Investigation of Effects of Material Dimensions on Wrinkling in Flexforming Process

Sheet metal forming processes are very common manufacturing and leading processes in automotive and aerospace industries. Flexforming is one of the sheet metal forming processes which is preferable due to its exible manufacturing capabilities and its ability to produce multiple parts simultaneously. Convex contoured shaped parts are very much used in aerospace structures which are mostly produced by exforming. Wrinkling is a characteristic defect for those kinds of parts. Prediction of wrinkling before manufacturing is highly crucial in order to reduce scrap rates, labor time, and other unexpected costs. In this research work, extensive amounts of experiments are conducted on exforming press, and the process parameters such as material condition, contour radius, ange length, and material thickness which induce wrinkling are investigated in detail. Results have shown that sheet thickness is the most effective parameter, and as the sheet thickness is increased, wrinkling tendency is reduced extensively. Besides, increasing convex contour radius decreases wrinkling occurrence. Experimental ndings are then used to generate wrinkling limit diagrams in which safety and failure zones are specied for different material conditions and sheet thicknesses. The developed diagrams might help to designer who can design defect free parts, reduce scrap rates, and reduce production costs signicantly.


Introduction
The fundamental principal of hydroforming process is to apply pressurized uid to a sheet to form a part [1]. Basically, pressurized uid is used to form the part and a single die is generally used. In diaphragm forming (or exforming), which is a special form of hydroforming, a force is produced by pressurizing uid in a closed circle uid system and this force is used to form the sheet. In the process, the sheet metal is placed onto a sheet metal die (usually the lower die) and the diaphragm including uid reservoir is located at top. Then the uid is pressurized and the diaphragm is moved towards the lower die. When the diaphragm is moved, the sheet is pressed between the diaphragm and the lower die. With the increase of pressure, the elastic diaphragm behaves like the upper die and the sheet metal is drawn into the lower die, completing forming process [2,3].
Flexibility of the diaphragm accommodate multiple parts production due to its behavior like the upper die.
For this reason, parts having different shapes can easily be manufactured [4]. It is a fact that aerospace industry has many different parts at limited quantity. Hereby, this method is widely used [5] and provide great advantages in terms of cost for aerospace industry. Parts produced by exforming process have better surface qualities than the parts produced by other methods [6]. Besides, the usage of hydraulic pressure instead of the upper die, helps complex and di cult shaped parts production. In a conventional sheet metal forming operation, a die set containing an upper die and a lower die is needed, whereas in this method, only a lower die is su cient. Elimination of the upper die and labor cost savings are the biggest advantageous of the process. Near to net shape geometry is almost achieved [7]. Thickness thinning is rather smaller with respect to conventional method and mostly homogeneous thickness distribution is obtained [8].
The common failure problem in exforming is wrinkling. Wrinkling is a critical issue for sheet metal parts and developed generally in the wall or ange of the part. It is mostly seen in convex contour anges because of sheet thickening or building up. The main reason is that the length of convex contour ange after production is much smaller than the original ange length [9].
Hatipoglu et al. [10] investigated the effect of some parameters, such as sheet thickness, die bend radius, ange length, and the rolling direction on forming of AA2024-T3 sheets both for straight and curved dies.
The nite element model of the exforming was constructed. They found that the ange length as well as the rolling direction for this aluminum alloy is seen to have insigni cant effects on the springback, whereas the die radius and the sheet thickness have signi cant effects. Besides the nite element results were in an agreement with these experimental results. Abedrabbo et al. [11] studied the effect of contour radius on the wrinkling. They found that an increase in contour radius decreases wrinkling. Wang et al. [12] determined a mathematical model that shows the relationship between the ange length and strain.
The results reveal that an increase in ange length raises wrinkles with a ratio. Sun et al. [13] experimentally studied the wrinkles occurred in rubber forming of convex shaped parts with Ti-15-3 material and examined them in nite element analysis. As a result of the study, they observed that the ange length and rubber have an effect on wrinkling. Liu et al. [14] examined the problem of wrinkling in hydroforming process both experimentally and nite element analysis. They determined the effect of material stress and diameter-thickness ratio of parts on wrinkling. Feyissa and Kumar [15] investigated the forming cryorolled AA5083 sheet material by hydroforming. They found out that parts with complex geometries formed by hydroforming method are more successful than the parts produced by conventional methods. Although wrinkling defect is critical for exforming process, studies on this subject are limited. More investigation should be done.
In this research work, wrinkling problem for convex contour parts was investigated experimentally. The wrinkling behaviors of AA2024-O and AA2024-W by exforming were determined for convex contour structures. Effects of thickness, ange length, and radius of convex contour on wrinkling in exforming were investigated. The purpose of the research was to predict wrinkling before manufacturing. Otherwise, tremendous amount of effort and costs could be wasted. If the parameters that in uence wrinkling are understood accurately, defect free parts can easily be manufactured. Designer can design the most accurate die with less modi cations, and with less time and cost savings.

Tensile Test
Tensile tests were performed in order to determine mechanical properties of sheet materials. All the tests were conducted by Zwick/Roell model Z100 machine with a 100 mm/min deformation speed at different directions. True stress vs. true strain diagrams for different alloy conditions were plotted in Fig. 1.
It is clearly seen that AA2024-W condition has higher strength than AA2024-O condition. The material was brought to W condition from O condition by heat treatment. W condition is an unstable intermediate condition which can also be observed in hardening behavior in the gure.
Mechanical properties such as yield strength, ultimate tensile strength, elongation, and anisotropy coe cients obtained from tensile tests are listed in Table 1.

