in the edge area is required.

ISF for hemming represents a promising alternative technology by means of joining by forming since ISF is characterized by local and incremental forming mechanisms and high formability [3]. Hems of closed concave flanges have already been successfully formed as part of research on process integration of ISF for joining closed flanges [4]. Therefore, the potential for joining with ISF along curved edges will be investigated further and in more detail within the scope of this paper to gain process knowledge with a special focus on the material flow in the flange area.

2 Background and state of the art

Hemming is an established process for joining visual parts in the car industry. Especially roller hemming is a commonly used technology in combination with flexible manufacturing systems. In roller hemming, the joining process is performed with robots and universal roller tools, where a robot guides the roller along the hemming line and bends the flange successively. This makes it a very flexible process in comparison to conventional hemming since no individual tools are required.

Similar to conventional hemming, the angle of the tool changes in several pre-hemming steps, as shown in Fig. 1, to form the final closed hem. With regard to the tool motion, there is only a translatory movement combined with a free rotation of the roller due to the friction between the flange and the tool.

Fig. 1
figure 1

Schematic roller hemming process in three hemming steps (left) and door hemming assembly (right)

Full size image

The hemming process often reaches its process limits for curved hem lines due to the resulting geometrical conditions in the hem. While for convex edges (shrink hemming), wrinkling is a common error, and in concave edges (stretch hemming), the material tends to crack, as illustrated in Fig. 2 [2, 5].

Fig. 2
figure 2

Typical defects for hemming of concave and convex contours after [2]

Full size image

Wrinkles mainly occur if the radius of the sheet contour is decreasing over the hem length for convex shapes since the change in surface area results in excess material. In case the shape is concave, the increasing radius over the hem length results in failure of the sheet metal by cracking.

An early instance of this correlation can be seen to develop at the flanging stage which is the prior process to hemming. The outer contours of the flange change from the initial sheet metal edge to the flange edge, when they are bent from the flat sheet (0°) to 90° flanges. For concave contours, the edge length of the flange is increasing and so is the surface area leading to the term of stretch flanging (Fig. 3, left). The same relation applies for shrink flanging where the edge length and surface area are decreasing (Fig. 3, right).

Fig. 3
figure 3

Schematic drawing of stretch flanging (left) and shrink flanging (right) after [6]

Full size image

The same geometrical dependencies apply for the following step of hemming. In the case of hemming concave edges (stretch hemming), every point on the flange undergoes an increase in radius, as shown in Fig. 4. This results in tangential tensile stresses which can result in cracks. For convex edges (shrink hemming), the opposite is the case and there is a decrease in radius causing tangential compressive stresses, leading to wrinkles.

Fig. 4
figure 4

Radius changes in stretch and shrink hemming

Full size image

The observed strains during the deformation in roller hemming are typical for a bending process and are localized in the bending radius [7]. In past studies on roller hemming of convex edges, wrinkling rather than a change in flange length or sheet thickness was observed [8, 9]. Additionally, local elongation of the flange in rolling direction causes tangential compressive stresses and results in wrinkling or surface waves during pre- and final hemming [7, 10]. Hence, this adds up to the geometrical effect for convex hem contours.

The most recent papers on roller hemming focus on optimizing the positioning and trajectory of the tools. Huang et al. implemented a hand-eye sensor module to perform vision-guided positioning of the hemming path in dies-free roller hemming. They managed to reduce roll-in/ roll-out/ errors to 0.2 mm thus meeting industry standards for die-assisted roller hemming [11].

Wang et al. developed an algorithm for determining roller pose and kinematic trajectory of industrial robot roller hemming. They conducted trials on curved surface-curved edge hems and used differential geometry theory to obtain the pose and trajectory of the roller. Verification was also conducted by FE-simulation, but deformation mechanism or material flow analysis was not part of the research [12].

There are also recent works which research alternative hemming techniques. For example, Pereira et al. and Kasaei et al. developed a novel hemming process called hole hemming. In this process, dissimilar materials are joined by flanging and hemming pre-drilled holes with two dedicated punches. This procedure shows a great potential for joining materials with a difference in ductility, as their further works with magnesium and aluminum alloy sheets showed [13, 14]. In contrast to conventional hemming methods, however, the joining partners are hemmed within the component and not along the edges. The areas of application are therefore probably more comparable with clinching than with roller hemming.