Determination of Forming Limit Diagram
Forming Limit Diagram (FLD) provides failure limits at the onset of necking. For both aluminum alloy conditions, forming limit diagrams were determined. The FLDs were constructed using the Nakajima method [16]. The test samples were formed by a 600 kN Zwick model BUP 600 test machine. Deformations were measured optically and FLDs were plotted as seen in Fig. 2. The FLD results states that AA2024-O has better formability than AA2024-W which is also a well-known fact from shop oor practice.

Experimental Methods and Equipment
The purpose of experiments is to determine the effects of parameters which are mainly part dimensions on wrinkling tendency. All experiments were conducted at 70 MPa forming pressure with a constant oil temperature of 45°C, in 34°C ambient temperature. The total duration of forming process took about 3 min. and the part was exposed to maximum pressure for 3s. Experimental parameters, namely, material thickness (T), radius of contour (R), and ange length (F) are displayed in Fig. 3.
The experimental parameters were chosen to be the most effective ones on wrinkling and their ranges were determined by their incidences in aircraft parts designs. AA2024 is the mostly used material for ex formed parts and parts are formed either in O, or W condition. Therefore, both AA2024-O and AA2024-W were used in the experiments. Experimental parameters are summarized in Table 2. Bending radius of a constant value, 6.0 mm was selected. Bending radius (mm) 6.0 Flat patterns of specimens were adjusted such that the desired ange lengths were obtained after the forming. Also for meaningful comparisons, specimen bending arc lengths (130 mm) are kept identical since they are very effective in wrinkle formation.
A single die was designed and manufactured for the experiments (Fig. 7a). This die was capable of forming all of the con gurations in 24 runs (2 material x 4 thickness x 3 ange length). Specimens are attached on this die by the help of guide pins and then formed. Figure 7b shows formed specimens.

Results And Discussion
In this study, the effects of determined experimental parameters on ex formed parts were examined and compared Fig. 8 shows one of the formed specimens as a sample which has AA2024-W material, 1.27 mm thickness, 125 mm contour radius, and 40 mm ange length. As seen, two wrinkles were observed after forming.
The size of a wrinkle can be measured by two parameters which are height T, and width W (Fig. 9).
After performing the 24 runs, wrinkle sizes were measured and tabulated in Table 3. Some samples could not be measured due to their crushed forms. For better visualization, the ange edges of specimens were drawn on plots in Fig. 10-13. Those plots show the effects of experimental parameters on wrinkling clearly.
In Fig. 10, one can see the effect of sheet thickness on wrinkling. As the sheet gets thicker, the number of wrinkles decreases, and the size of wrinkles get bigger, and they disappear eventually Wrinkling phenomena can be thought as a sort of buckling in which thicker structures resist more. Figure 11 shows the effect of ange length on wrinkling. Number of wrinkles increases as the ange length is increased. The reason behind this trend can be explained by considering the arc length of ange edge before and after forming. Before forming, in which the part was at, the arc length of ange edge was higher than in its nal form. As ange length increases, this difference gets higher in which more material is forced to take the nal shape, resulting more compressive stresses and more wrinkle formations inevitably.
In Fig. 12, specimens bent in different contour radiuses are seen. As contour radius gets larger, wrinkle severity decreases. Again the reason can be explained by referring to previous discussion about the change in arc length of ange edge before and after bending. The change becomes insigni cant as contour radius gets larger, converging zero in straight bending, which is wrinkle-free.
The effect of alloy condition on wrinkling is signi cant. Figure 13 shows the difference between W and O conditions of AA2024 for a speci c con guration W condition, which is harder, has a wrinkle, whereas O condition is awless. The reason is W is a heat treatment condition in which grains are small and homogeneous but they are also unstable causing bigger wrinkles on the formed part.
Above plots clearly reveal that at higher contour radiuses, wrinkles disappear, and when the ange length is increased, they appear again. The thickness is a positive effect on reducing wrinkles.
The aim of this study was to represent the experiment results in the form of wrinkling limit diagrams (WLD). These diagrams will provide safety and failure zones for a given geometry and material. Figure 14 shows a sample WLD for AA2024-O, 0.63 mm. The abscissa is contour radius, and the ordinate is ange length. Experiment results were plotted as dots. When sorted those dots as wrinkle -no wrinkle, a wrinkling limit curve may be tted. The region above this curve is failure zone, and below is safety zone.
This curve could be de ned mathematically as: This equation can be used to nd wrinkling results of different con gurations, and is also useful for con gurations not found in the experiments. Similar diagrams were drawn for all the other materials and thicknesses.
One can see safe and wrinkling zone on the above graph. The curve on the graph form a border between safe and wrinkling zones. In the safe zone, no wrinkling should be observed. Similar graphs can be constructed based on material types and thicknesses. A designer may use those graphs effectively to nd the optimum conditions.

Conclusions
In this study, the wrinkling defect in exforming process was investigated experimentally. Aluminum alloy conditions, sheet thicknesses, contour radiuses, and ange lengths were considered as experimental