Similar to roller hemming, incremental sheet forming (ISF) is a local and CNC-controlled forming process with no geometry-specific tooling. ISF is characterized by high flexibility and fast implementation from CAD design to finished component. The local and incremental forming character of ISF is one of the main reasons to achieve increased formability when compared to conventional forming processes [15].

While ISF is part of research for the at least past two decades, its applications are constantly being expanded. Voswinkel et al. for example showed that the limits of conventional flanging, as the pre-process for hemming, could be exceeded significantly for curved contours by using ISF. They conducted incremental flanging experiments with different tool path strategies on stretch and shrink flanges with radii of 20, 50, and 100 mm. They concluded that the flange length-to-radius ratio for stretch flanging could be exceeded from 0.725 up to 1 and in case of shrink flanging from 0.05 to 1 as well. In their further work, they were even able to reach flange length-to-radius ratios of 2 for both flange geometries by using a new adaptive blank holder [6, 16].

The newest work for incremental flanging by Lopez-Fernandez et al. addresses the failure behavior of convex flanges. Flanges with radii of 20 and 45 mm and flange widths between 20 and 72 mm were analyzed, and the influence on wrinkling was determined. A practical process window was derived by this. They identified that wrinkling occurs depending on the compressive stress at the edge of the flange. With the help of a numerical model and measurement of the minor principal stresses, they identified a wrinkling limit [17].

Silva et al. investigated the limit-forming ratio of hole flanges formed by means of ISF compared to conventional press-working flanging. They stated that incremental flanging does not always present higher limit-forming ratios but that there is a dependency on the material. In comparison to conventional flanging, incremental flanging is favorable if the fracture limit is significantly higher than the forming limit [18].

Cristino et al. extended the range of flange geometries and formed square flanges with the help of ISF. They varied the length and corner radius of square pre-cut holes and investigated the deformation history along different sections of the flanges. They proved the general feasibility and concluded that fracture in the corners is caused by sheet thinning under biaxial stress [19].

The summarized works demonstrate the extension of ISF applications towards flanging. The present work aims to build on this progress and go one step further by producing hems from the incrementally formed flanges. In theory, ISF offers a potential for such a hemming operation in terms of formability and wrinkling compared to conventional hemming processes. In contrast to roller hemming, the hem is created by the movement of a simple and generally non-rotating forming head (cylindrical metal rod with a hemispherical tip). While in roller hemming, the roller moves over the flange in one motion (Fig. 5, left); in ISF, the tool moves in parallel and perpendicular direction to the hemming line (Fig. 5, right).

Fig. 5
figure 5

Tool motions of roller hemming (left) and ISF-hemming (right)

Full size image

The tool movement on the sheet metal causes deformation through combined shearing and stretching in tool direction. The deformation is increasing with the successive forming. Shear also occurs in the direction perpendicular to the tool movement leading to a pile up of material in the direction towards the undeformed sheet metal. A material flow from thickness direction is proven to be present [3, 20].

The idea of using the benefits of high formability for a new concept of joining was initially investigated for ISF-formed hole flanges. Partial form-fit joining and complete form-fit joining were reviewed. Blanks with different initial hole diameters (d0) were first formed to flanges with an inner diameter (d1) of 100 mm using a 20 mm ISF tool and a vertical step size (zV) of 1 mm. Due to the different initial diameters, the resulting flange lengths varied between 11 and 22 mm. In the final step these flanges were formed to hems using the same tool, a horizontal step size (zH) of 1 mm and no offset respectively intrusion depth. The formed hems are shown in Fig. 6.

Fig. 6
figure 6

Results for hemming closed geometries for different diameters and flange length

Full size image

All joints could be realized without failure producing form-fit and partially force-fit joining. The main focus of the evaluation was on the geometry of the hems. As expected, the resulting length of the hem correlated with the initial hole diameter in a way that longer hems were formed with smaller initial hole diameters. Additionally, increased spring-back on the topside of the hem as well as thinning was observed with increased flange lengths. The authors argued that the geometrical deviations were related to tangential stresses, which are increased with greater outer diameters after forming. A correlation between the initial hole diameter and the gap between hem radius and inner sheet could not be established. Overall, the study on form-fit incremental joining for closed concave geometries showed a general feasibility of joining by ISF but did not further investigate process mechanisms regarding deformation and material flow during hemming [4].

As no other work on hemming using incremental forming has been found in the literature, the present work aims to continue research on this novel hemming technique and gain more insights into material distribution in incrementally formed hems. In the case of flanging and hemming concave and convex geometries, an increased material flow in thickness direction might help reducing cracks and wrinkling as mentioned above.

3 Approach and scope of work

The state of the art points out that the challenge in hemming lies in joining curved shaped contours because of a change in radii over the flange length. This correlates with a change in surface area of the flange, as depicted in Fig. 7, for both stretch and shrink hemming. To compensate this change in surface area, due to geometrical conditions, a material flow in the flange area is necessary. This could suppress winkles because of excess material or cracks because of a lack of material in the flange area.

Fig. 7
figure 7

Surface area changes in stretch and shrink hemming

Full size image

Since the tool in ISF moves along two axes, parallel to the hemming line and perpendicular, it is expected that with ISF, deformation can be applied in the whole flange surface area.

Previous work has shown the general feasibility of ISF for hemming [4]. But, the underlying material flow and deformation mechanism have not been investigated and will be studied within the scope of this work. The aim is to prove that a significant material flow in the whole flange area is present besides the straining in the bending radius. Therefore, the following assumptions should be proven:

  • For shrink hemming, it is expected to achieve a radial material flow, which allows the excessive material to be distributed in radial direction. Alternatively or additionally, an increase in thickness will be measured.

  • For stretch hemming, it is expected that a material flow from thickness into circumferential direction can be realized to distribute the material from thickness in circumferential direction to compensate for the enlarged flange surface.

To prove these assumptions, experiments on closed flange geometries will be conducted. They allow to analyze the material flow in the flange directly through measuring a change in thickness and flange length.

4 Material and experimental procedure

The material used in this work is cold rolled deep drawing steel DC04 (1.0338) with a thickness of 1 mm. DC04 is characterized by its high formability with adequate strength and is therefore used for internal and external automotive parts. The typical processing methods are bending and deep drawing, but it has also proven itself for ISF applications. The properties of DC04 are listed in Table 1, and the stress-strain curve is shown in Fig. 8.

Table 1 Data sheet for mild steel DC04
Full size table
Fig. 8
figure 8

True stress-true strain relation of DC04 sheet

Full size image

In this experimental study, joining by ISF is applied to shrink and stretch flanges of closed circular geometries. Joining of shrink flanges corresponds to a radius reduction, and therefore, the hem is formed to the inside of the circle. For stretched flanges, it is the opposite; therefore, the hem is formed to the outside. Figure 9 shows the geometry for both types of specimens.

Fig. 9
figure 9

Geometry of the specimens for stretch and shrink hemming

Full size image

The experiments are carried out on circular geometries with initial radii (r0) of 200 mm and 50 mm. In addition, two different flange lengths are investigated for each radius as shown in Table 2. A flange length-to-radius ratio is defined to indicate the degree of difficulty for joining. A greater ratio implies a greater challenge for joining, as greater flange lengths as well as smaller radii are expected to result in greater changes in surface area (cf. Fig. 7). The parameters of the analyzed specimens are defined in Fig. 10 and summarized in Table 2.

Table 2 Design of experiment
Full size table
Fig. 10
figure 10

Geometric definitions of the flanges

Full size image

The flanging itself, as a pre-step to hemming, is also carried out with ISF. Therefore, cut sheet metal blanks are fixed between a die and a blank holder as shown in Figs. 11 and 12. The continuous motion of the CN controlled forming tool along the part contour with an increasing depth forms the flange. The tool circles around the part incrementally, changing direction after each full circle. The incremental forming parameters are chosen based on the literature: Single-stage forming with a tool diameter of 20 mm and a vertical step size (zV) of 1 mm is set as the tool path strategy for flanging (cf. Fig. 13) [4, 6].

Fig. 11
figure 11

Experimental setup for stretch hemming with R50

Full size image
Fig. 12
figure 12

Experimental setup for shrink hemming with R200

Full size image
Fig. 13
figure 13

Exemplary tool paths for ISF-flanging and hemming

Full size image

To proceed with hemming, the sheets are clamped upside down, and an inner sheet is inserted after the flange is formed. The dimensions of the inner sheet are chosen so that a gap (gF) of 2 mm to the flange remains. Based on the literature, successive horizontal lanes with a horizontal step size (zH) of 1 mm and no vertical offset realize the hemming process (cf. Fig. 13) [4]. A tool diameter of 30 mm proved to be suitable for the larger geometries with R200. As preliminary trials showed, the specimens with a radius of 50 mm, and subsequently, shorter flange lengths required a smaller tool diameter of 20 mm for successful hemming. Figure 14 shows the general process steps for ISF-hemming of this work.

Fig. 14
figure 14

ISF-hemming procedure based on the example of shrink geometry with R50

Full size image

Based on the geometric considerations of closed circular specimens, a material flow should be expressed by a change of sheet thickness or flange length as long as no failure occurs. For this purpose, after forming, the samples are separated to investigate the cross-section under the microscope. The sheet thickness is analyzed for every flange and hem specimen. As shown in Fig. 15 on the left, the measurement is carried out between the outermost point of the flange respectively hem to the hemming radius. For analyzing the length variation, a reference point outside the forming zone is marked on the blanks, which allows reproducible measurements of the resulting flange length after each flanging and hemming including the bending radius. The measurement method is also shown in Fig. 15 on the right.

Fig. 15
figure 15

Measurement length for sheet thickness analysis (left) and measurement method for length variation analysis (right)

Full size image

5 Results

In order to start with the general appearance of the specimens, Fig. 16 shows some examples of specimens, which were successfully flanged and hemmed with ISF. All tested flanges and hems of both geometries were formed without failure. Nevertheless, the shrink hems with a radius of 50 mm and 10 mm flange length (R50FL10) and a radius of 200 mm with 20 mm flange length (R200FL20) show a rough surface but no wrinkling.

Fig. 16
figure 16

Examples of successful flanged and hemmed specimens by means of ISF with a radius of 50 mm

Full size image

To address the preceding assumptions and to investigate the material flow, the sheet thickness and change in flange length are measured in this work as stated in the previous section. Figure 17 shows the results for two hems of different r0, each with two flange lengths for stretch hemming. The sheet thickness is plotted over the measured length for both the flange and the hem. The initial sheet thickness of 1 mm and the forming direction are also marked in the diagrams. It should be noted that the resulting measured length is shorter than the initial flange length. This is due to the length lost to the bending radius, which is not part of the thickness measurement as mentioned above.

Fig. 17
figure 17

Sheet thickness distribution for stretch hemmed geometries with different radii and flange lengths

Full size image

Considering the flanges first, the measurements show a considerable reduction in thickness up to 15% at the outer edge of the flange for all specimens. Additionally, the respective longer flanges of both radii also show a tendency for thinning in the middle of the measured length. Sheet thicknesses between 0.9 and 0.95 mm are found there. Moreover, in the case of the rather long flange for R200FL20, a local drop of the thickness in the middle of the flange can be seen.

In all four cases, the hems display a tendency for further thinning in comparison to the corresponding flanges. In general, the graphs of the hems show the same characteristics as for the flanges. The thickness decreases in forming direction and reaches the minimum of ca. 0.7 mm on average at the outer edge of the hem. It is noticeable that there is, on average, more thinning for both the flange and the hem with a greater initial flange length for the same radius. In other words, a higher flange length-to-radius ratio leads to more thinning for stretch hemming.

Similarly, Fig. 18 shows the results of the sheet thickness measurements for the shrink hemmed geometry. As for the stretch hemmed geometry, the measured length deviates from the initial flange length because the arc length of the bending radius is not considered in the thickness measurement. For the larger specimens with a radius of 200 mm, the flanging process also results in thinning of the flange, as it is the case for the stretch hemmed geometry. Contrary to this, the smaller specimens with a radius of 50 mm show a trend towards thickening compared to the initial sheet thickness. This already leads to sheet thicknesses up to 1.1 mm after flanging. The curves show a drop in sheet thickness towards the outer edge of the flange respectively hem as seen for the stretch hemmed specimens. After the hemming process, the sheet thickness is increased in all four cases. For the larger geometry, the hems are ca. 0.1 mm thicker than their corresponding flange. In most sections, the thickness of the hem even exceeds the initial sheet thickness of the blank. For both radii, it becomes obvious that the extent of the thickening increases with a greater flange length respectively with a greater flange length-to-radius ratio. This is strongly evident in the hems with a smaller radius for both flange lengths. The final sheet thickness goes up to 1.3 mm and exceeds the initial thickness over almost the entire measured length.

Fig. 18
figure 18

Sheet thickness distribution for shrink hemmed geometries with different radii and flange lengths

Full size image

To further describe a possible material flow, the change in length for both the flanging and the hemming process is analyzed in detail. In order to obtain a general comparability between the different specimens, the measured length is then referred to the initial flange. This allows conclusions about the change in length regardless of the different initial lengths for each radius. The results for every specimen referred to the initial length are presented in Fig. 19.

Fig. 19
figure 19

Elongation of flanges and hems as referred to initial dimension for all specimens

Full size image

The flanges are considered first. Flanging by means of ISF results in an increase in the flange length for all specimens. The shrink flanges show a greater increase in length than their counterparts of the stretch geometry referred to the initial flange length. On average, a 12.5% increase in length after flanging is observed for shrink geometries, while a maximum of ca. 8% is measured for the stretched geometries. Referring to the hems, with the exception of the shrink hem R200FL10, the resulting length increases significantly due to the hemming process after flanging. Again, the specimens of the shrink geometry in general show a tendency for greater elongation than the stretched hemmed parts. The biggest difference is found for R50FL10, where the shrink hem is elongated by ca. 27%, while the stretched one is ca. 8% longer. A correlation between flange length and elongation can be seen, with the stretch hemmed specimens showing less elongation with greater flange lengths. Contrary to this, the shrink hemmed specimens lengthen more when the flange length is increased. Moreover, when comparing the two different radii, it becomes obvious that the flanges and hems with the smaller radius undergo a greater percentage elongation. This is similar for both geometries.

6 Discussion

The results support the hypotheses set out at the beginning of the work. The specimens showed no signs of failure, encouraging a discussion on the material flow behavior. As expected, hemming of concave and convex edges by means of ISF leads to different sheet thickness distributions in the hems. For stretch hemming, an overall tendency for thinning could be observed. This was expected due to the geometrical dependencies and volume constancy. Generally, high thinning is possible without failure in ISF due to the higher formability. This is an advantage for stretch hemming where surface area increase is a challenge.

On the other hand, in the case of shrink hemming, a thickening was measured in all specimens. This is surprising because, in contrast to deep drawing, where thickening can take place in a controlled manner due to the high tool contact at the blank holder, there is only local forming in ISF. This is happening to such an extent, that even though the flanging process induces typical thinning, as seen for the larger radius of 200 mm, the later hemming compensates for this. As a result, the final average sheet thickness exceeds the initial sheet thickness in most sections. Moreover, the observations become more apparent with more challenging flange length-to-radius ratio, such as smaller radii or greater flange lengths. Therefore, the maximum thickening up to 1.3 mm is seen for the R50FL10 specimen.

As mentioned for the larger specimens R200FL20, a drop of thickness in the middle section of the flanges and hems could be observed for both stretch and shrink hemming. This observation cannot yet be fully explained at this point. There may be a relation to the deformation mechanism of ISF. Jackson et al. stated that one component of the ISF deformation mechanism is shear perpendicular to the tool direction, which in their work also led to material being displaced with successive laps when forming a truncated cone [3]. The result was a bulge and increased sheet thickness of the part at the end of the tool path. On closer inspection of the cross-section, it is noticeable that the hems show the effects of a material drag. As seen in Fig. 20, the outer edge of the hems deviates from a perfectly straight edge. The affected zone is around one-third of the sheet thickness from the point of tool contact. The material protrudes over the theoretically straight edge of the hem as if it had been drawn. This can also be explained by the observations of Jackson et al. They pointed out that the shear in tool direction as well as perpendicular to it is greatest at the tool-contact surface [3]. This might then also explain the thinning in the longer flanges of R200FL20 because more consecutive tool passes are able to add up to this effect.

Fig. 20
figure 20

Exemplary material drag at the outer edge of hem

Full size image

Possible inaccuracies in the measurement method should also be addressed at this point. Due to the separation and sanding of the specimens, it is not possible to completely prevent a burr from remaining. This makes precise sheet thickness measurements difficult and requires a large number of measurement points to compensate for such deviations. In addition, the sheet thickness must always be measured perpendicular to the upper and lower edge of the sheet, which becomes more complicated with curved flanges. These factors lead to measurement uncertainties.

The analysis for the elongation of the flanges and hems also presents unexpected results. The stretch hemmed geometries show a slight elongation for the flanges and hems. This shows one of the advantages of ISF. Compared to conventional processes, where the flanges tend to remain short in order to prevent cracks, the higher formability of ISF even leads to a lengthening of the flanges and therefore exceeding the process limits of conventional forming. Nevertheless, the increase in flange length goes hand in hand with thinning as it was to be expected because of the geometric considerations as well as the volume constancy.

With regard to the assumptions made at the beginning of this paper, the results of the shrink hemmed geometry provide further confirmation. First of all, wrinkling was prevented. The length variation indicates a greater elongation for the shrink flanges and hems than for the stretch hemmed specimens. At the same time, an increase in thickness was detected for the shrink geometry. Both points support the hypothesis for a material flow in thickness and radial direction preventing wrinkles. This behavior increases with smaller radii and greater flange lengths. ISF therefore offers the potential for successful hemming of small radii with relatively large flange lengths.

At this point, it must be conceded that, even though a compensation of wrinkles by thickening and flange elongation is possible for a flange length-to-radius ratio of 0.1, greater ratios for shrink hemming as seen for R50FL10 (cf. Fig. 16, right) lead to visual imperfections. During the hemming, it comes to material accumulations as precursors for wrinkles. These accumulations of material are then flattened as the tool progresses. As a result, the hem edge turns out to be locally elongated in a regular pattern. This represents an inconsistency in the measurements, since the length variation for all specimens is compared at single, identical point.

7 Conclusion

In this study, incremental sheet forming (ISF) was utilized to implement a hemming process. Experiments were carried out on closed curved hem geometries, where the flanges as well as the hems were formed by ISF. To analyze a potential material flow, the sheet thickness distribution and elongation for both flanges and hems were investigated. In order to combine and summarize the observations from both analyses, the following points can be noted:

  • In the case of stretch hemming, the flange surface is enlarged, and hence the flange is expected to compensate this by thinning and shortening when considering conventional methods. However, hemming by means of ISF shows different results. Supporting the hypothesis of a material flow from thickness into tangential direction, a significant thinning and elongation without failure in the hems are observed for stretch hemming.

  • The thinning increases analogously with the increase in surface area as observed for longer flange lengths. Moreover, an effect of radially drawing the material with the ISF tool can be assumed.

  • The results become even more apparent when looking at shrink hemming. Because of the excess material, shrink hemming is sensitive to producing wrinkles when smaller radii or greater flange lengths need to be achieved. In the case of roller hemming, wrinkling is further increased by inducing tangential compressive stresses. It is hypothesized that wrinkling can be compensated either by a material flow from tangential direction into thickness or by elongation in radial direction, which conventional hemming processes cannot provide.

  • At this point, the results demonstrate the great potential of ISF-hemming process in which specimens where formed without failure. Thickening along the entire flange was detected together with a great increase in flange length.

  • Both observations support the hypothesis of an increased material flow induced by ISF. On the one hand, the presence of excess material is avoided through an increase in thickness, and on the other hand, wrinkling is prevented by an increase in the flange length allowing the material to flow to compensate for this increasing dimension. This could be attributable to both the material flow perpendicular to tool motion and the radial components of the tool path.

  • It is observed that a more challenging flange length-to-radius ratio (greater flange lengths or smaller radii) amplifies the changes in sheet thickness and flange length.